Crouch Colin - The Queens Gambit Declined 5.Bf4 - Cadogan 1998

The Queen's Gambit Declined: 5 ~f4! Colin Crouch

CADOGAN ('hess LONDON. NEW YORK

Copyright © 1998 Colin Crouch First published 1998 by Cadogan Books pIc, 27-29 Betwi~k St.. London WI V 3RF Distributed in North America by The Globe Pequot Press, 6 Business Park Rd, P.O. Box 833, Old Saybrook, Connecticut 06475-0833, USA. Telephone 1-800 2430495 (toll free) All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without prior permission in writing from the publishers. British Library Cataloguing in Publication Data A CIP catalogue record for this book is available from the British Library

ISBN 1 85744 207 5

Edited by Graham Burgess and typeset by John Nunn for Gambit Publications Ltd, London. Printed in Great Britain by BPC Wheatons Ltd, Exeter

CADOGAN CHESS SERIES

Chief Advisor: Garry Kasparov Russian Series Editor: Ken Neat Editor: Murray Chandler For a complete catalogue of CADOGAN CHESS books (which includes the Pergamon Chess and Maxwell Macmillan Chess lists) please write to: Cadogan Books pIc, 27-29 Berwick Street, London WI V 3RF Freephone 0800 328 0242

Contents Symbols Preface

5 6

1

Historical Introduction

8

2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

The Old Main Line Introduction Alternatives for Black on Move 10 1O...~e7: Alternatives to 11ll:)d2 e5 11 ll:)d2 e5: Introduction 12 ~g5 d4 13 ll:)b3 "d8 12ll:)b3 "b6 13 ~g5 d4 (or 12 ~g5 d4 13ll:)b3 "b6) 12ll:)b3: Alternatives to 12.....b6 13 ~g5 d4 Conclusion

20 20 20 23 29 30 41 44 47 50

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10

The New Main Line Introduction 1O...~e7 11 h4 1O...~e7 11 g4 1O... ~e7: White's 11th move alternatives 10...dxc4 11 ~xc4 without 11...~e7 1O...~d7 1O... a6 1O.. Jld8?! 1O...ll:)e4 General conclusion on 10 0-0-0

51 51 52 63 73 76 77 85 87 87 91

4 4.1 4.2 4.3 4.4 4.5

Alternatives to the Main Lines for White Introduction White's alternatives on move 10 White's alternatives on move 9 White's alternatives on move 8 White's alternatives on move 7

12~g3

93 93 93 102 113 144

5 5.1 5.2 5.3 5.4

Black Avoids the Main Line: ...0 Systems Black's alternatives on move 9 Black's alternatives on move 8 Black's alternatives on move 7 The 5 ...c5 system

149 149 159 165 168

6

Systems with ••• b6

172

7 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8

Systems with ...c6 7 'ii'c2ltlbd7 7 'ifc2: Black delays ...ltlbd7 or ...0-0 4ltlf3ltlbd7: Transpositions and independent lines The Exchange Variation, 7 cxd5 7 ~d3 7 :cl 7h3 Odds and ends

195 195 212 214 217 220 224 225 226

8 8.1 8.2 8.3 8.4 8.5

Miscellaneous Systems on Moves 5 and 6 Systems with ... a6 6 ...ltlbd7: Lines not already examined Miscellaneous 6th moves for Black Miscellaneous 6th moves for White Miscellaneous 5th moves for Black

229 229 238 244 247 250

Index of Variations

255

Symbols + ++ # !! !? ?! ? ??

+± ;t

= 00

=+=

+ -+ Ch Cht tt

Wch Ech Wcht ECC Ct

IZ Z

OL jr wom mem rpd COlT

1-0 1/2- 1/2 0-1 (n) (D)

check double check checkmate brilliant move good move interesting move dubious move bad move blunder White is winning White is much better White is slightly better equal position unclear position Black is slightly better Black is much better Black is winning championship team championship team tournament world championship European championship World Team Championship European Clubs Cup candidates event interzonal event zonal event olympiad junior event women's event memorial event rapidplay game correspondence game the game ends in a win for White the game ends in a draw the game ends in a win for Black nth match game see next diagram

Preface When Murray Chandler first suggested that I write a book on the .i.f4 Queen's Gambit, he noted that no book had been written on this system. To my surprise, he was right. The exclamation mark at the end of the title was also Murray's idea. At first I would have preferred !?, but I now feel that I can live with !. To cover more or less uncharted territory is difficult in any field, but it is also exciting. Where knowledge has not been properly systematized, there is so much more room for radical discovery. For example, one of Black's most interesting defences to the .i.f4 system is 5 ... dxc4, which will take us outside existing published theory as early as move 7. In the more established lines, several innovations have been introduced in the text, the general effect of which is to strengthen these lines for White. In particular, the New Main Line with 10 0-0-0, which appears to have lost some of its popularity in grandmaster play in 1997, has been given a new shine in several critical positions. Improvements for Black are more likely to be found in the lesser-known lines. The .i.f4 system is a young system, even though some of its key ideas go back to Hastings 1895. White is aiming for simple development, the objective being to get his pieces out on good squares before Black, and thus to achieve superior piece mobility; pawn structures tend to be fluid. Black's choice is either to develop quietly, accepting this disadvantage, or to attempt a counterattack against White's pawn centre, opening up lines of attack before White's development is complete. The counter-attacking approach proved more successful for Black, and the intense exploration of the Old Main Line after 6 ... c5 in the 1970s and 1980s appeared to indicate a wholly satisfactory defence for Black. Counterattack by Black versus normal play by White led to dynamic equilibrium. Then in 1988 a radically new idea for White was discovered. The apparently totally reckless idea of castling queenside, into the face of what appeared to be a ready-made attack, was not only possible, but positionally justified and strong. It has to be said that the quality of some of the earlier games by top players in this variation was by grandmaster standards atrocious, with Black in particular suffering some unnecessarily severe defeats. This itself can be taken as an indication that many ofthe basic ideas were wholly novel. The theory of the variation has settled down a bit now, but it is salutary to remember that White's most dangerous plan (10 ... .i.e7 11 h4) was introduced only as recently as 1995; before then, everyone was pushing the wrong pawn on the kingside! So maybe it is only now that it is possible to set down the full case for the .i.f4 variation for White. A book published before 1995 would have missed the most important variations in the main line, while ECO, published in 1987, missed the

Preface

7

main line altogether, and was instantly obsolete as far as critical assessment of 5 ~f4 is concerned.

It is a fallacy to assume that any openings book can be completely 'objective', in that discussion of variations necessarily involves consideration of what each side is aiming for. This book is generally written from White's point of view, but where I have found improvements for Black I have noted them. Neither have I attempted to keep secrets or to mislead the reader on any position; if there are mistakes, and surely there must be, they are genuine mistakes. One can never achieve perfect accuracy in writing; if the new ideas and corrections of mistakes of others outweighs the introduction of mistakes from my own hand, then some sort of contribution has been made to chess literature. I have tried to trust the analysis of others as little as possible. In one case, Kasparov produced some poor analysis, overlooking at one point a simple back-row check; I noticed that a strong and respected grandmaster recently published a book that repeated this analysis verbatim and on trust. My approach has been encyclopaedic rather than selective. In part this is out of necessity. Where there is no existing book on an opening, the author has a much bigger responsibility to show why some variations are more important than others. No doubt many of the second-rate lines discussed in this book will be discarded when others write on this opening in several years' time; for the moment, it is theoretically important to show clearly why they are second-rate. Again because of the lack of literature on the subject, there is no bibliography as such. The main, indeed almost the only, sources are a couple of shelves of Informator, ECO volume D (1987) and a computer database of around a thousand games. It is hoped that the reader will find the book useful. I have tried to explain the themes that arise, and have continued variations often deep into the middlegame to make it clearer why particular assessments have been made. The reader as White who grasps the basic themes should be rewarded by some smooth positional victories and some attractive attacking wins.

Colin Crouch Harrow Weald December 1997

1 Historical Introduction Early days Although in the early years of the twentieth century the .1g5 system established itself as White's main choice in the Queen's Gambit (with 2 ... e6), a few years earlier the debate was much more open as to White's most promising plan of action. At Hastings 1895 three basic plans were tried, and if anything the .1f4 plan was the most successful. Gunsberg was the main advocate of systems of development with .1b2, a plan that becomes justified if Black breaks with ... c5. However, the bishop bites on granite if Black is content to strongpoint d5. Gunsberg was famously dismissive of the .1g5 system, noting that "no good results from this early sortie of the bishop" after 1 d4 d5 2 c4 e6 3 tLlc3 tLlf6 4 .1g5 in the celebrated Pillsbury-Tarrasch game. Many later writers have seen this as rather a quaint point of view, but in fairness it has to be added that Gunsberg also advocated the queenside fianchetto for Black, which is now seen as critical in many defensive and counter-attacking systems, particularly in the Semi-Slav. What we are concerned with here is the third way, neither i.g5 nor .tb2, but .1f4. Why did .1f4 lose its popularity, and never regain it until recently? And what new theoretical developments have led to it being seen as a dangerous weapon, even at top

grandmaster level? These are the questions to which we try to provide outline answers in this introductory chapter. The first point to note is one of elementary move-order. Some of the Hastings games started 1 d4 dS 2 c4 e6 3 tLld tLlf6 4 .tf4 (D), reaching the following position:

This move-order is inaccurate; White should insert 4 tLlf3 .1e7 before playing 5 .1f4. Black has two ways of equalizing in comfort: 1) 4 ... dxc4!? 5 e3 (or 5 e4 .1b4) 5... .!Lld5 !. 2) 4 ....1b4!. The second line would be my own personal preference, leading in effect to a good Nimzo-Indian, but both systems leave the white bishop exposed and ineffective on f4. The variations analysed in this book all start with the modern moveorder 1 d4 dS 2 c4 e6 3 tLlc3 tLlf6 4

Historical Introduction

ffi .te7 5 .tf4 (D), reaching the position in the next diagram:

9

von Bardeleben - Em. Lasker Hastings 1895

(notes based on C.Crouch and K.Haines, Hastings 1895; the Centenary Book).

B

1 d4 d5 2 e4 e6 3 tOe3 tOf64 .tf4?! .te7?! 5 e3 0-0 6 ffi b6 7 .td3 cS 8 dxe5 bxcS 9 0-0 tOc610 adS exdS 11 .te6(D)

:c1

5 ....tb4? would now be a simple loss of tempo, so the moves most commonly played here are 5...0-0, and occasionally 5 ... c5 or 5 ... c6. Curiously, 5 ... dxc4!?, although natural and logical, remains almost completely unexplored, a defect we try to remedy in this book. Leaving aside the point of moveorder on moves 4 and 5, a recurrent theme in the .tf4 games from Hastings 1895 was the battle against the 'hanging pawns' on c5 and d5, with Black's b- and e-pawns missing. Black proved completely unable to handle this pawn structure, as the result summary shows: Round 1: Lasker-Marco, 1-0,29; Round 4: von Bardeleben-Lasker, 1-0,58; Round 9: Teichmann-Marco, 1-0, 32. The two games Marco lost were massacres, while even World Champion Lasker was unable as Black to find a good way to back up his pawns, in a game which is still of theoretical interest today.

The classic hanging pawn position. The two black pawns cover a lot of squares in the centre, but they are also open to attack, and if either of the pawns advances, Black's grip on the centre disintegrates, leaving White with a clear positional advantage. Black's pieces remain active, however, and ECO assesses the position as equal, a verdict I would regard as 'not proven'. 12 tOe2 1t'b6 13 ~ tOxeS?! 13... tOb4!? 14 .txe5 lDg4 15 .tel 1t'd6 Black aims to solve his difficulties by means of a kingside attack, but this plan must be treated with some scepticism, as the pieces diverted to the kingside do not provide much cover for the centre.

10

The Queen's Gambit Declined: 5J.f41

16 g3J.d7?! 16... lOeS 17lOf4 ;to 17lOc4 "h6IS h3lOc6 (D)

Endgame play in those days was unsophisticated however, and Lasker was almost let off the hook. 24•••:ac8 24 ... J.g4!? 25 ':c3 :CdS 26 ':del 27 J.e4 :a6 28 l:lxc5? White should play 28 a3!: 28 .. J:tb6 29 :lc2 :db8 30 :xcS :xb2 31 ':xb2 l:lxb2 32 :as +- or 28 ...:d2 29 :lc2 :xc2 30 :xc2 :as 31 J.dS!? and e4, f4, ~f2, etc., with a decisive advantage. 28..•':xa2 ± 29 :b5 J.e6 30 :c7

:c6

There now follows a standard simplifying sequence which takes the sting out of Black's kingside initiative, leaving White with a strongly favourable endgame. 19 J.xC6! J.xC6 20 lOxd5 J.xh3 (20 ... J.xb2 21 :c4!) 21lOxC6+ "xr6 22 "h5 'ii'h6 23 'ii'xh6 gxh6 24 :Cdl

(D)

B

as?! 30...l:Id2!? 31 J.C5 :'1+ 32 ~g2 J.a2 32... J.xfS 33 :xfS :f8 34 :a7 a4 3S :f4 :a2 36 :b4 +- Gunsberg. 33 b3a4 33... ':a8 34 J.e4 +- Gunsberg. 34 bxa4 :d2 35 as h5! 36 a6 J.d5+ 37 e4 :aa2! 38 cJilf3! Avoiding the trap set by the everresourceful Lasker: 38 :xdS? :xf2+ 39 ~gl (39 ~h3?? :h2#) 39...:g2+ drawing. 3S••.:a3+ 39 ~C4 l:Ixf2+ 40 ~e5 J.xe4 41 J.xe4 :txa6 42 ~d4 :h6 43 ':c8+ ~g7 44 ~e3 :ff6 45 :g5+ and Black soon resigned. For those who would argue that this is old hat, and that we now know how to keep the hanging pawns safe, this more recent game should act as a corrective.

A successful early outing for the system! White should win this easily, given that the cS-pawn will almost inevitably fall, leaving White with an active 2-1 queenside pawn majority.

Yusupov - Ljubojevic 1ilburg 1987 I d4lOc6 2 c4 e6 3 ffi d5 4lOc3 J.e7 5 J.C4 0-0 6 e3 b6 7 :el c5 S dxeS

Historical Introduction bxeS 9 .i.e2 .i.b7 10 0-0 lObd7 11 cxdS exdS?! (D) 11...lOxd5.

12lbes! Yet again White sets up a powerful attack against the hanging pawns, with .i.f3 in prospect. As in the von Bardelehen-Lasker game, it helps White to have his bishop on f4 rather than on g5, in that support is given for the knight to land on e5. 12.••1Ob6 12... lOxe5 13 .i.xeSlbd7 14 .i.g3 ± Yusupov. 13 a4! a5 14 .i.f3 :e8 15 lOb5! lIa6(D)

11

16lbd3 Black now has a lifeless position, and Yusupov points out that the quiet 16 b3!, preventing c4, would present grave difficulties. The text leads to interesting but unnecessary complications, which ultimately but not wholly convincingly favour White. 16...c4 17 .i.c7 "d7 18 lOe5 "c8 19 b3lba8! 20 bxc4lbxc7 21 cxdS A radical way of dealing with the hanging pawns! White certainly has full compensation for the piece, but no more than that, according to Yusupov. 21•••.i.d6! 22 lbc4 .i.eS? 22....i.b4 keeps Black in the game. 23 lbxaS! :xaS 24 :xc5 "d8 25 lbd6!! An attractive tactical resource; if 25 ...:xc5, 261Oxb7 "fIe7 27 d6 wins . 25.....xd6 26 :xaS .i.xd5 27 "d4 "e6 28 .i.xd5lbcxdS 29 :dl 30 :eS :XeS 31 "xc5 h5 32 as 33 h3 g6 34 "c6 "b4 35 a6 "as 36 "b7 "a4 37 l:[bll-O

:eS "e4

Going back to the early years, at Hastings 1895 Black was also called upon to handle a couple of isolated queen's pawn positions, a task which was achieved more successfully than that of holding the hanging pawns. The early part of Teichmann-Janowsky from round 8 has a modem appearance, with Black successfully liquidating with an early ...d4. Teichmann - Janowsky Hastings J895 1 d4 d5 2 c4 e6 3 lbc3 lOr64 .i.f4?! .i.e7?! 5 e3 0-0 6lbr3 eS 7 dxeS .i.xeS 8 cxdS!?

12

The Queen's Gambit Declined: 5 ll.f4!

Nowadays 8 'ii'c2 lOc6 9 a3 is regarded as the main line, but there were times during the 1980s and early 1990s when the main line was not looking too promising, with attention therefore switching to the attempt to make something out of the immediate pawn exchange. 8 •••exclS This is now firmly out of favour, but was one of Black's main defensive systems (via 7 ...lOc6 8 cxdS exd5) as recently as the early 1970s. 8 ...lOxdS 9 lOxdS exdS is now preferred, the positional threat of ... i.b4+ in most cases persuading White to spend a move on 10 a3. Then after 10...1Oc6 II i.d3 i.b6 12 0-0, Black can try 12 ... d4 or 12...i.g4, in the spirit of the Teichmann-Janowsky game, or 12...'ii'f6. 9i.d3 9 i.e2 is now regarded as a better way to keep the pressure on the isolani. 9 •••lOc6 10 0-0 i.g4! (D) Then as now the correct development of the bishop. The isolated pawn is a weakness, but if Black plays actively he can hold the balance. It is only when Black is passive that the d-pawn becomes a problem. Steinitz-Burn, Hastings 189S continued 1O...i.e6?! 11 l:lcl l:lc8 12 i.bl 'ii'a5?! 13 lOd2 i.e7 14 lOb3 'ii'd8 IS 'ii'd3 "d7 16 l:lfdl with a clear edge. Few now would agree with Nimzowitsch that "the BK3 [white e3-bishop or black e6-bishop] belongs to the PQ4 [white d4-pawn or black dS-pawn] as does a nurse to a suckling child"; often the 'KNS' development (i.g5 for White; ...i.g4 for Black) is necessary to maintain the momentum of the position. On

this particular point, Janowsky was ahead of his time.

11l:lc1 Chasing the bishop is always a possibility in this type of position. Here though, 11 h3 i.hS 12 g4?! is unconvincing after 12...i.g6 13 i.xg6 fxg6 14 g5lOhS 15 "xdS+ (1SlOxdS can be met by 15 .....d716lOf6+lOxf617 gxf6 GO or 15 ...Wh8) lS .....xdS 16 Illxd5 lOxf4 17 exf4 (17lOxf4 i.d6 =1=) 17...J:lad818l:ladlJ:lfS+. White's kingside pawns have been weakened more than Black's. 11•••d4! Nicely judged. 12 lObS After 12 lOa4 i.e7 13 exd4 lOxd4 White has problems, as 14 i.xh7+?? Wxh7 15 "xd4 i.xf3 16 "xd8 J:lfxd8 17 gxf3 l:ld4 wins a piece for Black, a trick which shamefully I missed in writing my 1895 book. 14 i.c7 "dS! IS i.xh7+?? also fails: Is ... lOxh7 16 'ii'xd4 i.xf3. 12.••i.b613lObxd4lOxd4 14 exd4

"cIS! (D) 14... i.xd4?! 15 h3! . The text-move is an elegant way of dealing with the IQP problem. Black

Historical Introduction

w

sacrifices the weak isolani in order to give White an even weaker one; furthermore, Black makes full use of the blockading square in front of the pawn.

15 ~e5 :ac8 16 .e2 ~xd4 17 ~xd4 ~xl3 18 .xl3 :xcll9 :Xci .xd4 20 :c3 :dB 21 h3 21 "xb7?? ~4! is winning for Black. 21...•b4 and Black is slightly better, the knight being able, unlike the bishop, to reach the dark squares. The game was later drawn after several inaccuracies. After Hastings 1895, the ~f4 system was to remain comparatively rare. Tarrasch-Maroczy, Budapest 1896 started 1 d4 e6 2 c4 d5 3 1Oc3 lDf6 4 ~f4, and now 4 ...dxc4, "a new defence which is perfectly good", according to Charousek. On 5 e3 lDd5 6 ~xc4 lDxf4 7 exf4 Black already stood comfortably, even after the passive 7 ...c6. Black later won the game. The secret was out; 4 ~f4 gives White nothing, and the clear-cut 4 ~g5 came to be preferred. We now know that 4 lDf3 ~e7 5 ~f4 is a perfectly valid try for an edge, but there are complications in that Black need

13

not meet 4 lDf3 with 4 ... ~e7. He could also choose a Semi-Slav with 4 ...c6, or a Semi-Tarrasch with 4 ...c5, while 4 ... ~b4 is another possibility for Black. Systems with 4 ~g5 or 4 cxd5 cut down Black's options, making preparation simpler. Naturally the problem of how to deal with 4 ...c6 or 4 ...c5 still has to be faced by the player who wants to try the 5 ~f4 system. It is not proposed to deal with this particular topic in this book, but it is something the reader will have to be aware of.

The Recent Revival For the types of reason listed above, the ~f4 system remained very much a secondary plan, a minor alternative to ~g5, for several decades. Connoisseurs of the Queen's Gambit such as Rubinstein and Capablanca played it occasionally, and Bogoljubow tried it more often, but the system never really moved out of the shadows. Yet results with the ~f4 system were good. Around the early 1970s, the typical pattern was that White gained an advantage in pawn structure and squeezed relentlessly. The next game is an example.

Gbeorghiu - Rubinetti Mar del Plata 1971 IlDf3lDf6 2 c4 e6 3M dS 4 d4 ~e7 5 ~f4 cS 6 dxcS lDa6 7 e3 At around the same time, Portisch had some smooth wins with 7 ~d6. 7.••lDxcS 8 adS exd5 9 ~b5+ ~d7 10 ~xd7+ .xd7 11 0-0 0-0 12 ~e5 (D)

14

The Queen's Gambit Declined: 5 ~f4! 22.••~gS 23 ~4lbxr3+ 24 ~xf3 25 j.b2 lIdd8 26 ~d4 'ifd7 27 'ifc7 (D) ~e4

B

The start of a prolonged positional campaign against the isolated pawn. Such a pawn is not necessarily in itself a decisive weakness, but if the player with the isolani cannot achieve active piece-play, the prospects are grim. This is the case here.

12•.•lIfd8 13 'ike2 86 14 lIfdl 'iff5 15 j.c7 lId7 16 ~d4 'ifg6 17 j.eS ~fe418:act lieS 19 ~3 'ifOO 20 b3 j.f8 21'ifc2 Meandering play, but also constructive and accurate play. White shifts his pieces around with no obvious grand plan in sight. but places reliance on the fact that it is difficult for the defender to cover everything adequately. White's first priority is to keep just a little extra piece flexibility, so that Black can never fully equalize; then he can be perpetually on the lookout for a way to take advantage of any momentary lapse in the defence in order to gain ground. It is this purposeful tacking around, rather than the attempt to follow a predetermined plan, that is the secret of successful play against a static pawn weakness.

21 •.. b622~2 Making real progress; the other knight gets a chance to head for d4.

White has kept a firm grip on the position, and now progresses towards an endgame in which all Black's queenside pawns, and not just the isolani on d5, are weak. White's knight outpost on d4 is more secure than Black's on e4, a point which becomes even clearer after queens are exchanged.

27••. j.d6 28 'ifxd7 lIxd7 29 ~S j.e5 30 j.xe5 :XeS 31 lIc8+ ~b7 32 ~4 ~f6 33 lIdct lieS 34 lISe7 lIed8 35 f3 g6 36 lIxd7 :Xd7 37 lIc8 ~g7 38 ~f2 ~b7 39 cRe2 40 ~d3 bS 4184 lIe7 42 a5 ~d7 43 'itd2 ~6 44

m

b4~e84S:d8

The first attempt for over 20 moves to make a direct attack on the d-pawn. Static weaknesses don't run away, so it is perfectly in order to be patient.

4S••• ~7 46 ~d3 cRf6 47 ~b3 ~e5 48~cS (D) . Black's queenside is now under clamps, and White has a completely free hand to use his kingside pawn majority. The immediate threat is f4+ followed by lId6.

48•..g5 49 g3 f6 50 b3 ~b5 51 f4+

Historical Introduction

gxf4 52 gxf4+ ~5 53 :xd5+ ~g6 54 l:[d7 l:lxd7+ 55 .fu:d7 Ci)a7 56 e4 tDc6 57 'iii'c4 ~7 58 e51-0 In/ormator references to the i.f4 system in the early 1970s were dominated by smooth wins for White, although it has to be acknowledged that the worst game of the 1972 FischerSpassky match (game 14, Fischer as White) finished as a draw after several serious mistakes by both sides. Players at the time were quick to draw the lesson that Black's play needed to be sharpened to avoid the risk of being slowly ground down. Sometimes indeed Black's play became overly sharp, as in this game from a later world championship match.

Korchnoi - Karpov Baguio City Wch (21) 1978

.c2

1 c4 Ci)f6 2 Ci)c3 e6 3 Ci)f3 d5 4 d4 i.e7 5 i.f4 0-0 6 e3 c5 7 dxc5 i.xc5 8

tDc6 (D)

This position remains theoretically critical for the i.f4 variation. Black develops rapidly and aggressively, hoping to ensure that White is not given the time to build up the sort of

15

clear long-term positional advantage that we have seen in previous games. Indeed, if White tries anything too elaborate there is the danger that he will find himself suddenly stuck with an exposed king in the centre and an exposed kingside. To give one example, the line 9 :dl WaS 10 a3 i.e7 II .J:d2 briefly aroused interest in the early 1980s, but it was soon found that 1l...C/)e4! 12 Ci)xe4 dxe413 Wxe4 IId8 followed by a timely ...e5 gives Black excellent attacking chances. 9 :dl .a5 10 a3 This is what we describe as the 'Old Main Line'. The usual move here is lO ...i.e7, which is satisfactory for Black. Karpov now tried a prepared 'improvement', which Korchnoi impressively demolished over the board. 10...:e8?! Preparing to sprint through the barricades with ...e5 and ...d4, but White's position is not taken so easily. lllDd2! A standard response in the mainline systems. The knight breaks the pin on the as-el diagonal (so that b4 is now a threat, without Black having ... Ci)xb4), blocks any ...C/)e4 move by Black, prepares a knight recapture on

16

The Queen's Gambit Declined: 5

c4 should Black exchange, and aims to play ~b3, dispersing Black's advanced queenside forces. 1l... eS 12 J.gS ~d4 Still prepared analysis. Black's attack runs out of steam after the other piece sacrifice 12...d4 13 ~b3 'ii'b6 14 ~a4 J.b4+ 15 axb4 Wxb4+ 16 ~d2 e4 17 J.xf6 gxf6 18 ""3. 13 "bl! White calmly ignores the sacrifice, avoiding such horrors as 13 exd4? exd4+ 14 ~2 ~g4 15 b4 d3 16 'ii'xd3 J.xt2#. After the text-move, Black faces a common problem in the old main line: once he has achieved ... e5, the d5square becomes particularly weak, especially after J.xf6 by White. Black must continue to play very aggressively to justify his earlier play. 13•••J.rS 13 ...J.g4 14 J.xf6 gxf6 15 ~xd5 J.xdl 16 ~xdl ± Speelman. 14 J.d3 e4 (D)

15 J.c2! Timman, in The Art of Chess Analysis, claimed 15 J.n as a win, suggesting that Kholmov's line of 15 ... ~g4, with the idea 16 ~xdS ~e5 17 exd4

~f4!

e3 18 'ii'xf5 ~f3+ 19 'ii'xf3 exd2#, is refuted by 16 cxdS. This claim is made without considering any variations, and is made on the hardly earth-shattering grounds that the c4-square is freed and the black knight remains trapped. But with all Black's pieces in play, and White undeveloped and uncoordinated, it would be amazing if Black could not create dangerous attacking possibilities. After 15 J.n ~g4 16 cxdS, Black plays as before 16 ...~S, and is ready to meet such direct approaches as 17 ~c4 ~ef3+ 18 gxf3 ~xf3+ 19 ~e2 Wa6, or 17 exd4 ~f3+ 18 gxf3 exf3+ 19 J.e3 (19 ~de4 llxe4+) 19... J.xbl with very sharp play, probably favouring Black. Maybe Timman had in mind 17 J.f4, threatening J.xeS followed by ~4. Then the attempt to be super-cool with 17... bS?! backfires after 18 J.xeS llxeS 19 WeI! when exd4 becomes a real threat, but the critical line would seem to be 17 ... ~d3+ 18 J.xd3 exd3 19 0-0 ~2+ 20 ~xe2 dxe2 21 Wxf5, and now perhaps 21...exdl'ii' 22 llxdl 'ii'b5, with the idea of ...We2, with unclear play. [Later note: In the recently published second edition of his book, Timman analyses 17 exd4 J.xd4 18 J.e3 to a win for White, but either ignores or overlooks the possibility of 17...~f3+.] It's certainly a long way from the win claimed by Timman. Korchnoi's judgement in avoiding this line is vindicated, in that he achieves a clear edge in the line he plays.

IS•..lDxc2+ 16 .xc2 Despite the exchange of White's bishop, Black still comes under severe pressure on the light squares.

16•.••a6!?

Historical Introduction 16... dxc4 17 i.xf6 gxf6 180-0 ± Thuman. 17 i.xf6 .u6 IS lDb3 i.d6 19 lhdS So White has won a pawn, but Black's bishop-pair gives him the opportunity to create trouble. With some regret, we give only brief notes to the rest of the game, which is a real heavyweight struggle, as this is an openings book not a game collection. 19...:e5 19 ...i.e5 20 lDd4 i.xd4 21 :xd4 "gS 22 g3! ± Andersson (cited by Timman). 20 lDd4 :cS 21lheS Timman gives 21 f4 as a clear win. Again one must dispute this. After 21...:xdS 22lDxdS Wh4+, Black is in the game on either 23 g3 Wh3 24 lDxfS WxfS 2S "b3 hS, when White's kingside still needs watching, or 23 Wf2 Wxf2+ 24 ~xf2 i.d7, which is if anything good for Black. 21 ..."xeS 22lDxfS .US (D)

view of his control of the d-file and queenside light squares. Snatching the e-pawn is bad: 23 Wxe4? .xe4 24lDxe4 :xc4 2s1Dc3 i.xa3 +, or 23lDxe4? b5 24 0-0 :xc4 :j: Filip.

23•••:xc4 24 :dl 'ii'eS 25 g3 a6 26 'ii'b3 bS 27 a4! :b4 28 .dS .xdS 29 lhdS i.f8 30 nbS as! 31 :dS :xb2 32ltaS fS 33:xaS (D)

White proceeded to win the endgame after a titanic struggle. 33•••i.b4 34 :as+ ~f7 35 lDa4 :bl+ 36 ~g2 i.d6 37 :a7+ ~f6 3S b6 i.bS 39 :as i.e5 40 lDcS .td6 41 b7 ~e7 42 :g8 .teS 43 f4 eu3+ 44 ~~f7(D)

230-0 Returning the extra pawn in order to complete his development. White can now look forward to a very pleasant late middlegame and endgame in

17

18

The Queen's Gambit Declined: 5 1./4!

45I:tc8?! Timman (based on joint analysis with Andersson) gives as winning 45 I:td8 ~e7 46 I:td7+ ~e8 47 e4 I:tb5 48 I:td5 fxe4+ 49 ~xe4 ~b8 50 ~d3 ~a7 51 ~c4 I:tb6 52 ~4! followed by I:tb5. 45••• ~e7 46 h3 h5? Andersson suggested that 46...I:tb5 probably still holds for Black. 47 I:tg8 ~7 48 I:td8 g5 49 g4 bxg4+ 50 bxg4 9:;e7 51 llg8 fxg4+ 52 ~xg4 ~7 53 I:tc8 .i.d6 54 e4 I:tgl+ 55 ~5 g4 56 e5 I:tn+ 57 ~e4 I:tel+ 58 ~d5 I:tdl+ 58 ...~xe5 59ll:)d3 +-. 59ll:)d3 :Xd3+ 60 ~c4 1-0

I d4ll:)f6 2 c4 e6 3ll:)f3 d5 4 M ~e7 5 .i.f4 0-0 6 e3 c5 7 dxeS ll:)c6 8 Wc2 ~xeS 9 a3"aS 100-0-0! (D)

However, it has to be emphasized that Black's results in both theory and practice proved satisfactory after 1O... ~e7, despite various attempts to improve White's play. The ~f4 system proved modestly fashionable during the early 1980s, but then interest gradually waned. Many of the side systems that were tried, such as 8 ~e2 or 8 cxd5, were modestly successful, but no more than that. Then suddenly Grigory Kaidanov's idea of castling queenside, into what appears to be the centre of the storm, rejuvenated the ~f4 system; White's king is safer than it looks, and it turns out that it is not so easy for Black to hold both centre and kingside. Many first-class scalps have been collected in double-quick time using Kaidanov's idea; several of these games are given in full in the theoretical section of the book. In the meantime, just consider what effect it had on the chess public when in a Candidates' Match Jon Speelman played the latest and freshest idea on an

Upping the ante considerably. In a system which had previously been considered quiet, positional and only dangerous if Black played insufficiently aggressively, White meets fire with fire. Clearly White has it in mind to launch a kingside pawn-storm, but this is not the only justification for castling queens ide instead of playing I:td 1. The point is that the white king is better placed on cl than on el, in that White no longer has to take such great care to deal with counterplay along the as-el diagonal. This means for

unsuspecting Short, scoring a critical victory. The story was that a critical game from the Soviet Championship happened to be published in a Norwegian newspaper, and Tisdall, a member of Speelman's team, happened to be travelling in from Oslo, while Short was not privy to this particular channel of information.

Speelman - Short London Ct (3) 1988

Historical Introduction example that White's pressure on the d5-pawn is also accentuated. White's whole idea is so counterintuitive that it was not even mentioned as a possibility in the 1987 edition of ECO; it is 'obviously' wrong ... until you take a close second look! 10••• j,e7 11 g4 Even this move has been superseded, in that ll...dxc4 12 j,xc4 e5 13 g5 exf4 is now regarded as a satisfactory reply for Black. Indeed there was a danger for a while that White would lose interest in this system, until Kasparov introduced 11 h4! in 1995. 11...:d8 12 h3 a6 13 ti)d2 e5?! As in the :dl variations, this has to be timed very carefully to avoid disaster on the light squares. Such timing is absent here. 14g5! Forcing a knight move, as 14...exf4? 15 gxf6 j,xf6 16 ti)xd5 is unbearable. 14••• ti)e815 ti)b3 "b6 (D)

16 ti)xd5 So the decision comes not through a kingside attack, but rather as a result

19

of the acceleration of the attack against Black's centre, made possible by getting White's king out of the central zone. 16•••:xd5 17 cxd5 exf4 18 dxc6 fxe3 19 fxe3 j,xg5 20 ~bl bxc6 21 j,c4 (D)

The liquidation of the centre has left White with a slight material advantage and, more importantly, a massive advantage in piece activity and coordination. 21 .••:a7 22 :hf1 j,f6?! Speelman gives 22 ...:e7 as the only chance to resist. 23 "e4 ~f8 24 "xh7 g6 25 e4 c5 26 e5 j,g7 27 e6 1-0 In very general terms, White's best approach in the j,f4 system is to try to create a slight but persistent positional edge if Black plays quietly, but to play very aggressively, with queenside castling, if Black aims for counterplay with the ...c5, .....a5 plan. We now move on to the theoretical analysis of specific variations.

2 The Old Main Line 2.1 Introduction 1 d4 d5 2 c4 e6 3 M lDf6 4ffi i.e7 5 i.f4 0·0 6 e3 cS 7 dxcS i.xcS S Wc2 lLlc6 9 a3 WaS 10 :dl (D) For 10 0-0-0 see Chapter 3. For other moves (10 l:c1; 10 lLld2!?) see Chapter 4.1.

not least because Black could consider 9 ... i.e7!? instead of9 .....a5, avoiding the 'new main line', and challenging White to find something other than 10 :d1 This point is discussed further in Chapter 5.1, Line H. From the diagram, there is a major branching point after 1O ... i.e7 11 lDd2 e5, when 12lDb3, 12 i.g3 and 12 i.g5 all need to be considered. Before considering these main lines, we examine alternatives on moves 10 and 11.

"as.

2.2 Alternatives for Black on Move 10

This position was the main focus of theoretical debate through the 1970s and most of the 1980s. White now threatens b4, so Black usually replies 10 ...i.e7. White is clearly ahead in development on the queenside, but it will take him a long time to get his king into safety. This allows Black time to build up a counter-attack; the thematic plan is, if White plays lLld2, to reply with ... e5 followed by ... d4. Play usually becomes quite sharp, but the general impression is that Black holds his own in the complications. As a result, 10 0-0-0 is currently preferred. The old main line remains important however,

1 d4 dS 2 c4 e6 3 lDc3 lDf6 4ffi i.e7 5 i.f4 0-0 6 e3 cS 7 dxcS i.xcS S Wc2 ~9a3WaS 10:dl Now 10... i.e7 is clearly best, and the alternatives are mostly inferior. We consider: A: 10.•.:eS?! 20 B: 10•••dxc4 20 C: 10...i.d7!? 21 D: 10...:dS 21 E: 10...lDe4?! 21 A)

10...:eS?! is probably too ambitious. See the Korchnoi-Karpov game in the Introduction for further comment. B)

10... dxc4, without even waiting for lDd2, gives White a free hand. 11

The Old Main Line i.xc4 i.e7 120-0 eS 13 i.gS i.g4 14 i.xf6 i.xf6 IS ltldS, TukmakovO.Rodriguez, Las Palmas 1978, gives excellent central control. If IS ...i.d8?, 16 b4 'ifxa3 17 :alltlxb4 18 'ifd2 and White wins.

C) 10•.• i.d7!? could be considered, since 11 b4? is met by 11...li:lxb4 12 axb4 i.xb4 13 :clltle4 14 i.eS f6 IS i.d4 eS. White should probably first play 11 li:ld2, when 11.. .i.e7 transposes into Chapter 2.3, Line C.

D) 10•••:d8 (D)

21

12li:lxc4 (12 i.xc4 i.e7 13 0-0 i.d7 14 i.e2 i.e8 IS ltlc4 "fS 16 'ifxfS exfS 17 i.f3 :ac8 18ltld6 is slightly better for White, Balashov-Averbakh, Moscow 1978) 12...:xdl+ 13 'ifxdl "d8 14 'ifxd8+ (14 'ii'c2li:ldS IS i.g3 li:lxc3 16 'ifxc3 'ife7 17 i.d3 i.d7 18 0-0 bS =Gallego-San Segundo, Spanish Ch 1996) 14... ltlxd8 IS i.e2 (winning a pawn with IS li:la4?! i.e7 16 li:lab6 axb6 17 li:lxb6 18 li:lxc8 l:tcS gives Black a big lead in development, Euwe-Kupper, Zurich 19S4) IS ... li:ldS!? 16 li:lxdS exdS BagirovMatanovic, USSR-Yugoslavia 1971, and now instead of 17 i.d6 b6 =, Matanovic suggests 17 li:ld6 with a slight advantage for White.

:a5

E)

10•••li:le4?! (D) has a poor reputation for the wrong reason.

w

This removes an important retreatsquare, as seen after 11 It:ld2: 1) l1...eS? 12 i.gS d413li:lb3 'itb6 14 i.xf6 gxf6 IS It:lds +- R.BatesD.Vergun, Bratislava U-14 Wch 1993. 2) 11...d4? 12ltlb3 'ifb6 13 ltla4 i.b4+ 14 axb4 'ifxb4+ IS li:ld2 'ifaS 16 b3li:lb4 17 'ifbl d3 18 f3li:lhS 19 i.g3 and Black has not got enough for the piece, Gheorghiu-P.Ostojic, Monte Carlo 1969. 3) 11 ... dxc4 is more sensible, but White ~an play for an endgame edge:

11 cxdS 11 i.d3?! h6 12 cxdS (12 i.xe4? dxe4 13 'ifxe4? i.xa3) 12... ltlxc3 13 bxc3 exdS 14 0-0 i.e7 (threatening ...'ifxa3) IS a4 i.f6 16 li:ld4li:les =Xu lun-Adianto, Dubai OL 1986. 11 ...exdS 12 :xdS li:lxc3 13 bxc3 'ii'xa3

22

The Queen's Gambit Declined: 51./4!

13 ...lDe7 can be met by 14lLlg5 g6 15 :d2 i.xa3 16 i.c4 i.f5 17 "b3 ± Gipslis-Shashin, Moscow 1970, or, perhaps less convincingly, 14 l%dl i.f5 15 i.d3 i.xa3?! (15 ...i.xd3 16 ':xd3 "xa3 ;!; Nepominshay) 160-0 i.xd3 17 :xd3lDg6 18 :al ± Nepominshay-Marciano, St Petersburg 1996. 14lDgS Who can resist threatening mate in one? (But see below for White's best move.)

14.••g61S i.c4 (D)

Black's a-pawn will be a tremendous asset for the endgame, if the game ever gets that far! Kingside disasters have tended to intervene, for example 15 ...lDd8? 16lDxh7 ~xh7 17 %lh5+ q;.g7 18 i.e5+ f6 19 ':g5 1-0 Portisch-T.Berger, Amsterdam 1964; or 15 ...~g7? 16 0-0 i.e7 17 e4i.f618 e5 i.xg5 19 i.xg5 i.e6 20 i.f6+ ~g8 21 'iWd2 lDe7 22 i.d3 :fc8 23 :a5 "xc3 24 "h6 1-0 G.Schroll-Wohlmann, Austria 1989. 15 ...i.f5?! led to desperate complications in Chekhov-Azmaiparashvili, USSR 1981: 16 :xf5 gxf5 17 O-O! lDe7 18 i.e5 i.d6 19 "e2 i.xe5 20 "h5 ~g7 and then 21 "xh7+ ~f6 22

f4 i.d4!! (22 .....xc3? 23lDe4+ fxe4 24 fxe5+ ~xe5 25 "g7+ f6 26 "e7#; 22 ...i.xc3? 23 g4! followed by lDe4+ and g5#) 23 exd4 "xc3 24 "h6+ lDg6 25lDh7+ ct;e7 26 "g5+ f6 (26 ...~d6? 27 "f6+ 1;c7 28 i.xf7 +-) 27 "xg6 with an unclear position, later drawn. Great fun to play through, but White can do better by delaying "xh7+ and playing instead 21 f4! (not mentioned by Chekhov). Then if 21...i.xc3 or 21.. ...xc3, 22 "xh7+ transposes into the variations given by Chekhov, while after 21...i.d4 22 exd4 "xc3 23 i.xf7! (though not 23 "xh7+ ~f6 24 i.xf7?? :h8) White should win. Finally, 21...h6 is answered by 22lDxf7. But the main improvement is Black's ... 15.•• b6! (D)

Black intends to defend by exchanging light-squared bishops. 160-0 16 lDxh7 ~xh7 17 :h5+ ~g7 18 i.h6+ ~f6 19 i.g5+ ~g7 is only a drawing line. 16... i.a6 17 i.xa6 W'xa6 18 h4 :ad8 19 :fdl "c4 20 :d7 i.e7 ; L.Ravi-D.Donchev, Calcutta 1996 Oater drawn).

The Old Main Line

23

Does this mean that 1O... lLle4 is a simple equalizing line, and that players with the black pieces need not read the rest of the chapter? Not so; return to the position after 13 ...Wxa3 (D).

w A: 11 J.e2?! 23 B: 11 :d2 24 c: l1lLld2 without 11•••eS

28

A) 14 lLlg5?! takes the knight to the wrong side of the board. Instead of chasing the king, White can chase the queen with 14lLld2! threatening lLlc4. After 14...Wal+ 15lLlbi J.e716J.d3 Black has massive difficulties with his queen; other lines also leave Black with problems. So 10... lLle4 is, after all, dubious, but not for the reasons given by existing theory.

11 J.e2?! lLle4 12 cxdS lLlxc3 13 bxc3 exdS 140-0 J.e6 (D)

2.3 10....i.e7: Alternatives to 11 lLld2 e5 1 d4 dS 2 c4 e6 3lLlc3lDf6 4 ffi J.e7 5 J.f4 0-0 6 e3 cS 7 dxcS J.xcS 8 Wc2 lLlc69 a3"aS 10 l:[d1 J.e7 (D) 11 lLld2 is very much the main line, aiming to harass the black queen, and tempting, ipdeed almost forcing, Black to unbalance the central pawn-structure with 11...e5. Our lines are:

The biggest surprise about this position is that White has achieved 75% from it in grandmaster play! His queenside pawns are too weak to allow him to play for an edge. Now there are two possibilities for White: 1) 15 l:tbl 'it'xa3 (15 ... b6?! 16 a4) 16 :xb7 :fc8 17 :cl, and instead of 17 ... J.f5 18 'it'dl a5 19 J.b5lLld8 20 l%c7 00 Miles-Van der Sterren, Wijk

24

The Queen·.s Gambit Declined: 5 j.f4!

aan Zee 1984, perhaps 17 ... ~!? 18 l:lc7 -*.f6:j:. 2) IS a4:tc816l:lbl-*.f617lhb7 "xc3(17...-*.xc3? 18ll:)gS; 17... lLlh4? 18 'ifb1-*.xc3 19lLlgS) 18 "xc3 -*.xc3 19 l:lcl, Conquest-Greenfeld, Hastings 198516, and if Black is content to liquidate to safety, then simply 19...d4 =(instead of 19...-*.f6 20 h3 h6 21 l:lc7 when White is slightly better). B)

11 l:ld2 (D) was introduced by Lajos Portisch in 1982, and had some initial successes through surprise value. However, Black has several ways of equalizing.

B

The rook move looks a bit strange. but it breaks the pin on the as-el diagonal, and does so without decentralizing the king's knight, and without allowing ... eS. And if White can develop in peace, he might well be able to double rooks profitably on the dfile. Black's main choices are 11...l:ld8, keeping the tension on the d-fiIe, and 1l...lLle4, a promising gambit line. A few other moves have also been tried, so our variations are:

Bl: 11••• l:le8?! B2: 11...h6 B3: l1...dxc4 B4: 11...l:ld8 BS: 11...lLle4!

24 24 24 2S 27

B1)

11...l:le8?! misplaces the rook. After 12 cxdS ll:)xdS 13 ll:)xdS exdS 14 i.d3 h6 150-0 -*.f6 16 l:lfd1 Black is worse off than in corresponding variations with the rook on d8 (Line B4 below). Not, however, 12ll:)b5? e5. B2)

The waiting move 11 ...h6, denying White use of g5, has been tried with success by Geller. 12 cxd5 exdS 13 -*.e2 (13 'ifb3?! i.e6 14 "xb7 l:lac8 15 -*.b5 d4! 16 exd4 ~S gave Black a massive attack in H.Grunberg-Geller, Sochi 1983) 13 ...i.e6 140-0 l:lac8 IS l:lfd 1 bS! 16 -*.xbS ~ 17ll:)xe4 dxe4 18 -*.xc6 exf3, and Black is OK despite the pawn count. Meduna-Geller, Sochi 1983 continued 19 "a4 (19 'ife4? -*.b3) 19... 'ifb6! 20 -*.dS (20 i.xf3 -*.b3) 20...-*.xdS 21 l:lxdS "xb2 22 "xa7 -*.h4 23 g3 i.f6 =. B3)

l1...dxc4 12 -*.xc4 and now: 1) 12...eS?! is bad; Black weakens himself too much on the light squares, for example 13 -*.g3 -*.g4 14 0-0 -*.xf3 IS gxf3 l:lac8 16 -*.a2 bS 17 ~S! ll:)xdS 18 -*.xdS ~419 "dl l:lcd820 -*.xf7+ lhf7 21 exd4 exd4 22lhd4 a6 23 l:lxd8+ "xd8 24 'ifb3 ± A.Petrosian-Ubilava, Telavi 1982. 2) 12 ... lLlhS may be considered. After 13 -*.d6 -*.xd6 14 l:lxd6ll:)f6, the game Adorjan-Tisdall, Lugano 1983

The Old Main Line continued IS .d2 CiJe7 16 e4 'ircs 17 .d4 .xd418 :Xd4 a619lDeS bS 20 ~b3 ~b7 =. The most likely place to look for an improvement for White is move IS. Not IS O-O?! .cS 16 lDbS lDeS! ao (16 ... a6?! 17 b4), but rather 15 :td2, for example IS ... a6 160-0 bS 17 ~e2 ~b7 18lDe4!? 84) 1l•••:td8 (D)

This was Black's reaction in the inaugural Portisch-Balashov game; these days White usually responds by exchanging pawn~ 12cxd5 . 12 h3?! is, as one would expect, a serious loss of tempo. In AgzamovLputian, USSR Ch 1982, Black continued vigorously with 12 ... lDe4! 13 lDxe4 dxe4 14 'iIIxe4 :txd2 (14 ... eS! ECO) IS lDxd2 eS 16 ~h2 (16 ~g3 ~e6 17 ~e2 :d8 18 .c2 ~fS +) 16 ... ~e6 17 g4 l%d8 18 'iIIc2 lDb4! with a strong initiative. The gambit with ...lDe4 is dangerous enough without the gift of an extra tempo. 12 ~e2 lDe4 13 lDxe4 dxe4 14 'iIIxe4 transposes to lines given under 11...CiJe4 (BS below).

25

The author toyed with the idea of 12 cS several years ago, but again 12•..lDe4! is a fully adequate reply, and 12...•xcS 13 b4 .b6 14 lDa4lDxb4 is also worth considering. 12••. lDxd5 12...exdS? 13lDb5! jams up Black's queenside, for instance 13 ...lDe8 14 ~d3 h6 IS 0-0 ±, with White ready to continue his assault with b4 and, if ..."'6, lDc3 threatening lDa4. Instead, Miles-Karpov, Plovdiv Echt 1983 continued tamely 13 ~e2?! ~g4 14 0-0 ~xf3 IS ~xf3 d4 Ih,_ l h. Curiously the analysis in ECO, under Karpov's name, does not cite the above game, and gives 13 lDbS as merely ·'unclear". Presumably it would not have been a wise career-move for a member of Karpov's analytical team to suggest that the great man could err. 13lDxdS :txd5 . 13 ...exdS 14 ~d3 h6 IS 0-0 ~f6 (IS ...~g4?! 16 lDes ;!; Browne) 16 'iIIb3 ~g4 17 :edl (17 'iIIxb1? ~xf3 18 b4 'iIIxa3 19 'iIIxc6 .xb4 -+ Browne) 17 ...:td7 and now, rather than 18 h3 ~xf3! (18 ... ~e6 19 'iIIbS;!; PortischBalashov, Toluca IZ 1982) 19 gxf3 d4 20 'ifb5 'iIIxbS 21 ~xb5 = BrowneKarpov, Tilburg 1982, 18 ~e2!? gives chances for a slight edge. 14 ~d3 (D) 14 ~c4 :txd2 15 .xd2 b6 with equality, Ree-Beliavsky, Plovdiv Echt 1983. 14...e5! Black must be prepared to abandon his h-pawn if he is to stay in the game. 14 ... h6?! is too slow: 15 0-0 ci>h8 (IS ...e5 16 ~h7+ ci>h8 17 :txdS .xd5 18 ~e4 'iIIe619 ~g3 ± Farago; White has plans of building on the light

26

The Queen's Gambit Declined: 5 1./41

squares with :dl, etc.) 16 .i.e4 :xd2 17 lDxd2 .i.d7 18 lDc4 ± I.FaragoAbramovic, Belgrade 1982. White has a comfortable advantage in space. IS .i.g3 It turns out that 15 .i.xh7+?! c;j;lh8 does not work: 16 .i.g3 (16 .i.g5?? :xd2) 16...g6 17 .i.xg6 fxg6 18 'ii'xg6 :xd2 19lDxd2 .i.d7, 00 according to Farago, + according to Karpov. White has too little to attack with. If 15 .i.g5, then 15 ... e4! (15 ...h6?! ;1;) 16 .i.xe4 :C5 (but not 16...:xg5? 17 lDxg5 'ii'xg5 18 .i.xc6 bxc6 19 "xc6 :b8 20 :d8+ .i.xd8 21 "e8# Polovodin and Fedorov) 17 "bl :xg5 18 lDxg5.i.xg5 19 .i.xh7+ c;j;lh8 20.i.e4 .i.e6 and Black's active minor pieces perhaps slightly outweigh the slight material deficit, Y.Zilbershtein-Polovodin, USSR 1983. IS•••.i.g4! It is best to repeat the pawn offer. On 15 ... h6, 16 .i.c4?! :xd2 17lDxd2 .i.d7 180-0 occurred in Yuneev-Polovodin, USSR 1983, when 18 ...:c8 = (ECO) improves on the game continuation 18....i.e8 19.i.a2;1;. However, White offered the exchange of rooks too early; 16 O-O! .i.e6 17 .i.c4 l:xd2 18 lDxd2 ;!;, Once Black has played

... e5, the exchange of light-squared bishops will favour White. 16.i.xh7+ 16 O-O? faces tactical problems after 16...:xd3!: 17 "iWxd3 e4 18 "iWb3 (18 "xe4 .i.xf3) 18...exf3 19 "xb7 "xd2 20 "xa8+ tt)d8 21 gxf3 .i.h3 + (ECO), or 17 :xd3 e4 18 :d4lDxd4 19 lDxd4 Wd5 :j: Agzamov-Ubilava, USSR Ch (First League) 1982. 16•••c;j;lf8! The obvious 16... c;j;lh8 is also playable, but this move, bringing the king closer to the centre for the endgame, might just give Black the edge. Mestel-Vaganian, Hastings 198213 continued 17 .i.e4 :d6 18 "c3 "xc3 19 bxc3 :xd2 20 c;j;lxd2 f6 21 a4lDa5 22 lDh4 :d8+ 23 ~c2 .i.c5 24 .i.f5 .i.h5, when Black was comfortably equal despite being a pawn down, in view of White's weak queenside pawns and offside minor pieces. 17 .i.e4 :d6 18 Not an exchange that White particularly wants to offer, but 18 O-O?? loses a piece to 18 ... .i.xf3 19 :xd6 .i.xe4, while 18.i.xc6?:xc619"h7? :Lcl+ 20 c;j;le2 "b5+ 21 :d3 "xb2+ 22 :d2 "b5+ 23 ltd3 :c2+ gives Black a mating attack. 18.•••xc3 19 bxc3 :xd2 20 c;j;lxd2 20 lDxd2!? Polovodin and Fedorov. 20•••f6 As in the Mestel-Vaganian game, but with Black's king two squares closer to the centre. I.Kalinsky-Polovodin, USSR 1983 continued 21 :bl lDa5 22 a4 :d8+ 23 c;j;lel b624lDh4 :c8=i=. After 11. ..:d8 Black is comfortably holding the balance, ifhe plays actively.

.c3

The Old Main Line

27

85) 1l•.•~! (D)

Even better! This thematic pawn sacrifice aims to replace the awkward pin on the c3-knight with an even more awkward pin on the d2-rook. Black hopes to build up a major lead in development. 12lDxe4 dxe4 13 Wxe4 l:d8 (D) The gambit was originally played in conjunction with B ... fS, and then 14 Wc2 e5 15 .i.g5 .i.xgS 16 ~xg5 h6 17 ffi (17 ~h3 .i.e6 + Nei) 17...e4 (l7...l:d8? 18 c5! ~h8 19 WcI! .i.e620 b4 ± Browne-0gaard, Gj(6vik 1983) 18 ~d4 ~xd4 19 exd4 f4 20 Wxe4 (20 Wc3?! Wg5 + Nei) 20 ....i.g4 21 Wd5+ 'W'xd5 22 cxd5 l:fe8+ 23 .i.e2 f3, liquidating into a drawn endgame, Bareev-Goldin, USSR 1983. The text gives White even more opportunities to go wrong! 14c5 An untested suggestion by Miles, the point of which is to clear the c4square. Alternatively: 1) 14 Wc2? e5 15 .i.g3 e4 16 'W'xe4 .i.f5 -+ 17 Wf4 l:xd2 18 ~xd2 l:d8 19 e4 .i.g4! (threatening ....i.g5 and ....i.b4) 20 c5 ~b4 21 f3 g5 22 axb4

(22 Wxg4 lLlc2+) 22 ...Wal+ 23 ~f2 gxf4 24 .i.xf4 .i.e6 25 .i.e5 'W'cl 26 .i.c3 .i.g5 0-1 Miles-Beliavsky, Wijk aan Zee 1984. A devastating example of the disasters that may occasionally befall White if he allows his king to get stuck in the centre in the .i.f4 system. 2) 14g4g5!15b4Wxa316l:xd8+ ~xd8 17 ~xg5 'W'xb4+ 18 ~dl 'W'b3+ 19 ~el (19 Wc2 'W'xc2+ 20 ~xc2 e5 wins a piece) 19.....c3+ 20 ~dl Wa1+ 21 ~d2 "b2+ 22 ~dl, Agzamov-Alzate, Bogota 1984, and now 22....i.d7 (ECO) wins. More problems for the white king. 3) 14 .i.e2 and now, rather than 14... e5?! 15 b4! (better than 15 .i.g3 f5 16 Wc2 e4 17 lLld4 ~xd4 18 exd4 .i.g5 '; De Roode-Van der Sterren, AmsterdamlArnhem 1983) 15 .....xa3 16 l:xd8+ ~xd8 (l6....i.xd8 170-0 g6 18 ~xe5 .i.f5 19 'W'f3 "xb4 20 ~xc6 bxc6 21 'W'xc6 +- Baka1ar-Harasta, Czechoslovakia 1989) 17 0-0 "xb4 18 lLlxe5 ;t Meduna-Velikov, Trnava 1983. ECO suggests as an improvement 14...l:xd2! 15 ~xd2 e5 16.i.g3 .i.e6 17 Wc2 l:d8 ';; White will have difficulty completing his development.

28

The Queen's Gambit Declined: 5 1./41

14•••:xd2! 14...1.xcS? IS b4 +-; 14...fS IS 1i'c4 ;t Miles. 14...WxcS IS 1.d3 g6 (1S ... fS 16"c4;tMiles) 16Wc4(I60-0?eS=F ECO) and White forces the exchange of queens, but with no real sign of an advantage. 15 tQxd2 eS16 .td3!? 16 .txeS iDxe5 17 Wxe5 .te6 18 Wc3 Wxc3 19 bxc3 1.xcS 20 a4 1.e7 =F ECO.

16...g617 .th6 .tfS1S "c4 .txd3 19 "xd3 :dS 20 "c2 l:dS ;. After I1...iDe4, the onus is on White to equalize. It is difficult to see any prospects for a white revival in the 11 l:d2 line. If in one of the main lines (l1...l:d8) Black is comfortably equal, and in the other (l1...iDe4) Black is aiming for an advantage, it is unlikely that White will find the necessary improvement in both lines. Ultimately, II l:d2, though interesting, is a little too artificial.

C) lliDd2: alternatives to ll ...eS 11...eS is the natural reply to the knight retreat, but it radically destabilizes the centre; complications follow. Black's only realistic alternative is to play ... .td7 and aim for counterplay along the c-file. ll...a6 is possible, but if ...1.d7 is played next move, it is merely a transposition (thus, for instance, the game Petrosian-Spassky below went 11 ... a6 12 .te2 .td7 rather than 11.. ..td7 frrst). 1l•••.td7 (D) 121.e2 The pawn is not really worth winning. After 12 cxdS exdS 13 iDb3 Wb6, 14 .tgS .te6 IS 1.xf6 .txf6 16

iDxdS .txdS 17 l:xdS l:ac8 18 l:cS iDe7 19 .td3 l:xcS 20 iDxcs Wa5+ gave Black good play in Dreev-Kuporosov, Tallinn 1986, while if 14iDxdS, Geller gives the entertaining drawing line 14...iDxdS IS lbdSiDb4 16 axb4 .ta4 17 .td3 Wxb4+ 18 ~e2 J.xb3 19 1.xh7+ ~h8 20 l:h5 .txc2 21 J.xc2+ ~g8 22 J.h7+. 12.•.l:fcS So that the queen can go back to d8 without breaking the coordination of the rooks. 12 ...l:ac8, suggested by Geller, might also be worth a try. After 12 ... a6!? 13 0-0 l:ac8, 14 .tg3?! bS IS cxbS?! axbS 16 .txbS iDb4 17 Wb3 1.xbS 18 axb4 Wb6 19 :rei 1.d3 11l-1/2 was Petrosian-Spassky, Buenos Aires 1979. After 20 iDxdS iDxdS 21 Wxd3iDxb4 Black's active pieces amply compensate for White's extra isolated pawn. ECO suggests 14 1.f3 as an improvement for White, citing 14... eS IS 1.gS ;t, but once Black has played ...1.d7, it seems a little illogical to revert so suddenly to the ... eS plan, with White having had time to complete his development. I4... bS!? is more natural, e.g. IS cxdS exdS 16 iDb3 (16 1.xdS b4!?) 16...Wb6 17 iDxdS iDxdS 18 1.xdS iDb4 19 We4

The Old Main Line lDxdS 20 'iFxdS IIfd8 (20 ...J.e6?! 21 'iFd4!) with ample compensation for the pawn. 13 0-0 "d8 14 cxd5lDxd5! After 14...exdS?! IS lDf3 h6 16 lDeS ;!; White had the standard good play against the isolani in KarpovSpassky, Montreal 1979. Despite the general principle that the player opposing the isolated queen's pawn will tend to want to simplify, in most cases in the J.f4 system Black will want to exchange the f6-knight for the c3-knight before accepting the isolani. There are two main reasons for this. First, an inflexible defensive piece is being exchanged for a flexible attacking piece, which makes it easier for Black to create active play. Secondly, the c3-knight provides important cover for White on both the el-aS diagonal and the c-file, and the removal of this knight adds to Black's ability to create threats. These themes will be further illustrated when the 'Exchange Variation' is discussed in Chapter 4. 15lDc4 Evidently deciding there is nothing in 15 lDxdS exdS, when the queen has to take cover. 16 'iFb3 J.e6 17 'ii'xb7?? lDaS would be like taking cover in a gorse bush. 15.....e8 16 J.g3 b5 17lDd6 J.xd6 18 J.xd6lDaS 19 l:[d4lDxc3 20 bxc3 lDc4 21 J.xc4 l:[xc4 In this opposite-colour bishops position, Black's grip on the light squares is more convincing than White's on the dark, and Black duly won in TaborovLerner, Yalta 1982. The general impression is that the ... J.d7 line is under-researched, and

29

provides interesting possibilities for Black. But we lb.ust examine the main lines to see whether Black has any need to avoid them.

2.4 11 ltld2 e5: Introduction 1 d4 d5 2 c4 e6 3lDc3lDf6 4lDf3 J.e7 5 J.f4 0-0 6 e3 c5 7 dxc5 J.xc5 8 "c2 lDc6 9 a3 "as 10 l:[dl J.e7 UlDd2 e5(D)

w

This is Black's most direct method of dealing with the Old Main Line, and by the mid-1980s, with most of the recent theoretical improvements running in Black's favour, was regarded as fully satisfactory. Indeed, as we noted in the introductory chapter, there was even a lull in the popUlarity of the J.f4 system for a few years, until interest was revived by the 'New Main Line' (10 0-0-0) in 1988. 11.. .eS gains a tempo, but ensures that the d-pawn is totally unsustainable on dS. Since ...dxc4 is usually positionally inadvisable, especially when Black has already weakened the light squares with ...eS, Black will want to

30

The Queen's Gambit Declined: 5 1./41

play ...d4 at the fIrst opportunity. White can then exchange in order to isolate the d-pawn, but this opens up the e-flle at a time when White's kingside development is severely lacking. Indeed the general theme of battle in these variations is that White will try to establish control of the light squares, and maybe try to surround an isolated d-pawn, while Black will try to make something of his lead in development. From the diagram, White can play 12 ~b3, 12 i.g5 or 12 i.g3, with the knight move being able to transpose into either of the other lines. To take into account these transpositions, we consider: 2.512 i.g5 d4 13 ~b3 Wd8 (generally regarded as the main line); 2.6 12 i.g5 d4 13 ~b3 Wb6 (or 12 ~b3 'ifb613 i.g5 d4); 2.7 12 ~b3 Wb6 13 i.g5 i.g4 (White avoids 2.5 and Black avoids 2.6); 2.S 12 i.g3 d4 13 ~b3 "ifb6 (or 12 ~b3 Wb6 13 i.g3 d4).

2.5 12 1.g5 d4 13 lLlb3 .dB

.c2

1 d4 d5 2 c4 e6 3 ~c3 ~f6 4 ~f3 i.e7 5 i.f4 0-0 6 e3 c5 7 dxc5 i.xc5 S ~c6 9 a3 .a5 10 lIdl i.e7 11 ~2 eS 12 i.gS d4 13 ~b3 .dS (D) Black avoids the simplifIcation (and hence danger of a slight but steady edge to White) that follows 13 ...Wb6 14 i.xf6 i.xf6 15 ~d5. Instead, he allows his d-pawn to be pinned. The popUlarity of this variation for Black increased when Geller, who has uncovered several new ideas for Black in

the i.f4 system, demonstrated that 14 i.e2 as! was a good equalizer. Before discussing 14 i.e2, we consider a few alternatives. Our lines are: A: 14 i.xf6 30 B: 14 i.d3?! 30 C: 14exd4 31 D: 14 i.e2 33 A)

14 i.xf6 i.xf6 is strictly a minor alternative. The obvious 15 ~5leaves White a tempo down on 2.6 (12 i.g5 d4 13 ~b3 Wb6 14 i.xf6 i.xf6 15 ~5), while 15 i.d3 g6 16 ~5 i.g7, Fta~nik-Rosandic, Vinkovci 1995, should by the same analogy hold no terrors for Black. B)

14 i.d3?! was tried in KaidanovMurugan, Calcutta 1988, and after 14... g61! 15 exd4 exd4 16 ~e2 ~g4 17 i.xe7 "ilxe7 180-0 "ile5 19 ~g3 White stood well. As so often though, Black does best to abandon his h-pawn and get on with his counterplay. Thus 14...~g4! 15 i.xh7+ (15 i.xe7 .xe7 16 ~5 .h4;) 15 .. /~)h8 16 h4! (the only move; 16 i.xe7 .xe7 17 ~d5 "ilM +) 16... g6 17 i.xe7 (17 i.xg6

The Old Main Line fxg61S Wxg6 WeS 19 WxeS :xeS 20 lLldS :bS +) 17...lLlxe7! IS hS ..tg7! and Black will eventually eat up the bishop (not, however, IS ... Wxh1? 19 hxg6+ ..tg7 20 l:lh7+ WgS 21 g7). C)

14 exd4 (D)

Now Black may capture either way, although he may well be obliged at some stage to sacrifice the isolated pawn for active piece-play: Cl: 14•••exd4 31 C2: 14...lLlxd4 32

31

game. A simpler try would have been 16... lLlxd4 17 .txd4 (17 :xd4?! .tcS +) 17 ... .txa3+ IS .te2 .te7, but after 190-0 .te6 20 lLldS! White's lead in development is cause for concern for Black. One light-hearted tactical possibility is 20... lLlxdS 21 cxdS .txdS?! 22 .txg7 .txg2? 23 .txfS .txfl 24 .txe7 .txe2 25 Wxe2 and White safeguards his extra piece. This is worth savouring, as it is one of the few unexplored possibilities in the Old Main Line that offers White the chance to make headway. However, it must be remembered that Black can also vary with 14...lLlxd4 (Line C2 below). 2) IslLlbS is another untried suggestion in ECO, where Is ... lLlg4 16 i.f4 .tgS 17 .tg3 .tf6 "" is cited, but by analogy with the 14...lLlxd4 IslLlxd4 exd4 16 lLlbS .tg4 line (C2 below), Black might want to try IS ....tg4!? Is...lLlg4 16 i.xe7 'fIxe7 (D)

C1)

14..•exd41S .te2 I) IS .te3!? is given without analysis as an untried suggestion in ECO. Then the obvious IS ... lLlg4 16 .txd4 lLlxd4 17 :xd4 gives White a sufficient grip on dS to allow him to count on an edge. IS ... WeS!? may improve for Black, but probably does not equalize. After 16lLlxd4 the post-ECO game Bewersdorff-A.Arnason, Reykjavik 1990 continued 16... lLlg4?! 17lLlxc6 bxc6 IS .te2lLlxe3 19 fxe3 .th4+ 20 ~d2! .tgS 21 Wcl and after castling by hand White had a clearly good

This position can also be reached after 14 .te2lLlg4 IS .txe7 Wxe7 16 exd4 exd4!? (Line D below). In this particular move-order, 16... Wh4 is usually preferred, but the simple pawn recapture is also fully satisfactory. 17lLlds

32

The Queen's Gambit Declined: 5 j"f4!

17 itlxd4 could provoke Black into an unsound piece sacrifice: 1) 17 ...itlxd4?! 18 :xd4itlxt2 19 ~xf2! (not, however, 19 O-O??, as recommended by Azmaiparashvili in Informator 33, in view of 19... We3, threatening the standard smothered mate with ... lLlh3++, ... Wgl+, etc.) 19...Wf6+ 20 ~e3 1.f5 21 Wd2 :ad8 22itld5 :fe8+ 23 ~t2 :xd5 24 I[xdS 1.d3+ 25 1.f3 :e2+ 26 Wxe2 1.xe2 and now the killer zwischenzug 27 .:tel! left White a clear piece up in Cu.Hansen-Ahlander, Malmo 1994. 2) The correct reply is the thematic 17 ...Wh4 18 g3 Wh3: 2a) 19 We4? leaves White's king in the crossfire after 19 ...1.d7 20 lLlf3 Wg2 21 :n :ae8 22 'ii'f4 lLlce5 23 lLlh4 itlxh2! (very pretty; 24 itlxg2 lLlef3+ 25 Wxf3lLlxf3# leads to a less stereotypical smothered mate) 24 I[xd7 Wh3 25 :d4itlg6 26itlxg6 Wxfl + 27 ~d2 :xe2+ 28 ~d3 fxg6 0-1 DreevKruppa, Fronze 1988. 2b) 19 lLlxc6 bxc6 20 'ii'e4 and now, rather than 20...:b8 21 :d21.e6 (2l...itlf6 22 Wf4!) 22 nitlf6 23 Wd3 :fe8 24 ~t2 ± Sher-Ahlander, Farum 1993, Black can play more resolutely: 20...itlf6! 21 Wf4 :e8 and the initiative is with Black after, for example, 22 ~d2 i.f5 23 ~c IlLle4!. 17••:iVeS! This time the direct attack is less convincing: 17...Wh4?! 18 g3 Wh3 19 itlf4! (19 'ii'e4 1.d7 20 ~d2 GO Petraitis-Tomson, corr. 1984) 19 ...'ii'h6 20 lLlxd4itlxd4 21 :xd4 :e8 22 h4! and Black had nothing to show for the pawn in the game Dreev-Gelfand, Uzhgorod 1987. 18(4

18 h3 J.fS 19 'it'd2 d3 20 hxg4 dxe2 21 :cl (21 Wxe2 'ii'xe2+ 22 ~xe2 1.c2 =F) 2l...J.xg4 =F. 18•.•1.(5 18...itle3 19 fxeS itlxc2+ 20 ~t2 lLle3 21 lLlxe3 dxe3+ 22 ~xe3 itlxeS 23lLld4left White nicely placed for the endgame in Forintos-Radulov, Sofia 1981. The text-move is more logical in that Black is developing a piece while attacking the queen. 19 1Wd2 1We4 20 0-0 itle3 21itlxe3 dxe3 22 'ii'c3 Assessed as equal in ECa, but Black was looking good in Dreev-Gelfand, Riga 1987 after 22 ...Wc2 23 :del Wxc3 24 bxc3 :fe8 25 : n :ad8 26 1.n 1.c2 27 lLld4 itlxd4 28 cxd4 :xd4 29 :fxe3 :xe3 30 :xe3 ~f8 31 g3 i.e4. In this endgame, both Black's pieces are more actively placed than their counterparts, and White's queenside pawn structure is broken. Black duly won some 40 moves later. Unless White can make something out of IS i.e3, one gets the impression that it is White who is struggling to equalize. C2) 14••• lLlxd4 15 itlxd4 exd4 (D)

The Old Main Line 16~bS

White must play for material gain; it is too late to play quietly with 16 .i.e2? Black won nicely in J.ScbroerDolgitser, USA 1984, after 16...•a5 17 .i.xf6.i.xf6 18 b4 "e5 19 li:)d5 .i.f5 20 "b2 (20 ~xf6+ gxf6!) 20 ... l:[fe8 21 f4 "xd5! 22 cxd5 d3 23 "b3 llxe2+ 24 ~f1l:[ae8 25 .c4 .i.h4. ECO gives another promising queen sacrifice after 16 .i.e3?! .i.c5 17 .i.e2 dxe3! 18l:[xd8 exf2+ 19 ~fll:[xd8 20 ll)dl g6 21 fuf2 .i.f5 22 "cl bS! +. 16••.i.g4 Likewise, it is a little too late for Black to try retaining the d4-pawn. 16 ... .i.c5?! 17 b4 "e7+ (17 ...a6 18 li:)xd4 .i.xd4 19 .i.e3 ± Cu.Hansen) and now 18 .i.e2 d3 19 "xd3 .i.xf2+ 20 'it'fl WeS 21 .i.xf6 "xf6 22.f3;t Cu.Hansen-Thorsteins, Kiljavajr Wch 1984; instead 18 'ife2?! is tempting, but Black has the resource 18 ...l:[e819 i.xf6? (hoping for 19... gxf6 20 ll)c7) 19.....d7! 20.i.e5 f6. Ivanchuk-Vaganian, USSR Ch 1988 was agreed drawn after 16... .i.d7 17 li:)xd4 "a5+ 18 "d2 "b6 19 i.e3, but Ivanchuk analyses further with 19 ... .i.a4 20 b3, and then 20....i.xb3 21 l:[blli:)e4 22 :xb3li:)xd2 23l:txb6 li:)xfl 24 :xb7li:)xe3 25 fxe3 ±. This does not exhaust the possibilities, and there are many entertaining variations, the most theoretically relevant perhaps being 20... li:)e4! 21li:)f5 (21 "c2 Wa5+; 21 li:)b5 li:)xd2) 21.. ..i.c5 ! (21 .....xb3 22 ~xe7+ ~h8 23 'ifcl ±), when my initial view was that Black stood well. However, Burgess then suggested 22 'ifb2! f6 (22.....a5+? 23 l:[d2! ±; 23 b4?! .i.xb4+ 24 1fxb4 "xfS 25 "xa4li:)c3 00) 23 .i.d3l:[ae8

33

24 .i.xe4 l:[xe4 25 0-0 .i.xe3 26 fxe3 "xb3 (26 ... l:[xc4? 27 l:[bl) 27 "xb3 .i.xb3 28l:[d7, with one possible line being 28 ....i.xc4 29 l:[xg7+ ~h8 30 l:[f3! (30 l:tcl?! .i.e6) 30...l:[e5 with an unclear endgame; if 31 l:[h3, then 3l...i.g8. One gets the feeling that, if anything, it was Vaganian with Black, and not Ivanchuk, who agreed the draw prematurely. Certainly 16... i.d7 deserves serious consideration as an alternative to 16....i.g4. 17l:txd4 17 f3 encounters the same reply. 17•••l:[eS! Gaining an important tempo with the threat of ....i.b4#. IS.i.e3.b6 Another queen sacrifice; 19l:txg4?? li:)xg4 20 .i.xb6 i.b4+ 21 ~dll:[el#. 19 .td3 .tcS 20 0-0 .txd4 20... a6? 21l:[d6. 21 .txd4 .dS 22 .i.xf6 gxf6 23 h3 .thS 24 .txh7+ ~g7 with an unclear position perhaps slightly favouring Black, Piket-Van der Sterren, Lyons Z 1990.

0) 14 .te2 (D)

34

The Queen's Gambit Declined: 5 i,.f4!

14••.aS This move of Geller's is currently the most popular, and is backed up by a number of tactical resources. Black aims to dissolve White's pressure against the black centre with a timely ... a4. Of the alternatives, the old 14...lOg4 is satisfactory, but others offer White an edge. 1) 14 ... lOg4 IS i.xe7 'iIIxe7 16 exd4, and now 16...exd4!? transposes into lines considered under 14 exd4 exd4 (Line C 1 above). 16...'iIIh4 is safe, but is liable to fizzle out into a draw. There follows 17 g3 (White hasn't so far risked 17 i.xg4 i.xg4 18 l:[d2 exd4 19lOxd4, and indeed 19 ... l:ae8+ 20 ~n {20 lOde2 lIe6 gives a strong initiative} 20 ...'ille7 is strong, for example 21 'iIIc 1 lOxd4 22 lIxd4 i.e2+ 23 ~gl i.xc4) 17 ...'illh3 18 dSiOd4! (18 ...Wg2? 19l1niOd420fud4exd4 21 'iIIe4 ±) 19 li.)xd4 exd4 20 :lxd4 :est (20..."g2? loses control: 21 'iIIe4! 'iIIxf2+ 22 d2lOf6 23 'iIIe3 'iIIg2 24 'iIIgl 'iIIh3 2S 1Ih4 'iIId7 26 'iIId4 lIe8 27 i.d3 b6 28 l:tn 1-0 Petrosian-Filip, Cura~ao Ct 1962) 21 l:te4 i.d7 and now various games have subsided to repetition here, for example PortischSpassky, Havana 1966, 22 i.n 'iIIhS 23 i.e2 'iIIh3, etc., or Ree-Pfleger, Lone Pine 1981, with Black playing 22 ...'illh6 instead of 22 ...•hS (probably a slightly safer move-order, as it cuts out the possibility of 23 h3 i.fS 24 hxg4 .txe4). Black's initiative fully compensates for the two pawns, and White is well advised to accept the repetition. The one attempt by White to try for more backfired in Chiburdanidze-Azmaiparashvili, USSR 1982: 22 ~d2? .h6+ 23 f4 lOf2 24 lIxe8+

lIxe8 2S lin .xh2 with a clear advantage for Black. 2) 14 ... h6 IS i.xf6 i.xf6 160-0 as (by analogy with 14... aS variations, but White is fully developed and therefore ready; 16... i.e6 17lOcs .e7 18 lOxe6 .xe6 19 lOdS ;t KorchnoiKarpov, Baguio City Wch (9) 1975) 17 .tf3 i.d7 (17 ... a4 IS1Oc5 i.e7 19 lO3xa4 'iIIaS 20 b4 'iIIa7 21 i.xc6 bxc6 22 exd4 +- Abramovic; this trick of Black's can work in 14 ... aS lines, but here White is too far ahead in tempi) 18 lOc5 i.e8 19 lOd5 i.e7 20 lOd3 dxe3 21 fxe3, and White has achieved his dream of complete control of the central light squares, Hiibner-Short, Brussels 1986. 3) 14... g6 15 exd4 exd4 160-0 and now: 3a) 16... i.f5 gives White two attractive choices: either 17 i.d3 i.e6 (17 ... i.xd3 ;t) 18 lOe2 ;t YusupovKlovans, USSR 1984, or, more imaginatively though not necessarily better, 17.cl d3?! (17 ...•b61SlOb5! threatening c5; 17 ... a6!?) 18 i.xd3! i.xd3 19 lIfel :le8 20 i.xf6 i.xf6 21 lIxe8+ .xeS 22 lOd5 ± Forintos-I.Zaitsev, USSR 1978. 3b) Black may do better with 16... a6!?, keeping ...i.f5 in reserve. 4) 14... i.e6?! makes the mistake of delaying active play until White has finished his development and is ready to react. Goehler-Koglin, Bundesliga wom 1997 continued IS 0-0 lIc8 16 exd4 lOxd4 17 lOxd4 exd4 18 lOb5 .as 19 i.xf6 i.xf6 20 lOxd4 and White had safely picked up the isolated pawn. After 14... a5, play branches, with the most popular choices being IS

The Old Main Line exd4 and 15lDa4, though the two other lines covered here are perfectly acceptable: 01: IS 0-0 35 02: IS .txf6!? 35 03: IS exd4 36 04: IS lDa4 37 15 lDb5 is briefly mentioned at the start ofD4.

D1)

IS 0-0 84 161Dc1 (D)

35

17.td3! In the inaugural game of the 14...a5 variation, I.Farago-Geller, Novi Sad 1979, White mistakenly tried to mix things up with 17 f4?, and Black stood slightly better after 17 .....b6 18lDxa4 "a7 19 f5 .td7 (19 ....tc8! =t Geller) 20 b3 lDg4 21 .txg4 .txg5. Afterwards Geller showed that 17 ...lDg4! 18 .txe7lDxe7 19 .txg4 .txg4 20 :ct3 "b6! would have been winning; the weakness on e3 costs White dearly. Blocking the centre with 17 e4 gives Black an unchallenged edge. Geller gives 17 .....b6! 18 .txf6 (18lDd5?? lDxd5 -+) 18 ...dxc3! 19 .txe7 "xb2 20 "xb2 cxb2 21lDd3lDxe7 22lDxb2 1Dc6 ;. Black has the better bishop, the superior pawn structure, and the d4square.

17•••g6 Not 17...dxc3? 18 .txf6.

18exd4 Correctly releasing the pressure in the centre. This has had rather a poor reputation, since the uninterrupted thrust of the a-pawn has allowed Black to consolidate the position of his d-pawn. However, a recent game by Ivan Sokolov suggests that the position is not the straightforward; that for example ECO assumes. 16•...te6 With the point 17lDxa4?"a5 18 b3 .txa3 =t. 16.....a5 17lDb5 .te6 18 b4 axb3 19lDxb3 "b6 20 exd4 exd4 21lD3xd4 ;!; Bog~- Wlodarczyk, Copenhagen 1991, has also been tried, but the indirect defence is clearly more economical and does not misplace the queen.

18•••lDxd4 19 Wd2 lDhS 20 .txe7 Wxe7 21 We3 f6 22 .te4 White possibly has a slight edge, although the awkward position of the knight on c 1 reduces his prospects somewhat, I.Sokolov-Portisch. Jakarta 1996. Black must avoid 22 ....txc4?? 23 l:lxd4 exd4 24 .td5+. D2)

IS .txf6!? .txf616 0-0 (D) Exchanging on f6 first is a refinement of White's play; the exchange has not only weakened Black's grip on d5, but has also provided the b3knight with a 'retreat' on c5, a considerable improvement on the c I-square.

16•••84

The Queen's Gambit Declined: 5 i.j4!

36

As given by Geller in his analysis of his game with Farago, but Black might well do better to try 16... g6, a thematic move which increases the flexibility of both bishops. 17lLlcS 18lLl3xa4 i.e7 19 b4 "a7 Beliavsky is quoted in ECO as advocating 19 ...•c7 with compensation for the pawn, but this assessment seems rather optimistic. 20 b5 b6! 21 i.f3 bxc5 22 i.xc6 i.b7 keeps Black in the game, but the immediate 20 i.f3 is strong, as 20...i.xc5 can be met with 21 bxc5!, preparing lLlb6. This pawn capture would not be possible if Black had doubled on the a-file. 20 i.f3 dxe3 21 fxe3 i.xc5 22 lLlxc5 "xa3 (D)

"as

w

This position provides an excellent object lesson against taking someone else's analysis on trust. In Informator 28 Geller assesses this position, reached after a long forcing variation, as being much better for Black. And a year later, Hubner, in a critical Candidates' Match game against Korchnoi, played down Geller's analysis. Korchnoi, however, had seen beyond the two weak pawns on b4 and e3, and saw that with correct play he could force a favourable endgame. Thus: 23.d2! 'ifxb4 23 ... lLlxb4 24lLlxb7 is also good for White. 24'ifxb4 24 "'d6 ':a5 25 lLlxb7 i.xb7 26 "'xb4 lLlxb4 27 i.xb7 ':c5 = Quinteros-Chen, Hanover 1983. 24•••lLlxb4 25 lLlxb7 The passed c-pawn, well backed up by White's active pieces, is now a major factor in the position, KorchnoiHubner, Merano Ct (6) 1980. White soon established a winning position, but Black slipped to safety just before the time control. The line with 15 i.xf6 merits further investigation. D3) 15 exd4 a4 (D) This leads to a tactical interlude, resulting in positions where Black seems no worse. 16lLlxa4 16 dxe5? is well answered by 16... lLld7 17 i.xe7 axb3! + - Geller, ECO. Geller's original analysis suggested 16... axb3 17 "'xb3 lLld7, but White can cut his losses with 17 ':xd8 (ECO).

The Old Main Line

37

:xa4 26 :d7 b6 and soon agreed drawn. 2O•••:xaS 21 .i.xf6 .i.b4+ 22 ~n gxf6 23 l:[xd4 Equal, and indeed one of the games on the database was agreed drawn here. White is a pawn up, but he is behind in development and his knight is sidelined. Korchnoi-Karpov, Merano Wch (11) 1981 continued 24 g4 bS 2S cxb5 .i.b7 26 f3 :fe8 27 .i.dl :xbS, with a draw soon following. Korchnoi noted that the exchange sacrifice 26 ':xb4 .i.xhl 27 f3 leads to a more elaborate draw. After 27 ...:fe8 28 tbc3 :e3 29 tbdS :xe2 30 tbxf6+ ~f8 31 tbxe8 :xe8 32 ~f2 :b8 33 b6 :b7 Black cannot escape with his cornered bishop, but White cannot win it without shedding enough pawns to concede the draw. The line with 19 .i.d2!? might well be theoretically critical for Geller's 14... aS.

:eS

16•••tbxd4 Black aims to illustrate the old proverb, "Knights on the edge get stuck in the hedge". Diving into the hedge with 16 ...:xa4? is unwise, as after 17 dS tbaS 18 d6.i.xd6 19 tbcs (Beliavsky) White is about to trim the a-me. 17 tbxd4 exd4 18 b3 18 0-0 "as 19 .i.xf6 .i.xf6 20 b3 .i.d7 is given as equal by Beliavsky. If anything, this is generous to White. 18••:.aS+ 19 'iVd2 Agreeing to the exchange of queens and the return of the pawn. A more ambitious and riskier plan would be 19 .i.d2 .i.fS! 20 "b2. After 20.....eS, Portisch-Beliavsky, Moscow 1981 continued 21 .i.b4?! .i.xb4+ 22 axb4 :fe8 23 :xd4 .i.c2 24 :d2 .i.xb3 2S "xb3 "al+. White could try instead 21 .i.e3!? .i.xa3 22 "xd4 .i.b4+ 23 ~f1 "e7, which Beliavsky assesses as giving compensation for the pawn. However, 24 c5! quickly reduces the appeal of Black's position. 19•••.i.xa3 20 WxaS 200-0 tbe4 21 "xaS :xaS 22 .i.d2 tbxd2 23 :xd2 .i.d7 =Korchnoi, was an assessment confirmed by the game Bagirov-Vaganian, USSR Ch (First League) 1982: 24 :xd4 .i.xa4 2S bxa4

D4)

15 tba4 (D) The most popular move in this position, immediately squelching Black's ambitions with the a-pawn. However, White's knights are moving far enough to the left to cause concern among spin doctors. Another knight move deserves a brief mention: 15 tbb5 a4 16 tiJd2 dxe3 17 .i.xe3 tbg4 18 lLle4 lLlxe3 19 fxe3 "as+ 20 "d2 fS with an equal position, Portisch-H.Olafsson, Reykjavik 1987. After IS tba4, Black's choice is as wide as a move earlier, with IS ... g6 and IS ... h6 being the main possibilities:

38

The Queen's Gambit Declined: 5 /J..f4! 043) 15•••li)g4 (D)

w

041: 042: 043: 044: 045:

15•••.i.g4?! 15....i.d7 15...tOg4 15.••b6 15...g6

38 38 38 39 39

041) 15..•.i.g4?! is perhaps a little primitive. 16 0-0 .i.xe2 (16 ... d3 17 :xd3 .i.f5 18 :xd8 .i.xc2 19 :xa8 :xa8 20 .i.dl ± Yusupov) 17 "xe2 :e8 18 c5 and now, rather than 18 .....c7?! 19 .i.xf6 .i.xf6 20 tOb6 :ad8 21 e4 ± Yusupov-Beliavsky, Montpellier Ct 1985, Yusupov suggests 18 .....d5 as an improvement; then maybe 19 "c2 (instead of the suggested 19 lOcI e4!?, etc.), again with thoughts of .i.xf6 and e4. 042) 15... ~d7 is a sensible development for the bishop, but it is more flexible to insert 15 ... h6 16 .i.h4 first. 16 li)bc5 :b8 (16 ... .i.c8!?; 16... b6?! leaves too many weaknesses: 17li)xd7 fud7 18 .i.xe7 'fIxe7 19.i.f3 ± Ivanchuk-Beliavsky, Moscow 1987) 17 ~xf6 ~xf6 18 li)xd7 "xd7 19 ~f3 li)e7 20 .i.e4 g6 21 c5 ;!;; Yusupov-Ubilava, USSR 1983.

Arguably with less point than a move earlier, in that White's knight is out of range of the d-pawn. 16 ~xe7 'jIxe7 17 b3! After 17 exd4?! "h4, 18 ~xg4?! ~xg4 19 :d2 exd4 =i= was YusupovKorchnoi, Tilburg 1987, while 18 g3 (Korchnoi) is possible, but promises no edge. 17•••li)b6 The piece sacrifice is not quite good enough. 17 ... li)xe3 18 fxe3 Wh4+ 19 Citd2 Wg5 20 ~c1 "xe3+ 21 ci>bl e4 22 .tg4 ;!;; Korchnoi. After 17 ... 'ii'h4, 18 g3?! li)xe3 19 "d3 li)g2+ 20 ~f1 e4 21 Wd2li)e3+ 22 ci>gl "g5 23 :h2 'ii'e5 24 fxe3 d3, as given by Korchnoi, is messy and unclear. He notes that the simpler 18 0-0 li)h6 19 exd4 exd4 20 .td3 is good for an edge. 18exd4 18 ~f3 is probably more accurate . 18...exd4 Korchnoi gives 18 .....g5! co. If 19 "e4 ~f5 20 Wf3 exd4 21 li)xd4? then 21...li)xd4 22 :xd4 'ii'cl+ 23 :dl"c2=t:.

The Old Main Line 190-0"g5 Now: 1) 20 'iWd2 'ii'g6 21 ~hl .tf5 22 lLlxd4i.Llxd4 23 'ii'xd4 .txh3 24 gxh3 'ii'c6+ = Av.Bykhovsky-Kharitonov, USSR 1983. 2) ECO suggests that 20 f4 stabilizes the position in White's favour; if 20.....e7 then 21 .td3 g6 22lLlb6 ±. 044) 15•••h6 (D)

w

16.th4 16 .txf6 .txf6 170-0 "c7 18 e4 b6, Grabowski-Sapis, Poland 1986, is hannless; Black is already taking over the initiative. 16....td7 After 16...i.Llg4 17 .txe7 'fIxe7 18 h3 the black knight lacks the retreat on h6. 17i.Llbc5 Not very convincing as a try for an edge, but neither is anything else: 1) Certainly not 17lLlac5?? a4. 2) Portisch-Tal, Montpellier Ct 1985 went 17 O-O?! lLlb4 18 axb4 .txa4 19 exd4 exd4 20 :d3 (White must aim to unpin the knight; 20 .txf6 .txf6 21 b5 'ii'e7 22 .tf3 "b4 23 :d3 :ac8 =t, or

39

20 bxa5 i.Lld7! aiming for c5 - Tal) 20 ... lLld7 21 .txe7 "xe7 22 :xd4, and now 22 .....xb4! 23 'ii'd2 'ii'xb3 24 .tdl "b4 25 .txa4lLlc5; (Tal) would have been more precise than 22... axb4 23 'fId3 ltfe8 24 .td 1 112-112, as actually played. 3) Tal also analysed 17 exd4lLlxd4 (17 ... exd4!?) 18lLlxd4 exd4190-0(19 ltxd4?? .txa4) 19....txa4 20 'ii'xa4 "b6=. 17....tCS A challenge to White; can he do better than to repeat moves? 17... b6?! is less reliable; 18i.Llxd7 lLlxd7 19 .txe7 'ii'xe7, and now by analogy with the Ivanchuk-Beliavsky game (Line 042 above), 20 .tf3! with a light square bind. Gorelov-Polovodin, Kharkov 1982, unnecessarily interpolated 20 'fIe4 'ii'f6, but after 21 .tf3 :ac8 22 "g4 :fd8 Black was on the point of consolidating. The white queen needs to attack the queenside ('ii'c2-b3 perhaps) rather than the heavily defended kingside. IslLlb3 With a draw by repetition following, Kharitonov-Polovodin, Tallinn 1983. 18 .tg3 lLld7 19 lLlb3 dxe3 20 fxe3 .th4leaves White's pawns too mangled to give any advantage, while on, for example, 180-0, the exchange of various minor pieces after 18 ... i.Lld7 leaves the a4-knight stranded. 045) 15•••16 (D) The most popular choice. Black takes advantage of the displacement of White's knight, and aims for immediate pressure along the h7-bl diagonal. 16.txf6

40

The Queen's Gambit Declined: 5 ~f4! liJd2 .tg5 21 liJf3 .th6 22 .tc4 ;t Thessaloniki OL 1988) is most accurate, exchanging the knight as soon as it reaches d2. Then F~nik-Ubilava, Belgrade 1988 concluded 20 i.c4 l:tac8 21 l:tfe1 ~h8 22liJd2 i.xd2 1/Z-1f2. 17....tg718 0-0 'it'b419 e4.te6 (D) Fta~nik-Kotronias,

w

w So as to loosen Black's grip on the e4-square. On 16 exd4 White suffered a drastic setback in Hiibner-Short, Tilburg 1988: 16....tfS 17"c1 exd4 (17 ... liJxd4 18 liJxd4 exd4 19 0-0 l:tc8 is equal, Gutop-Vitolin~, corr. 1983) 180-0 l:te8 19 :lfel?! (19 .tf3 is answered by 19 ... :lc8 with equality, but not 19 ... .te4 ? 20 .txe4 liJxe4 21 i.xe7 "xe7 22 "f4 ± Bareev-Zotkin, USSR 1982/3) 19 ...:lc8 20 i.e3, and after the thematic queen sacrifice with 20...dxe3 21 ':xd8 exf2+ 22 ~xf2 i.xd8 23 c;tgl liJeS, White, demoralized, resigned prematurely. It is clear though that Black is covering all the important squares, and has the best minor pieces. White cannot really afford a sudden opening of the position when his knights are so off-centre. 16..•i.xf617 cS Gaining some queenside space, and hoping to bring the knight back into the game. 17 0-0 .tg7 18 cS transposes, but 17 cS is slightly more flexible, in that 17 c5 "c7?! 18liJb6 is inadvisable for Black. After 17 0-0 "c7 18 c5 i.e6 19 e4, 19 ... i.g5! (improving on 19 ... liJe7 20

20 liJb6 After the alternative 20 i.d3, Schneider and V.Gurevich give 20....th6 21liJd2 .tg4 22 :ldel i.f4 23 g3 'ii'hs 24 gxf4 i.f3 25liJxf3 "g4+, with a draw. 2O•••l:tad8 21 i.c4 i.h6 Now in 0stenstad-Bonsch, Novi Sad OL 1990, White unwisely exchanged with 22 i.xe6?! fxe6 23 "d3:lf7 24 llJc4 :ldf8, and Black was already gaining pressure on the kingside, while White had lost his grip on dS. Several possible improvements come to mind, for example 22 "d3!?;t.

General conclusion on 12 .tg5 d4 13liJb3 'it'dB Black has little to fear. The pin on the d-file looks awkward, but White's kings ide development takes a long

The Old Main Line time. If Black plays sharply, and builds up counterplay before White can castle, White will not be able to dominate the light squares as he would wish. That having been said, there is still the suspicion that Black does not quite equalize with Geller's 14 .te2 as, the main focus of debate during the 1980s, and older lines involving ...~g4 seem more reliable.

2.6 12 l2Jb3 'ii'b6 13 i.g5 d4 (or 12 .tg5 d4 13l2Jb3 "b6) 1 d4 d5 2 c4 e6 3 M ~r6 4 ~r3 .te7 5 .tr4 0-0 6 e3 c5 7 clxc5 .txc5 8 "c2 ~c6 9 a3 "as 10 :Ldl .te7 11 ~d2 e5 12 ~b3 "b6 13 .tgS d4 (D)

Black's queen is no longer pinned in this line, but is open to a hit with ~dS. The critical question is whether White can establish an edge with the exchange on f6. 14.txr6 Almost invariably chosen. 14 cS?!, by forcing another queen move, 'gains' a tempo, but the net effect is to weaken

41

White's grip on dS. After 14.....dS IS .tbS ~g4 16 .txe7 "xe7 17 exd4 exd4+ 18 ~2, 18 ...'ilh419 g3"h3 20 ~f4 :LeS+ 21 ~d2 'iht6 22 ~cl .td7 '±' Doroshkevich-Plisetsky, USSR 1979 is the reference given by ECO, but 23 h3! (instead of23 .tc4?!), looks very unclear. However, 18 .....f6! is extremely strong for Black. 14•••.txf'6 15 ~d5 "d8 16 .td3 16 exd4?! exd4 17 .te2 .te6 ;t; gives Black an easy and harmonious development. Gupta-Lengyel, Dortmund 1988 finished abruptly: 18 ~xf6+ 1t'xf6 19 0-0 :Lac8 20 :Ld2 ~eS 21 ~xd4 .txc4 22 1t'fS :LcdS 23 "c2 .txe2 24 ~xe2 ~f3+ 0-1. 16•.•g6 Or 16...h6: 1) White gets nowhere by following the same plan as against ... g6: 17 exd4 ~xd4 IS ~xd4 exd4 190-0 .te6 20 .th7+ ~h8 21 .te4 .txdS 22 .txdS 112_112 Miles-D.Gurevich, San Francisco 1987. 2) 17 ~xf6+ "xf6 18 exd4 exd4 19 0-0 .te6 20 f4 :Lad8 21 ~S (without a black pawn on g6, the move fS is pointless) 21.....e7 22 ~xe6 1t'xe6 23 :Ldel 1t'f6 24 :Le4 :LfeS with equality, Tolstikh-laworski, Ceske Budejovice 1995. 3) Since 16 ... h6 to some extent abandons the light squares, White does better to focus on these squares with 17 0-0 .te6 18 .th7+ ~hS 19 .te4;t, when I.Farago-Bayer, Luxembourg 1986 continued 19 ...:Lc8 20 ~xf6 Wxf6 21 ~S dxe3 22 fxe3 WgS 23 ~xe6 1t'xe3+ 24 Wf2 1t'xf2+ 2S :Lxf2 fxe6 26 :Lxf8+ :Lxf8 27 b4 :Lf4 28 :LeI ±. 17exd4~d4

42

The Queen's Gambit Declined: 5 1./4!

17...exd4 IS ~xf6+ 'ilxf6 19 0-0 l:[dS (after 19... a5, White should try 20 ~c5!? rather than 20f4?! ~e7 21 'ilf2 :ldS 22 .i.bl .i.e6 23 :lxd4 a4 24 lIxdS+ lIxdS 25 ~a5 l:[d4 26 c5 i.d5 27 f5 'ilg5 +2S f6 lId2 29 fxe7 'ilxg2+ 30 'ilxg2 lIxg2+ 31 ~hl lIg4+ 0-1 Cu.Hansen-Bonsch, Tilburg 1994) 20 i.e4 i.g4 21 f3 i.e6 22 ~5 lIacS 23 ~xb7 d3 112_1/2 Savon-Geller, Tiraspol 1994, although 24 'ilf2 seems strong. There is scope for further testing here, but on the whole Black would probably prefer to exchange his passive knight for White's active knight. 18 ~xd4 exd4 19 ~xf6+ 19 O-O?! i.g7 20 i.e4 i.g4 leads to an equal position, Ruderfer-Zhidkov, USSR 1979. 19.....xf6 20 0-0 (D)

The critical position, in which White tries to play for a slight positional edge. Black's d-pawn will always require protection, while White has a queenside pawn majority, and also the possibility of pushing on the kingside with f4-f5. We have: A: 20•••i.e6 42 B: 2O•••i.d7 42 C: 20...i.g4 43

A) 20•••i.e6 21 f4 21 :lfel lIacS 22 b3 lIfdS =Korchnoi-Karpov, Baguio City Weh (23) 1975. 21 ...i.d7 21...:lacS?! 22 f5 i.xf5 23 i.xf5 gxf5 24 b3! (24 :lxf5 lIxc4) 24 ...lIfdS 25 lIxf5 'ile7 26 1If3 lIc6 27 'ilf2 "xa3 2S "g3+ lIg6 29 "c7 ± Forintos-Doroshkevich, Lipetsk 1965. Allowing f5 with gain of tempo is dangerous, while with the text-move Black hopes to have provoked a significant weakening of the e-file. 22i.e4 22 f5 g5 is watertight. 22•••i.c6 23 i.d5 23 i.xc6 bxc6 24 f5 looks more promising. 23•••lIfe8 24 "d3 i.xd5 25 cxd5 lIac8 Black gets his rook into play just in time to offset the loss of the pawn. Cu.Hansen-Yusupov, Reykjavik 19S5 continued 26 g3 ~f8 27 1112 lIe3 2S 1i'xd4 1i'xd4 29 lIxd4 ~e7 30 d6+ ~d7 31 lIfd2 lIc6 32 :lb4 1Ib6 33 ~f2 lIe6 and soon drawn. This was reached via 20 ... i.g4 21 f3 i.e6 22 f4, so in fact each player had taken an extra move.

B)

2O•••i.d7 (D) 2111d2 Aiming to double rooks. 21 b4 lIacS 22 1i'b2 lIfdS 23 1Id2 112-112 Korehnoi-Schiissler, Haninge 19S5. ECO cites 21 i.e4 i.c6 22 i.d5, and then 22 ... lIadS 23 1i'e4 ±, Forintos-Smederevac, Wijk aan Zee 1970,

The Old Main Line

but 22 ... lIac8, aiming for counterplay on the c-file, looks much more natural, e.g. 23 'ile4 .txdS 24 cxdS lIc4 or 23 Wd3 lIfd8 24 Wxd4 Wxd4 2S :xd4 .txdS 26 cxdS (26 :xdS lIxdS 27 cxdS ~f8 will round up White's extra pawn) 26...:c2 27 :bl :d6 and Black holds. As in the Hansen-Yusupov game, winning the isolated pawn does not mean a win in the double rook endgame when White's extra pawn is isolated and Black's rooks fully activated. 2l•••lIac8 Again, counterplay on the c-file is the most reliable method for Black. Kallai-I.Almasi, Hungary 1993 continued 21.. ..tc6 22 b4 'ii'f4 (threatening ....txg2) 23 h3 :fe8 (23 .. :ii'gS 24 f4 Wg3 2S bS .td7 26:f3 Wel+ 27 h2 ±) 24 lIfdl :eS 2S.tfl l:tgS 26 :xd4 Wf3 27 :g4 ±. 22 b3 .te6 23 :tfdl :fd8 24 .tn :e7 2S cS b6 26 ab6 axb6 27 b4 .f4 28 .b3 lIed7 Beliavsky-San Segundo, Madrid 1995.

=

C) 20•••.tg4 (D) 2llld2 21 f3!? is also possible:

43

1) 21.. ..te6 (maybe best) 22 f4 transposes into the 20....te6 variation (Line A, above). 2) 21.. ..tfS proved a little too experimental in I.Farago-Arlandi, Forli 1992: 22 .txfS gxfS 23 Wd3 :ad8 24 l:tfel b6 2S lIe2 l:td6 26 f4 ±. Having four isolated pawns is not a recipe for success in a positional struggle. 3) After 21.. ..td7, 22 :f2?! :fe8 23 :fd2 :ad8 24 .te4 .tc6 2S .txc6 bxc6 26 cS :e3 27 W'c4 'ii'eS 28 ~f2 d3 ~ was Tukmakov-Balashov, Kislovodsk 1982. The manoeuvre :f2-d2 looks out of place in this line though. The queen belongs on f2, where it attacks the d-pawn, and the rooks should double on the e-fiJe rather than on the d-file, forcing Black either to concede the open file or to go into an awkward queen and minor piece endgame. Thus 22 'ii'f2! ;t. White has also tried 22 b4, when 22...:fe8 23 .te4 .tc6 24 bS, Forintos-Genovese, Catania 1994, keeps an edge for White. 22 ...:ac8! is better. 2l•••lIae8 The most direct, although 21 ...l:fe8 22 .te4 :e7 23 Wd3 (23 h3? .txh3) 23 ...:ae8 24 f3 .tfS 2S .txfS gxfS 26 'ii'xd4 Wxd4+ 27 :xd4 :e2 also

44

The Queen's Gambit Declined: 51;./4!

proved sufficient to hold in GorelovAndrianov, USSR 1984. 22 ~e4 ~e6 23 c5 b6 24 c6 :fd8 25 :el ~cIS 26 ~xd5 :xd5 27 b4 b5 = L.Spassov-Bonsch, Polanica Zdroj 1987. This is another case where the singleton passed pawn is easily blockaded.

Conclusion White can aim for a microscopic edge in the ..."b6line, and probably gets it. Black should be holding the margin of the draw comfortably enough though. He must be prepared in some lines to let the pawn drop, and get a rook on the seventh.

2.7 12 tDb3: Alternatives to 12...'ifb6 13 .i.g5 d4 1 d4 cIS 2 c4 e6 3llk31Of6 4 tbf3 ~e7 5 ~f4 0-0 6 e3 c5 7 dxc5 ~xc5 8 ~c6 9 a3 "as 10 l:ld1 ~e7 11 ~2 e5 12 tbb3 (D) If White plays 12 ~g5 first, then Black is more or less obliged to play 12 ... d4. The trouble for White is that, as we have seen, 13 tbb3 Wd8 is satisfactory for Black. So White can try 12 ~b3, aiming to force the queen to b6, and then attempt to transpose into 2.6 above with 12.....b6 13 ~g5 d4. However, Black is not obliged to play 13 ...d4, and 13 ... ~g4 is a genuine alternative. One point is that if White has already played ~b3, ...dxc4 is a more likely possibility, particularly if Black has already played ...~e6. 12.....b6 12...Wd8? loses a pawn to 13 tbxd5.

"c2

12 .....c7 is rarely played, but deserves consideration: 1) 13~g3dxc414~xc4,andnow not 14... ~b4? 15 We2llk2+ 16 Wxc2 "xc4 17 ~xe5 ± K.Grigorian-Spassky, USSR 1973, nor 14...~g4? 15 f3 ~d7 16 0-0 :ad8 17 tbd5 tbxd5 18 ~xd5 'iVc8 19 e4 with a light-squared bind, Tukmakov-Lputian, Tbilisi 1980, but rather 14... ~d8! 15 tbb5 Wc6 160-0 ~e617 :cI ~xc418 .xc4 Wxc4 19 l:lxc4llk6 =Lputian. 2) 13 ~g5. Now ECO gives 13...d4 (maybe best) 14 ~xf6 ~xf6 15 tbd5, transposing to 2.6 above if Black plays 15 .....d8, but 15 ...•d6!? enters new territory. ECO also gives 13 ...~g4 14 f3 ~e6 15 cxdS ~xd5 16 ~xd5 ~xdS 17 ~xe7 ~xb3 18 .xb3 ;to 13~g5

For 13 ~g3, see 2.8 below. 13••• ~g4 Aiming to provoke a weakening pawn move. 1) 13 ... ~e6?! immediately saw a spectacular but objectively not fully deserved success in Seirawan-Short, Amsterdam 1992: 14 ~a4 'iVa6 and then 15 cxdS Wxa4 16 ~xf6? tbb4! 17 "e4 :ac8 18 axb4 ~xb4+ 19 ~e2 Wxb3 20 i.xe5 :c4 21 l:td4 :xd4 22

The Old Main Line 'fIxd4 .i.xd5 0-1. A crushing exploitation of White's underdeveloped kingside, but Seirawan noted afterwards that White could have turned the tables with 15 lLlac5! i.xc5 16 cxd5 .i.b4+ 17lDd2 i.xd2+ 18 ':xd2 "a5 19 dxe6 :ad8 20 .i.xf6 gxf6 21 exf7+ ~g7 22 .i.c4 ±. Pieces are level again, but White has two extra pawns and is close to completing his development. Also, by analogy with Chekhov-Kengis (discussed under 13 ....i.g4 14 f3 i.e6), 15 cxd5! 'fIxa4 16 dxe6 should be strong. 2) 13 ... dxc4?! 14 i.xc4 .i.g4 15 l:1d2! ;!; :ad8 16 .txf6 .i.xf6 17 lLld5 was a move-order lapse by Black in Naumkin-Slezka, Prague 1989. 14f3 14 :bl 112-112 Portisch-Beliavsky, Tunis IZ 1985, did not test Black's idea'to destruction. 14... d4 is the natural reply. After the text-move, Black can choose between: A: 14.•. dxc4 45 B: 14....te6 46

A) 14...dxc4 15 .i.xc4 (D)

45

15...i.e6!? A recent attempt by Yusupov to sharpen Black's play. He accepts the weakness of the doubled isolated epawn, but in doing so enhances his control of all the key central squares except for e4. White's e-pawn will itself be weak, as if he ever plays e4 he loses control of d4 and foregoes the possibility of occupying the blockading square. The old move is 15 ....i.h5, with the equalizing plan being to challenge the d-file: 1) 16 'fIe2?! was met by 16...:ad8 170-0 :xdl 18 :xdl l:1d8 19 :xd8+ 112-112 in I.Farago-Kurajica, Amsterdam 1976, but even in this line Black can play more sharply: 16... e4! 17 i.xf6 exf3 18 gxf3 i.xf6 19lLld5 'fId8 20 lLlf4 i.h4+ 21 ~f1 1i'g5:j: Naumkin-Klovans, USSR 1985. Then 22:d5 lbe5 23 :gl i.xf3 24 :xg5 .i.xe2+ 25 lLlxe2 i.xg5 26 :xe5 .i.f6 followed by ...i.xb2, a variation given in In/orma-

tor. 2) 160-0 leaves the king less exposed: 16...e4 17 ~hl exf3 18 .txf6 i.xf6 (18 ... fxg2+ 19 1i'xg2 i.xf6 20 :xf6 i.xdl 21 lDxdl .d8 is also slightly better for White) 19lLld5 Wd8 20 gxf3 ;!; I.Farago-K.Grigorian, Erevan 1982. Black must avoid 16... h6?? 17 .i.xf6 1i'xe3+ 18 ~hl.i.xf6 19 g4 i.g6 20 .xg6, but 16...:ac8 may be considered. 3) 16 Wf2 ':fd8 17 0-0 :xdl 18 :xdl :d8 19 :cl :c8 (19 ... lLlg4? 20 fxg4 i.xg5 21 :el! i.xg4 22 lLld5 wins the exchange) 20 i.xf6 i.xf6 21 lLld5 ;!; I.Farago-Eolian, Erevan 1982. We conclude that Black is not fully equal after 15 ....i.h5.

46

The Queen's Gambit Declined: 5 i.,f4!

16.i.xe6 After 16itJa4, 16.....c7? 17 .i.xe6 fxe6 18 .i.xf6! l:txf6 19 0-0 l:th6 20 itJac5 saw White in control in GulkoYusupov, Novi Sad 1990, but 16...itJb4! equalizes after 17 axb4 "xb4+ 18 ltJd2 .i.xc4 19 "xc4 ji'xc4 20 itJxc4 b5, 17 "c3 itJa2 18 "c2 itJb4 or 17 "e2 .i.xc4 18 "xc4 "a6 - Gulko. 16•••fxe617lbd2 As this time 17 .i.xf6? is not possible in view of 17 ...Wxe3+. 17•••lOds 18 iOOI Not 18itJxd5? in view of 18 ...exd5 19 .i.xe7 "xe3+. 18•••"c7 19 itJb5 "d7 20 .i.xe7 "xe7 Black is comfortable, with wellentrenched knights, Kamsky-Short, Tilburg 1991. B)

14....i.e6 (D)

w

This simple retreat is also possible, the tactical point being that 15 .i.xf6 can be met by 15 ...dxc4. 15 itJa4 15 c5?! takes too much pressure off the black centre. Following 15 .....d8, 16 J.d3 d4 left Black already on top in

Yurtaev-Lputian, Tbilisi 1980. Neither does 16 .i.xf6 .i.xf6 17 itJxd5 .i.xd5 18 J.c4 work, in view of 18 ... ltJd4! 19 itJxd4 .i.xc4 -+ Lputian. 16 e4, as suggested in ECO, still looks pleasant for Black after 16... d4. 15.....c7 Cited as the main line in ECO. 1) The only recent Informator reference is 15 .....d8 16 J.xf6 .i.xf6 and then 17 itJac5 d4 18 .i.d3 g6 19 itJxe6 fxe6 20 .i.e4itJe7! =Tukmakov-Klovans, USSR 1981. ECO, however, gives 17 cxd5 .i.xd5 18 ltJc3 .i.xb3 19 "xb3 ;t, a fair enough verdict given that Black is likely to face problems on the light squares. . 2) 15 .....a6?! is interesting but inadvisable, analogous to the SeirawanShort game (13 ... .i.e6 14 itJa4 Wa6), but with White's extra f3 move as likely to be a handicap as an advantage. Chekhov-Kengis, USSR 1981 continued 16 cxd5 "xa4 17 dxe6 l:tac8 (17 ...fxe6 18 "c4 ±) 18 exf7+ ~h8 19 .i.xf6itJb4 (19 ... gxf6 20 "c4itJb4 21 "e6! ±) 20 "xc8 l:txc8 21 .i.xe7 ltJc2+ 22 ~d2 and White will emerge with a good spread for the queen. 16 .i.xr6 dxc4 A vital gain of tempo. 17 .i.xc4 .i.xc4 18 'ii'xc4 .i.xf6 19 0-0 19 itJbc5 l:tfd8 20 l:txd8+ l:txd8 21 itJe4 .i.h4+ 22 g3 .i.e7 23 ~f2 Wd7 24 ltJac3 'ii'h3 and White's kingside pawns are ragged, Piket-Ivanchuk, Monaco 1997. This was a quickplay game, and one should be cautious about drawing conclusions from such games, but the text-move definitely looks more solid. 19....i.g5 20 e4 l:tacH = Analysis by Kengis.

The Old Main Line

2.8 12 ~g3 1 d4 dS 2 c4 e6 3 M ~f6 4 ~f3 J.e' S J.f4 0-0 6 e3 c5 , dxc5 J.xc5 8 Wc2 ~c6 9 83 WaS 10 l:ldl J.e' Ulbd2 eS 12 J.g3 (D)

Lines with J.gS do not promise all that much, mainly because the bishop is exposed to tactical tricks and simplifying manoeuvres. The quieter 12 J.g3 makes it more difficult for Black to exchange the bishop, but also removes the pressure from the f6-knight, and hence lessens the pressure on d5. 12...d4 Still the most effective reply. 12 ...dxc4? is unthematic. After 13 lbxc4 Wc7 14 lbbS Wb8 IS lbd4! lbd7 16 .i.d3, Ivkov-Pfleger, Bamberg 1968, White's pieces were in picture positions. 12... J.g4 involves a queen sacrifice: 13 f3 J.e6 14 ~b3 "d8 IS cxdS J.xdS 16 .i.c4 J.xc4 17 l:lxd8 l:lfxd8 18 ~f2 J.d3 19 "cl ~ Toth-Barbero, Switzerland 1986. 13 lbb3Wb6 13 ...Wd8 gives White a choice: 1) 14 exd4 exd4 IS lbbS J.g4 16 J.e2 (16 f3 ~) 16... J.xe2 17 "xe2

47

.txa3180-0J.b419lb3xd4lbxd420 lbxd4 We7 21 "f3 We4 22 "b3 ~ Glek-Donchenko, USSR 1983. 2) 14 J.e2!? g6?! (14...J.e6!?) IS exd4 lbxd4 16 ~xd4 exd4 17 lbbS J.fS 18 Wd2 ~e4 19 Wxd4 "a5+ 20 b4 .txb4+ 21 axb4 "xb4+ 22 ~f1 l:lad8 23 J.d6 ± Plaskett-Cools, Skien jrWch 1979. After the text-move, we have reached a branching-out point, with White's choices being: A: 14 c5?! 47 48 B: 14lbbS C: 14 exd4 48 14 lba4?! has also been tried, but presumably only as a move-order slip. C.Wagner-Goldgewicht, Cannes 1997 continued 14....tfS IS fub6 J.xc2 16 lbxa8 .txb3 (16 ...l:lxa8! +) 17 lbc7 J.xdl 18 ~xdl dxe3 and Black was slightly better.

A) 14 c5?! (D)

Attempting to drive the queen to an inferior square, but Black unleashes a typical tactical resource. 14•••J.xc5!

48

The Queen's Gambit Declined: 5 ~f4!

Also 14 ....tfS!? IS .td3 .txd3 16 :xd3 Wa6 17 exd4 exd4 IS lDe2 lDe4 =1= Klovans. 15 ~a4 .tb4+ 16 axb4 Wxb4+ 17 ~.tg4

With two pawns, an attack, and a massive lead in development for the piece. IS:al Or: 1) 18 f3 dxe3 19 fxg41Oe4! keeps White in a bind - Ciolac. 2) 18 .te2 .ixe2 19 'lixe2 d3+ 20 ~xd3 "bS+ 21 "c4 .J:lfdS+ 22 'lic3 "as+ 23 ~c2 bS -+ Ciolac. 3) 18 'ikb3 (relatively the best move) IS .....xb3 19 lOxb3 .ixdl 20 ~xdl .J:lfdS 21 ~c1 :acS gives White an extremely nasty endgame to defend, Footner-Dijkstra, corr 1984. IS...:ScS 19 Wb3 Queen sacrifices for Black in the Old Main Line are highly thematic, and 19 e410aS 20 "d3 'ii'xa4 (Ciolac) is another. 19•••:fdS 20 f3 .te6 21 'ira3 dxe3 22 -'xe3 lOd4 23 .td3 e4! 24 fxe4 lOg4 25 Wf41Ob3 -+ Bondoc-Ciolac, Romania 19S4.

are fun to play, but should not be ventured on if there is a simpler and better alternative. Analytically, it is enough to show that Black is not better in this line, and 19 :c1, preparing ~2, is the way to start. 15 c5 WdS 16 exd4 nbS 17 dxeS 'ife8! IS .ixbS Perhaps it is better to allow Black to do his worst on the e-file: IS exf6 .ixf6+ 19 .te2 .te6 20 .id6 .txb3 21 1i'xb3 lOd4 =Klovans. l8•••lObS 19 .te2 19 f4!? Klovans. 19•••lOxg3 20 hxg3 h6 ; Kaidanov-Klovans, Pinsk 1986.

C) B) 14~b5(D)

With obvious designs on the queen. Black must respond vigorously. 14•••a6! A simple, direct and effective move. 14... dxe3 leads to complications after IS cS "a6 161Dc7 (161Oc3 exf2+ 17 .txf2 bS IS .txbS "b7 is unclear, possibly favouring Black) 16 ... exf2+ 17 .txf2 "a4 IS ~aS .ig4 with compensation for the sacrificed material, according to Kaidanov. Such positions

l4exd4 The main line. Black's choice lies between the natural 14... lOxd4 and the less obvious 14....tfS: Cl: l4...~d4 48 C2: l4••..tfS 49 C1) l4•••lOxd4 15 lOxd4 exd4 16 lObS (D) As so often, the d4-pawn appears doomed, but Black can seek compensation through active piece-play.

The Old Main line

B

49

White is a safe pawn up, KaidanovLysenko, USSR 1983) 21 b4! (21l1d2?! .i.a4 22 1Wxa4 .i.xd4 112-112 ArlandiInkiov, Banco 1985) and Black is in trouble. 2b) 18 ...lIad8 19 .i.f2 "a5+ 20 'ifh5 21 "c3 (21.i.e2 might improve) 2l....i.cS 22 .i.e2 lIfe8 23 lId2 "g6 24 ~f1 and Black has compensation for the pawn, H.Griinberg-Lobron, West Germany 1983. 17li)c7 The obvious capture on d4 is weak. 17 lDxd4? :dS 18 lDb3 (IS lDf3? "a5+ favours Black), and now instead of IS ...lIxdl+? 19 ~xdl.i.e6 20 ~cl co YrjOlii-Westerinen, Gausdal 19S5, Black could have kept up the pressure with IS ....i.fS! 19"xfS"xb320.i.d3 "xb2 21 .i.e5 "xa3 22 .i.xf6 g6! 23 "f3 .i.xf6 + YrjOlii. 17•••lIb8! 17...lIa7?! IS .i.e2 .i.d7 19 b4 ±. The text-move is assessed as unclear by Yrjolii, without giving supporting analysis. 18 lDds lDxdS 19 .i.xb8 lDb4! 20 1We4 20 axb4 .i.xb4+ 21 lId2 (21 ~e2? "e6+ -+) 21...:e8+ 22 .i.e2 .i.f5! 23 "xfS .i.xd2+ -+. The position after 20 "e4 is unexplored. One feels that Black ought to have plenty of compensation for the exchange. Maybe 20 ... lDc6!? is the way to proceed.

"d2

16.•.a6! Putting White's undeveloped kingside under great pressure. 1) 16....i.d7 is also good. 17lDxd4 :fe8 18 .i.e2 'ii'a5+ 19 'ii'd2 (19 :d2 lIad8 20 0-0 .i.a4 21 b3 {21 .i.c7 'ii'xc7 22 'ilxa4 'ii'f4 23 lIfdllDe4 -+; 21 "d3lDe4 22 b4 "b6 23 "xe4 .i.g5 +} 21. ...i.c5 22 :fd 1 lDe4 23 lId3 lDxg3 24 hxg3 lIxd4 25 lIxd4 .i.xd4 26 lIxd4 .i.xb3 27 "xb3 lIxe2 = is analysis by Abramovic) 19 .....xd2+ 20 lIxd2lDe4 21 lId3 (21 lid 1? .i.f6 22 .i.f4 .i.a4! 23 lId3 lDc5 24 lId2 .i.xd4 2S lIxd4li::le6, and Black, having ftrst won a tempo with ....i.a4, now wins the exchange, Hai'k-Abramovic, Montpellier 1986) 21.. ..i.f6 22 .i.f4 lDcS 23 lId2 .i.xd4 24 lIxd4 lDe6 2S lIxd7lDxf4 26 lId2lDxg2+ =Abramovic. 2) 16....i.g4 17 f3 .i.d7 is less convincing than in variations where White has played .i.g5; Black lacks the tactical tricks resulting from the exposure of the bishop, and voluntarily relinquishes the possibility of playing ... lDe4. 18lDxd4 and now: 2a) 18 ... .i.c5?! 19 .i.f2 lIfe8+ 20 .i.e2 lIad8 (20 ... a5 21 0-0 lIad8 22 lId2 .i.a4 23 'ii'xa4 .i.xd4 24 "dl and

C2) 14•••.i.fS Aiming to get stuck into White's queenside before White can complete his kingside development. IS.i.d3

50

The Queen's Gambit Declined: 5 i.,f4!

.a2

15 .xf5 .xb3 16 dxe5 .xb2 17 lLlb5 a6 18 :bl 19.i.d3 g6 20.f3 lLld7 21lLlc7lLldxe5 22 .e2 .xa3 23 .txe5 lLlxe5 24 0-0 lLlxd3 25 lLlxa8 :xa8 26 :xb7 and after some fastmoving but relatively straightforward play, Black is perhaps a little better, Arkhangelsky-St.Nikolov, Bulgaria 1986. Black played passively with 26 ...:e8 and later lost, but 26 ... i.c5 27 .f3 lLle5 would leave him well placed. IS•••.i.xd3 16 :xd3 (D)

16.••lLlxd4 16...exd4?! 17lLlb5;t. 17 lLlxd4 exd4 18lLle2 18 lLlb5 .e6+ keeps the white king in the centre, unless he is prepared to surrender his c-pawn. ECO assesses the position after 18 lLle2 as even, while Averkin regards it as slightly advantageous for White. We leave it to the reader to decide, from the last few moves of the game; a microscopic advantage perhaps? 18.••:t'd8 19 0-0 lLle4 20 :fdl (20 i.e5 i.f6) 20•••i.f6 21 b4 as 22 eS "a6 23 i.e7 :de8 24 i.xaS lLlxeS 2S

bxeS 'i'xaS 26 lLlxd4 "xeS Glek-Averkin, USSR 1983.

l/z_l/,.

2.9 Conclusion We have now reached the end of our survey of the Old Main Line with 10 :dl. Time after time it becomes clear that White has nothing or next to nothing, and while there are several sharp variations, these are just as likely to favour Black as White. This chapter is perhaps of most practical use to the player with the black pieces attempting to find a reliable method against 10 :dl. For such a player, the most reliable choices would seem to be: 10 :dl i.e7 and then: 1) 11 :d2 lLle4 12 lLlxe4 dxe4 13 .xe4:d8. 2) I1lLld2 e5 and here: 2a) 12 i.g5 d4 13 lLlb3 .d8 14 i.e2lLlg4. 2b) 12lLlb3 'il'b6 13 .tg5 .i.g4 14 f3 .te6. 2c) 12 i.g3 d4 13 lLlb3 Wb6 14 exd4lLlxd4 15lLlxd4 exd4 16lLlb5 a6. Such a selection is slightly subjective, based at least in part on the author's own style of play, but the good news for Black is that in each of these lines of play, Black has a satisfactory and valid alternative. This means that it is unlikely that any improvement for White in a particular line will challenge the basic soundness of Black's position. It is time to tum to the New Main Line, and to see what Black has got there.

3 The New Main Line 3.1 Introduction 1 d4 cIS 2 c4 e6 3lLlc3lLlc6 4lLlc3 i.e7 5 i.C4 0·0 6 e3 cS 7 dxcS i.xcS 8 "c2 lLlc6 9 a3 10 O-O-O!? (D)

"as

We now move to the cutting edge of theory. Queenside castling was not even mentioned as a possibility in ECO, published in 1987, yet from 1988 onwards it has been very much the main line. Even so, the move is so visually implausible that it can be difficult to convince oneself that White's play is sound. After all, Black has already started to open up lines on the queenside, and lines will be opened up still more upon an exchange of pawns, while as yet White's kingside attack is not even in its infancy. It takes a definite leap in imagination to appreciate that there are several positive points to the move, which outweigh the manifest defects. For example:

• The pin on the knight is broken, and White need no longer worry about his king getting stuck in the centre. These are far from minor considerations, given that in so many otherwise promising lines in the .l:tdl variation, the initiative disintegrates because the white king comes under attack. The king, surprising as it may seem, is safer on c1 than on el. • Associated with this, White need no longer have worries over the development of his king's bishop. The fl-bishop moves out when it is ready to play its part in the game, and not as an emergency measure to speed up castling, or to block the e-file. • Again associated with this, White no longer needs to play the excruciatingly slow lLld2-b3 purely in order to relieve the pressure on the a5-el diagonal. This diversion of the knight in the Old Main Line allows Black time to play ...e5 and ...d4. Sometimes White will still want to play lLld2-b3 to embarrass the black queen, but this will be a matter of choice, an element in an attack, rather than an obligation. These points together suggest that White's development is quicker and more flexible in the new main line than in the old - and efficient development is one of the chief objectives of opening play. If we now consider the positions of the respective kings, we

52

The Queen's Gambit Declined: 5 j.,f4!

may first of all note that Black's king has not too many friendly pieces around it, giving White legitimate hopes of building up an attack. Secondly we may note that White's pressure in the centre makes it particularly difficult for Black to develop the c8-bishop, which makes it more difficult for Black to open up lines effectively on the queenside. So far, general experience has favoured White in this position, and grandmasters of the calibre of Short, Timman and Vaganian have suffered crushing defeats as Black. One would not be surprised if reliable equalizers were found for Black in several lines, and many of the verdicts in this chapter are best regarded as provisional. It must be reassuring for White though that this line has proved its worth for a decade now; it is not one of those lines, such as 10 l:ldl and 11 l:ld2, thilt briefly flits into fashion before being packed off to the lumber room. One may feel reasonably confident that White will be winning grandmaster games from this position in ten years' time, provided of course Black allows the position to occur. The main lines in this chapter are as follows: 3.2 1O... .i.e7 11 h4; 3.3 10... .i.e7 11 g4; 3.4 1O....i.e7 with neither 11 h4 nor 11 g4; 3.5 1O...dxc4 11 .i.xc4 without 1l....i.e7; 3.6 1O... .i.d7; 3.7 1O... a6; 3.8 1O... l:ld8; 3.9 1O... lOe4.

3.2 10....i.e7 11 h4 1 d4 dS 2 c4 e6 3 M lLlr6 4lLlr3 .i.e7 5 .i.r4 0-0 6 e3 cS 7 dxcS .i.xcS 8 'it'd lLlc6 9 a3 .as 10 0-0-0 .i.e7 11 h4 (D)

The older 11 g4 is considered in 3.3, and other moves, such as 111L1d2, 11 ~bl and 11 lLlbS, in 3.4. With so many of the examples so far coming from the 1980s, it is refreshing to jump straight into the latest theory. White naturally intends some sort of kingside pawn-storm, but which comes first, the g-pawn or the h-pawn? 11 g4 is the most obvious move, with the straightforward plan of booting the knight with gS and then taking control in the centre. However, Black can respond with an immediate counterattack, 11...dxc4 12 .i.xc4 eS! 13 gS exf4 14 gxf6 .i.xf6 IS lLlds being a critical line (3.3 below).

11 h4 has the merit of avoiding this possibility. The h-pawn will always be useful in a kingside attack, not least because it will support any piece landing on gS, but 11 h4 does not threaten anything just yet. Neither, though,

The New Main Line does it provide any target for Black to attack, which is important as it makes it difficult to him to develop quickly with the aid of threats and tactics. We consider: A: 1l...dxc4 53 B: 11..•:d8 60 c: 11••.a6 61 D: 11 ••..i.d7?! 63 A)

1l•••dxc4 12 .i.xc4 Now we have: AI: 1l...b6?! 53 A2: 12•••86 54 A3: 12... l:d8 58 An inferior option is 12...e5?! 13 .i.g5 ±, and if 13 ...h6?, 14 .i.xh6 gxh6 15"g6+~hS16"xh6+~h717 :d5 "ikc7 IS ~g5 .i.f5 19 .i.d3, etc. A1) 12•••b6?! (D)

This was Black's first try, in Kasparov-Vaganian, Novgorod 1995. Black didn't last long! 13~gS?!

Although Kasparov won the stem game very nicely, I find it hard to

53

believe in the exclamation mark that he awards this move, especially given that 13 .i.b5! places the queen in mortal danger. Kasparov then gives 13 ....i.b7 14 ~d2 a6 15 ~c4 axb5 16 ~xa5 ~xa5 (l6 ... l:xa5 17 .i.d6 .i.xd6 18 :xd6 b4 19tile4 ~xe4 20 'ii'xe4 bxa3 21 bxa3 :xa3 22 :hdl gives White a clear advantage) 17 .i.d6 .i.xd6 IS :xd6 b4 19 axb4 ~4 20 l:ddl :al+ (20 ....i.e4! is stronger - Burgess) 21 ~bl :cS "with compensation for the material". Yet it would be extraordinary if White, with queen and pawn for two minor pieces, did not have an outright win in all this. 20 :d4! instead of the limp 20 :ddl? must surely bust Black: 1) For example, after 20...:fcS 21 :hdl ~5 (Black must watch out for back-row mates after, for example, 21...:al +) 22 "b3 Black's attack has only limited resources. 2) 20...J:a1+ 21 ~bl b5 (Burgess) with the idea of ...e5 is a tougher defence. Then 22 :hdl e5 23 :dS .i.e4 24 :xf8+ ~xfS 25 :d3 :a2, and now Nunn gives 26 "dl :xb2 27 ~3 ±; White's lack of queenside pawns will make it difficult to force the win. 26 f3! looks even stronger, and if 26 ....i.xd3, 27 'ii'xd3 :xb2 28 'ii'dS+ tiles 29 ~3! and now either 29 ...:xb4 or 29 ...:d2 would be decisively met by 30 ~d5. After 26 ...:xb2 27 fxe4 :xc2+ 2S ~xc2 ~xe4 29 ~c3 ~f2 (29 ... ~ed6 30 ~b3 +-) 30 :d7 ~xe3+ 31 ~b3 and ~xb5 White's passed pawn will easily outpace Black's. Could it be that Kasparov was writing on the principle that "the winner's moves are always correct; doubly so if the wi~ner is Kasparov"?

54

The Queen's Gambit Declined: 5 1£./41

13...~a6 14 IOce4 g6 15 IOxf6+ ~xf616lOe4! ~e7 17 ~xa6 "xa618 ~b1"b7?

Kasparov annotated the phase from moves 16-1S in considerable depth in Informator 63. Since this phase is of no great significance theoretically, given White's improvement on move 13, no attempt is made to repeat or summarize his analysis here. It is worth noting, however, that Kasparov regarded this as the critical mistake, preferring instead IS ...lIacS 19 ~h6, giving 19... lOe5 20 Wb3 as;t. De BoerTondivar, Netherlands 199516 saw instead 19 ... lIfdS 20 ~g5 1Ob4 21 l:xdS+ ~xdS 22 Wd2 i.xg5 23 hxg5 IOd5 24 /Od6 ':'c6 25 IOxf7 ~xf7 26 e4 iOe7 27 lIxh7+ (27 'ti'dS h5 28 gxh6 Wc8!) 27 ...~e8 2S l:thS+ ~f7 29 lIh7+ with a perpetual. The interesting point of this is that presumably both players took Kasparov's analysis at move 13 on trust! 19 hS lIac8 20 hxg6 IOb4 21 gxh7+
A2) 12•••a6 13 lOgS (D) Both sides limber up for their attacks on the opposing king. But should Black just get on with his attack, or should he do something about White's

kingside pressure? We consider three possibilities: A21: 13•••lOe5?! 54 A22: 13••• b5 55 A23: 13...h6 56 None of these moves seem to be adequate, leading to the conclusion that 12... a6 does not solve Black's problems.

A21) 13...lOe5?! Given the honour of the exclamation mark in Informator 67, but unconvincing. 14lOce4? Black's last move is to be regarded with suspicion, as it places the knight on an exposed square in a position susceptible to tactics. White's reply is even more questionable, however, as it exposes his weaknesses on the c-file, without adding anything substantial to the kingside attack. The recommended move is the more discreet 14 i.a2!, threatening /Ods, and if 14...lOg6, 15 ~g3 IOh5 (15 ...i.d7 16 IOxh7) 16 i.h2, when Black's counterplay is fruitless. My first impression was that the forcing 14iOd5?! would have been strong, but after 14... exd5 15 lIxd5

The New Main Line lLlxc4 16 l:xaS lLlxaS 17 .i.eS l:e8! (Burgess and Nunn) the invasion on h7 is relatively harmless. Nunn gives 18 .i.xf6 .i.xf6 19 "xh7+ ~f8 20 lLle4 .i.e6 21 lLlxf6 lLlb3+ with Black giving perpetual check. For what it is worth (3 Elo points to the Scottish GM; perhaps not so much theoretically), after 14 lLlce4?, Motwani-Bartels, Tilburg 1996 finished 14 ... lLlg6? (14 ... g6! IS .i.b3 .i.d7 Motwani) IS hSlLlxe4 16 "xe4 .i.xgS 17 hxg6 h6 (17 ...hxg6) 18 .i.d6 l:d8 19 gxf7+ ~xf7 20 f4 .i.f6 21 l:xh6! gxh6 22 "h7+ .i.g7 23 .i.eS l:g8 24 "fS+ .i.f6 2S .i.xe6+ .i.xe6 26 "xf6+ ~e8 27 "xe6+ 1-0. In both the Motwani-Bartels and the Kasparov-Vaganian games, we see short and snappy wins by White, but only after openings in which both players made mistakes. This is alas typical of what happens in new and complicated opening systems where the theoretical material is lacking, or has not had time to be properly evaluated. The rest of the chapter provides several further instances of play which comes well below meeting the needs of the position, with even leading grandmasters making big mistakes. It is the modest hope of the author that the fmt proper codification of the new system will make it easier for players, at both grandmaster and club levels, to avoid making the elementary mistakes.

55

w

=

A22) 13••. b5 (D) This would be fine if White had to move his bishop. But White is not so obliged.

14lLlce4! g615 .i.b3 Hjartarson suggests IS .i.d6, without offering further analysis. There might follow: 1) IS ...bxc4 16 .i.xe7lLlxe7 (after 16... lLlxe4 17 .i.xf8lLlxgS 18 .i.d6 the knight is trapped) 17 lLlxf6+ ~g7 18 lLlfxh7 l:h8 19 hS with a massive attack. 2) lS ...lLle8 is complicated and critical, with pieces flying allover the place: 2a) In an editorial comment, Nunn suggests 16 .i.xe6!?, and perhaps wisely leaves the author to sort out the resulting chaos. It turns out, however, that 16... lLlxd6 17 lLlxd6 .i.xe6 18 lLlxe6 fxe6 19 "xc6 b4! is, contrary to all notions of natural chess justice, apparently satisfactory for Black: 2al) The main point is that if the queen aims for e6 (20 "c4; 20 "e4; 20 'ii'd7) Black has 20 ... bxa3 21 "xe6+ ~h8, and if 22 "xe7 then 22... axb2+, when perhaps it is wisest for the author to leave the reader the analytical exercise of working out whether Black is winning after 23 ~bl "al + 24 ~c2. Black's attack should be good enough for a draw at least if he plays 24 .. :.a4+.

56

The Queen's Gambit Declined: 5 JLf4!

2a2) If instead White tries 20 1&4, Black has 20...•f5 21 e4 .f4+ 22 ~b1 :tacS 23 'ii'xe6+ .f7 24.xf7+ :txf7 and Black saves the day by a rapid counterattack on b2, for example 25 lbe5 :txf2 or 25 lOb6 :tbS 26lOdS bxa3. 2a3) It is not clear either that 20 a4 b3 makes progress for White. 2b) My original note suggested 16 i.xe7 lOxe7 17 h5 h6 (17 ... bxc4 IS hxg6 is extremely unlikely to be tenable for Black, for example IS ... hxg6 19:th7 i.b720':dhl.e5 21 f4i.xe4 22 fxe5 i.xc2 23 :thS+ ~g7 24 :tlh7#) IS hxg6 hxg5, and now I mistakenly suggested 19 gxf7+? :txf7 20 .e2, when 20...•c7! (Burgess) is good for Black. 19 g7! improves, but is still only good for a draw after 19...~xg7 20 :th7+ ~xh7! 21 lOxg5++ (not 21 lOf6++? ~g7) 21...~h6 22 .h7+ ~xg5 23 f4+ CiPf6 24 .h4+ CiPg7 25 :thllOg6 26 .h6+ ~f6 27 .g5+. 3) 15 ... lOxe4 16 .xe4, and now what? Variation '2' is critical for the soundness of 13 ... b5. On my flfSt draft of the book, I felt that White was much better, and that 13 ... b5 was inferior. It is with some amazement that I conclude that Black may well be OK after IS ...lOeS, though there is scope for further analysis. For example, after Nunn's 16 ..txe6, can White find an improvement on move 20? At the very least, Hjartarson's IS i.b3 is not, as I first assumed, a case of bottling out of the critical lines. IS•••i.b7 16lOxf6+ 16 hS? lOxe4 17lOxe4 :tacS with a strong counterattack - Hjartarson. 16•••i.xf6 17 lOe4

Hjartarson analysed 17 :td7? :tacS IS :txb7lOes +and suggested, without further analysis, 17lOxh7!? ~xh7 IS hS. It is hard to believe that the piece sacrifice is well timed. 17•••..teS! Improving on 17.....te7? IS ':d7 eS (1S ... lOb4 19 i.c7! ± Hjartarson) 19 :txb7 ':acS 20 ..th6 lOb4 21 .xcS ':xcS+ 22 CiPb1 lOc6 23 :tc1, with White having an excellent mixed bag in return for the queen, SchandorffOlesen, Copenhagen 1995. The game finished abruptly: 23 ...b4?? (23 ...•dS ±) 24 .:xc6 ':xc6 2S :txe7 1-0. 18lOeS i.xf4 19 exf4 'fIe7 20 ~bl :tad8 21lOxb7 'fIxb7 22 'fIe4 b4 According to Hjartarson, 23 axb4 .xb4 24 'ilxb4lOxb4 2S i.c4 should be played, with White having fractionally the better endgame. Instead, the game Hjartarson-King, Winterthur 1996 continued 23 a4 .b6? (23 ... lOaS 24 .xb7 lOxb7 Hjartarson) 24 hS with a clear plus for White. But IS ..td6!? is the move that has to be looked at.

=

A23) 13.•• h6 (D)

The New Main Line Black might not be threatening the knight straight away, but it is useful to keep it attacked. This position may also be reached via the move-order IO...dxc4 11 .i.xc4 a6 12lOg5 .i.e7 13 h4 h6, which explains why we have a couple of pre-1990 references in the notes that follow. 14.tal! Untried, but probably very strong. White takes the bishop out of harm's way, and prepares to bring it to bl to intensify the attack on h7. Indeed, White is attempting to make a nonsense of Black's whole plan of action; 11...dxc4 develops White's bishop, then 12 ...a6 aims to push it to a different square, so that Black has used up two tempi to help White with the lengthy but dangerous manoeuvre .tfl-c4-a2-bl. Other moves have been tried here. 1) 14 IOce4!? is good for a slight edge, but not so ambitious. A.SokolovBeliavsky, Erevan OL 1996 continued 14 ... lOxe4 15 Wxe4 Wf5 (following 15...hxg5? 16hxg5, 16...Wf517Wxf5 exf5 18 .:th2!, Ruban, gives White a mating attack down the h-file, while the refutation of 16... g6, given by Sokolov is even prettier: 17 l:r.d5! exd5 18 .te5 IOxe5 19 Wh4) 16 Wxf5 exf5 17 IOf3 .tf6, and now according to Beliavsky, instead of the quiet 18 ~bl .te6 =, 18 .td6 would have kept a slight edge. Black's pawn structure is suspect, and his development lags a bit. 2) 14 g4 is violent, but arguably less logical, in that the pawn is obstructed by the knight. 2a) Ruban-Kotronias, Sochi 1989 saw some brutal hacking after 14 ... b5?!

57

15lOce4 g6 161Oxf6+ .i.xf6 17.td3 (17lOxf7!?, Ruban, is also worth a try: 17 ... ~xf7? 18 h5 yields a winning attack, while 17...:xf7 18 Wxg6+ .tg7 19 .td3 also gives Black some scary moments) 17... b4 18 .txg6! hxg5 19 hxg5 .i.e5, and now White blew it with 20 .txf7+?? :xf7 21 Wg6+ ~f8 -+. Instead 20 .te4! bxa3 21 .txe5 Wxe5 22 .txc6 Wxb2+ 23 Wxb2 axb2+ 24 ~bl (Ruban) would have left White with a good endgame. 2b) 14 ... lOe51eads to complicated play: 2bl) Ruban blew another fuse in Ruban-A.Maric, Palma de Mallorca 1989: 15 :d5? exd5 16 .txe5 .:td8 (16 ... dxc4! 17 .txf6 .tf5 -+, Ruban, would have been even stronger) 17 .ta2 .txg4 +. 2b2) After 15lOd5 exd5, 16.txe5 :d8 is given by Ruban as unclear, while 16 :xd5? IOxc4 17 :xa5lOxa5 18 .i.e5 g6 19 Wc3 .i.xg4 20 Wxa5 l:ac8+ 21 ~b1 (21.tc3 hxg5) 21...:C5 would be distinctly unwise. Compared with the variation to the MotwaniBartels game (13 ... lOe5?! 14IOd5?!, etc., Line A21 above), White has the extra move g4, which attacks nothing, and Black has the extra move ... h6, which attacks the knight; this swings the variation in Black's favour. 2b3) Ruban also analyses the line 15 IOce4!? lOexg4 16 :hgl IOxe4 (l6 ...hxg5?! 17 hxg5 gives Black hfile problems) 17 IOxe4 f5 18 lOd6 with compensation for the pawn. Since White need not go into any of these double-edged lines, it is not necessary to analyse further to overturn these 'unclear' verdicts. 14•.•:d8

58

The Queen's Gambit Declined: 5 J..f4!

A sensible idea, but it would have been better a couple of moves earlier. Other moves lead to disaster: 1) 14 ......f5? 15 "'e2 followed by .tbl would walk straight into White's attack. 2) 14...hxg5? 15 hxg5 'iff5 16 'ife2 is similar. 3) 14... lDe5? 15 ll:xIs works for White this time, as Black no longer has the ... lOxc4 resource after 15 ... lOxdS 16l:txd5. 4) 14....txa3 has no real right to work: 15 bxa3 'ifxa3+ 16 'ifb2 'ifxb2+ 17 ~xb2 hxgS 18 hxgSlOg4 19 l:r.h4 lOxf2 20 .l:.d2 eS 21.l:.xf2 exf4 22 exf4 and Black's kingside will soon buckle under. S) 14....td7?? IS l:r.xd7. IS .l:.xdS+ .txdS White now has several ways of playing for a lasting edge, simplest perhaps being 161Oge4. Thus 12... a6 does not solve Black's problems. A3) 12•••.l:.dS (D)

This is the basic formation which Kharitonov has favoured on several

occasions, against both g4 and h4. One possible drawback is the lack of flight squares for the black queen. We discuss: A31: 13 ~bl S8 A32: 13 lObS!? S9 A31) 13 ~bl old7 14 lOgS 14 g4 transposes to positions usually reached via 11 h4.l:.d8 12 g4 .td7 13 ~bl dxc4 14 .txc4 (Line B, below). After 14 ....l:.ac8 Black has good chances of equalizing. 14 lObS lOeS is comfortable for Black. 14•••h6!? This move leads to massive complications. 14....te8 defends much more simply, and prepares favourable simplifications. Kharitonov then gives IS lOce4lOxe4 16 'ifxe4 'iffS 17 "'xfS exfS as ;1;, but it is difficult to see White making much progress. 151Od5!? Kharitonov gave both IS g4.te8 and IS lOge4lOeS as unclear. 15•••exdS 16 J:.xdS 1i'b6 16...olc5?? 17 .l:.xc5 +-; 16..:iVxdS? 17 .txdS hxgS 18 hxg5 +- Kharitonov. 171Oxf7 (D) 17•••lbxd5 Better than 17 ... ~f8 18 .l:.b5! (in Korotylev-Kharitonov, Moscow 1996, White chickened out with 18 .l:.ddl?, and Black was able to counterattack to great effect with IS ....txa3 1910xdS .l:.xd8 20 olb3 ole7 =+= 21 f3 loaS 22 .ta2 J:.cS 23 "f2? .tfS+ 0-1) Is ...lOd4 (Kharitonov gives the line 18 .....xbS 19 .txbS ct;xf7 as unclear, but 20 .tc4+ seems to favour White) 19 :xb6

The New Main Line

ll)xc2 20 ll)xdS! (in his analysis in Informator, Kharitonov considers only the less incisive 20 lIxb7 .i.f5 21 ll)xdS ll)xe3+ 22 ~a2 ll)xc4 23 ll)c6 .i.e4 24 lhe7 .i.xc6 25 lIe7 .i.d5, when if anyone has the edge, it is Black) 20 ... axb6 (20 ....i.xdS? 21 lIxf6+) 21 ll)e6+ .i.xe6 22 .i.xe6 and Black's extra knight is trapped, leaving White with excellent chances in the endgame, thanks to his strong bishop-pair and passed e-pawn. 18.i.xdS lSll)xh6+ ~f8 19 "h7 .i.f5+! 20 ll)xf5 .tf6 21 "hS+ ~f7 22 "h5+ ~gS 23 .txd5+ lIxd5 24 ll)h6+ ~h7 (24 ... gxh6 00) 25ll)f5+ is a perpetual check. 18•••~ 19l1)g5! Not 19 "h7 .tf5+! (Burgess) 20 .xf5 .tf6 and White is in trouble. 19•••.txgS The lines 19....txa3?? 20 ll)h7+ q;e7 21 and 19...ll)a5?? 20 Wh7 .tf5+ 21 "xf5+ .tf6 22ll)h7+ ~e7 23 "e4+ ~d7 24ll)xf6+ both win for White, while 19... hxg5 20 hxg5 opens the h-file for White's attack. 20 hxg5 ll)e7 21 gxh6 .tC5 21...ll)xd5 22 "h7 is extremely dangerous for Black, e.g. 22 ... gxh6 23

"e4+

59

.txh6+ ~eS 24 .tg5! ~fS 25 "hS+ ~f7 26 lIh7+ ~e6 27 lIh6+ ~f5 2S "h7+! ~xg5 29 "g7+ ~f5 30 lIh5+ leads to mate. 22 hxg7+ ~xg7 23 .te5+ ~g6 24 e4l1xd5! The alternatives 24 ....te6 25 "dl, 24 ....tg4 25 "d2 ll)gS 26 "f4 and 24 ...ll)xd5 25 exf5+ ~f7 26 "e2 lIgS (26... ll)f6 27 lIh6) 27 lIh7+ ~f8 2S "h5 all win for White. 25 exf5+ ~5 26 .tel "b5 with a likely draw. A32)

13ll)b5!? (D)

Unexplored, yet I suspect this is White's most promising idea. 13•••ll)d5 Black's queen is out of play after 13 ...lIxdl+ 14 lIxdl.td7 15 .tc7 b6 16ll)g5, etc. 14.txdS 14 ll)g5 g6 15 h5 ll)xf4 is unconvincing, so White plays for a positional edge. 14•••exd5 14... lIxd5 15 lIxd5 exd5 16ll)fd4;!; a6? 17 .tc7 ll)b4 18 .txa5 ll)xc2 19 ll)c7ll)xd4 20 exd4 (20 ll)xa8? ll)b3+)

60

The Queen's Gambit Declined: 5 i.f4!

20.. Jlb8 21lLlxdS and White is winning. 14...•xbS IS .i.e4 i. 15 .!Drd4lLlxd4 16lLlxd4 ;t White has the standard active pieceplay against the isolated d-pawn, with Black's bishop-pair not counting for much. If 16....i.g4, 17 .i.c7 .i.xdl 18 l:[xdl .cS 19 .i.xd8, and the simplification process is well in hand. The general conclusion is that the exchange 11 ...dxc4 is premature, as it assists White's development without enhancing Black's counterplay to a corresponding degree. B)

11 .•.l:[d8 (D)

w

This is much better timing than in the variation just considered, but it is probably still not sufficient to equalize. Black can quietly meet 12lLlbS? with 12 ... l:[d7, and White has succeeded only in misplacing his knight, but there is a better knight move. 12 .!Lld2! Still chasing the lady. The attempt to make her a widow with 12 g4 is less convincing; in effect White is playing

the g4 variation with h4 thrown in, which leaves unresolved the question of how to make gS a killer blow when the knight still has an escape square on hS. After 12 g4: 1) Kasparov-Ehlvest, Novgorod 1995 continued 12....i.d7 13 ~bl dxc4 14 .i.xc4l:[ac8 IS gSlLlhS 16 .i.d6 g6! (correctly preferring good piece coordination to early simplification; Black collapsed remarkably quickly in Lugovoi-Komeev, Elista 1995: 16....i.e8?! 17 .i.xe7lLlxe7 18l:[xd8l:[xd8 19l:[d I i l:[c8 20 .i.b3lLlfS 21 'it'd3 g622lLld4 ± lLlxd4 23 exd4 .i.c6 24 dS lLlf4? 2S 'it'd2 1-0; if 2S ...lLlxdS, then 26lLlxdS Wxd2 27lLlf6+ wins a piece) 17 .i.e2 .i.xd6 18 l:[xd61:iJe7 19 Wb3 .i.c6 20 l:[xd8+ ':xd8 21 ':dl WfS+ 22 ~al ':f8 23 e4 WcS 24 'it'b4, forcing a level ending. 2) 12...dxc4 13 .i.xc4 .i.d7 14 gS lLlhS IS .i.d6 ':ac8 would transpose after 16 ~bl. Instead, 16 .i.e2 .i.e8 17 .i.xe7 ':xdl+ 18 ':xdllLlxe7 1/2- 1/2 was Rustemov-Kharitonov, Russia 1996. 12•••':d7 This looks awkward, but Black is creating a flight square for the queen without losing his grip on dS. 1) 12...dxc4 was recommended in Informator, both by Agdestein and by Cvetkovic, and indeed Black is doing reasonably well after 13 .i.xc4?! a6. However, after the more natural 13 lLlxc4! l:[xdl + 14 Wxdl Wd8 it is hard to concur with Cvetkovic's view that Black is equal. After IS Wxd8+ followed by .i.e2, White has a pleasant advantage in space and development to take through to the endgame. 2) The familiar plan of 12...eS 13 .i.gS d4 was seen in Cifuentes-Van der

The New Main Line Sterren, Amsterdam 1995, but the setting seems wrong for Black. After 14 liJb3 "b6 IS cS "c7 16liJbS "b8 17 exd4 a6, White let things slip to equalitywith 18liJd6?! i.xd619cxd6 "xd6 20 dxe5 "xdl+ 21 "xdl l%xdl+ 22
61

The text and notes follow analysis by Cvetkovic. 14.tgSd4 14...e4 15 .te21eaves Black's pawn centre badly over-extended. 15 liJb3 15 .txf6? dxc3 =t. IS..:W'd8 16 exd4 16 .txf6 dxc3! ;1;. 16•••liJxd4 16...exd4 17 .txf6 .txf6 18liJd5 ±. 17 liJxd4 exd4 18 .txf6 .txf6 19

liJds! Cvetkovic suggests instead 19 .txh7+ ~h8 20 liJds g6 21.txg6 fxg6 22 "xg6 '±', which is unconvincing as Black has 22 ...l%xdS 23 cxd5 "c7+ 24 ~bl "h7, exchanging queens and ensuring that Black has the better supported passed d-pawn. The text-move gives White a significant positional advantage without any risk whatsoever. White is about two or three tempi ahead of corresponding lines with 10 l%dl .te7 11 liJd2 e5 12 .tgS d4 13 liJb3 "b6 14 .txf6, etc., given the misplacement of Black's rook and the better placement of White's king. So another of Black's tries proves unsuccessful.

C) 11... a6 (D) Black has done well with this so far, but it looks a bit slow. 12liJgS Threatening a double capture on dS. 12...l%d8 12...dxc4 transposes to the ll...dxc4 line (A2 above). 13.td3

62

The Queen's Gambit Declined: 5 i.f4!

This direct attacking move is probably best. 13 cxdS exdS is the alternative: 1) In a game Van Wely-Me.Sharif, Linares Z 1995 (in which the exchange on dS actually took place on move 12), White lost the plot: 14 ~bl ? h6 16 lLlf3 i.g4 17 i.e2 l:lacS, and Black, having gained a couple of tempi, was powering ahead on the queenside. 2) I spent a lot of effort trying to make 14 e4 work: 2a) One attractive possibility is 14...dxe4?! IS i.c4 lLld4 16 i.xf7+ ~f8 17 l:lxd4 l:lxd4 IS i.a2! l:ld3 19 lLlcxe4 "iffS 20 "ifxd3 "ifxf4+ 21 ~bl and White should win. 2b) However, I finally had to concede defeat, after a few editorial suggestions, on the attempt to prove . k'.'. something for White after 14... lLlxe4!. \ ,. ""':' •• >- Black is happy after IS l:I.xdS l:lxdS 16 (.\ (it' ~:~ "ifxe4 l:lxgSl or IS lLlcxe4 dxe4 16 '. ~p /.p l:lxds+lLlxdSI7"ife4 "iffS!.ISi.c4is ! '/ . .. an interesting try, but after IS ... dxc4 ",' 16 l:lxdS+ "ifxdS 17 l:ldl "as! IS l:[dS? (IS "ifxe4 i.fS 19 "ifxc4 i.xgS 20 hxgS l:ldS Nunn) IS ...i.xgS! Black is on top. Black also probably has nothing to fear after IS ...i.fS 16

=

lLlxdSlLlg3 (or 16...i.xgS - Nunn) 17 'ii'b3 lLlxhl (Burgess). 13.••h6 It is not so pleasant for Black to have to make a pawn move in front of the king. 13 ... g6 14 hS leaves White with good attacking chances, and if 14...lLlxhS?! then IS l:lxhS gxhS 16 i.xh7+ followed by lLlxf7. 14 g4! With ideas of a knight sacrifice followed by gS. Naturally, capturing the knight on gS is extremely hazardous; if immediately 14... hxgS IS hxgS, White's plan is to play f3 to bring the queen to h2. 14•••eS!? White is obviously planning i.h7+ followed by lLlxf7, so 14...dxc4 doesn't really slow him down. Attacking the bishop might just make the difference; if it retreats the i.xh6 sacrifice will not be on. 15 i.h7+ Wf8 16lLlxf7 cJlxr7 16... exf4 17 lLlxdS keeps White's attacking chances for negligible sacrifice in material. 17"g6+ 17 i.xh6 e4! blocks White's lines of communication. 17•••~f818 i.xh6 gxh619 "xh6+ ~7

Running with the king does not really help, for example 19 ... ~eS 20 j,g6+ ~d7 21 lLlxdS and Black's queen is in trouble. 20lLlxdS I found plenty of perpetual checks, but nothing conclusive, after direct measures on the kingside. But taking control of the centre might do the trick. "ifg6+ is now the big threat. 2o•••lLlxh7 21 "xh7+ ~e8

The New Main Line 21...fB leaves time for White to run his g-pawn with 22 g5, etc. The attempt at counterattack with 22 ....txa3 23 g6 .txb2+ 24 xb2 leaves Black with no playable checks. 22"'g8+~d7

Or 22....tfB 23 "g6+ Wd7 24 Wt7+ Wd6 25 lDb6+ Wc5 26 lOxa8 with an easy win. 23"'r7! and Black is helpless against the threat of 24 lOxe7+ c7 25 lOxc6+ xc6 26 Wf6+. It is perhaps not surprising that Black has avoided, on general grounds, the line with 11...a6. Even so, conducting the kingside attack for White involves having to play very accurately.

0) 11•••.td7 (D)

63

12g4:fc813g5lOh514.th2.te815 lOd2 a6 16 cxdSlOb4? 17 axb4 "'al+ 18 lOdbl .ta4 19 b3 .txb4 20 .te5 f6 21 Wb2 1-0. Black's one-shot attack didn't work. Also, White can try to steer towards the types of position reached in the 10....td7 11 bl dxc4 12 .txc4 .te7 line (Chapter 3.6, Line B) with, for example, 12lOg5 or 12 bl.

Conclusion: At the moment, 11 h4! looks a very good prospect against the 1O ... .te7 system. The big plus that it has in comparison with 11 g4 systems (to be discussed shortly) is that White builds up pressure on the kingside without creating weaknesses. And if White has no structural pawn weaknesses in his position, that implies that he can shift seamlessly to positional lines if Black makes compromises in order to defend his kingside.

3.3 10...i.e7 11 g4 1 d4 dS 2 c4 e6 3lDc3lOr6 4lDf3 .te7 5 .tf4 0-0 6 e3 c5 7 dxc5 .txc5 8 "'c2 lOc6 9 83"'aS 100-0-0 .te7 11 g4 (D)

If Black wants to play the ... .td7 defence, he should do so on move 10. The inserted 1O....te7 11 h4 favours White, one important detail being that White's bishop has a retreat on h2. Practical experience is limited, but the game H.Olafsson-0gaard, Gausdal 1996 is unlikely to appeal to Black:

The Queen's Gambit Declined: 5 1./4!

64

This was what White usually played when the castling line first came into fashion. White makes an immediate and direct lunge at the king, but he must ask himself, first what he does when the pawn reaches gS, and secondly whether he is weakening himself on the dl-hS diagonal. The direct approach with 11...dxc4 12 J.xc4 eS 13 gS exf4 is Black's most popular, and will be considered first. The pressure on dS is so strong that it is difficult to see convincing ways for Black to avoid the exchange on c4. OUT lines are:

A: 1l•••dxc4 B: 1l ...:d8

64 72

A)

1l...dxc4 12 J.xc4 Now:

AI: 12...eS A2: A3: A4: AS:

64 69 12...lbxg4?! 69 12..':d8 70 12...bS?

12...a6

71

A1) 12...e5 13 gS exf4 14 gxf6 J.xf6

(D)

ISlbds Bringing up the reserves. White's score with IS :dS is an unimpressive 115 on my database, but the move is not as bad as all that. lS ...•c7 and now: 1) 16 J.d3 g6 17 :gl J.g7 18 h4 lbe7 19 :dgS fxe3 + Cais-Ciolac, Schwabisch Gmund 1996. 2) 16lbe4 .e7! 17 h4 g6 18 h5 J.fS 19 lbxf6+ .xf6 20 e4, Nasybullin-Klovans, corr. 1988-90, and now Black played 20...J.e6 co, missing the startling defensive idea demonstrated by Nasybullin: 20...J.g4! =+= 21 hxg6? J.xf3 22 gxf7+ ~h8 23 :xh7+ 'itxh7 24 eS+ J.e4!! (24...•g6? 2S J.d3 +-) 2S .xe4+ .g6 -+. This line features a problem theme, with the bishop sacrifice drawing the queen over the critical d3-square. 20... lbe7! is also strong, ifless spectacular, as 21 exfS is met by 21...lbxdS 22 J.xdS :ac8 23 J.c4 .c6. Agrest-Sanden, Stockholm 1992 continued instead 21 :d4 J.g4 + 22 eS .c6 23 hxg6 hxg6 24 e6 (24 :d6 .xc4! -+) 24 ...•xf3 2S exf7+ ~g7 26 :e1 :ac8 and White's attack had run out of steam. 3) 16 l:lhS g6 17lbdS .d8 (after 17...•d6 18 :gl fxe3 19 fxe3 b5 20 J.a2 J.g7 21 lbf4, GogichaishviliTabatadze, Bmo 1991, was agreed drawn just in time to prevent massive complications breaking out) 18 :gl J.e6 (18 ... J.g7? 19 :xg6 hxg6 20 .xg6! fxg6 21lbe7#) 19lbgS J.xgS (19 ... J.xdS 20 J.xdS .xdS 21lbxh7 "'d8 22 :xg6+ fxg6 23 "'xg6+ J.g7 24lbgS +-) 20 :hxgS J.xdS 21 J.xd5 ~g7 co Libeau-A.Hoffman, Bie11993. We conclude that 15 :d5 .c7 16 :hS deserves further exploration.

The New Main Une 15•••~e7 Many of the apparently more natural moves lose very quickly. 1) IS ... .ie6? 16 ~xf6+ gxf6 17 :hg1+ ~h8 18 .ixe6 fxe6 19 :d7 +Kruppa. 2) IS ....ie7? 16lO
65

alive, although there is still much to be explored. Ultimately, I would expect the assessment to run in White's favour. 16~xf6+

16 :hgl ~h8 17 ~xf6 gxf6 transposes. 16 h4 ~xd5 17 :xdS "c7 18 ~gS .ixg5 19 hxgS g6 Kruppa. 16•••gxf6 (D)

+

w

21~.

4) It is a little surprising that IS ....id8 has not been tried more often. At the cost of a tempo, the bishop is preserved, and the g-file is not fully opened against the king. 4a) The sole Informator reference is a brief note by Kruppa (Informator 46) who suggests 16 h4, intending ~gS, with an attack. But Black is allowed to counter-attack: 16... .ig4! 17 ~gS g6 18 f3 (18 :dgl ~) 18 ...~S! 19 fxg4 :c8 20 b3 (20 :d4? :xc4 21 lbc4 .xdS 22 :e4 ~d3+ 23 ~bl .ixgS -+) 20...•xa3+! (20 ... bS? 21 • b2 ±) 21 "b2 ~xc4 22 bxc4 :xc4+ 23 ~bl "xb2+ 24 ~xb2 fxe3 and Black has liquidated and won enough white pawns to feel secure against White's extra piece. 4b) 16 :hgl is critical, a recent fiasco for Black being 16...h6? 17 "e4 ~h8 18 :xg7 f5 19 .xf4 ~xg7 20 :g1+ ~h7 21 "g3 1-0 SummerscaleMednis, Cannes 1997. 16... fxe3 17 fxe3 g6! (17 ....ie6 18 ~gS ±), and if 18 h4, then 18...b5 keeps the position

The main subject for theoretical discussion in the early 1990s. White hopes to make use of the exposure of the black king, while Black hopes to get his pieces out quickly to expose White's weaknesses on the queenside. After several practical tests it seems that Black is doing OK. 17:hgl+ The natural move, although White can also hold fire, and indeed might well be advised to do so if his intention is lO
66

The Queen's Gambit Declined: 5 i..f4!

and White finds there is no way to take advantage of the wayward position of Black's knight since 23 rle4 is met by 23 ... fS. The clearest way to the draw is given by Van der Sterren as 23 lIe7 lLld4 24 lIxb7lLlf3! 2S .i.dS lLleS. If White had given a check on gl, ... lLlO would have gained a tempo. 17 e4 will usually end up with White throwing in a check along the g-file, for example 17 ... i.h3 18 ~bl llac8 19 l:hgl+. 17.••~h8 Play now branches out into: All: 18 We4 66 AU: 18lLld4 67 A13: 18e4 67 A11) 18 We4 (D)

B

19••• Wb6 Forcing the exchange of queens. White has not had much luck trying to win this position. 20Wxb6 20 exf4 Wxd4 21 l:xd4 .tf5 22 ~d2 rlac8 23 l:cl.tg4 24 l:c3 l:cd8 =Gulko-Rodriguez Talavera, Seville 1992. 20...axb6 21 l:d6 .th3 Or alternatively 21...fxe3 22 fxe3 l:aS! (22 ... i.fS?! 23 rlxf6 i.e6 24 .td3 l:fd8 2S lLld4 ± AgamalievSmagacz, Koszalin 1997) 23 d2 (23 .tdS bS! 24 ~d2 l:a6 =; 23 rlxb6 rlcS 24 b3 .te6 2SlLld2lLleS =Khalifman) and now: 1) 23 ... l:fS 24 rln rlcs 2S .td3 l:c6 is equal, Kasparov-Khalifman, Reggio Emilia 199112. 2) 23 ...l:cS 24 .tdS (! Khalifman) 24 ... 'it>g7 25 lLld4 rle8 is more adventurous for Black, but after 26 e4 l:e5 27 l:n lLle7 28 l:dxf6 lLlxd5 29 l:xf7+ ~g8 30 exdS rlexdS, White can try 31 l:f8+ ~g7 32 l:1f7+ ~g6 33 l:f4! instead of 31 ~e3?! l:d8 32 l:e7 rlh5 ~ Lputian-Van der Sterren, Wijk aan Zee 1993. 22~b1

18•••lLlg6 18 ...•c5? 19 .xf4 .tfS 20 b4 'ii'c6 and now 21 ~b2 ± Khalifman, rather than 21 l:d6? Wa4 22 .t..d3 .xa3+ 23 ~dl l:fd8 + Ambartsumian-Magomedov, Frunze 1989. 19Wd4 19 ridS? Wc7 20 Wd4 i.e6! 21 l:cs We7 +Shabalov-Kruppa, Frunze 1989.

After 22 .tdS!?, 22 ... rlac8+?! 23 'it>bl l:cd8 24 rlxd8 l:xd8 25 .t..xb7 fxe3 26 fxe3 l:d3 27 lLld4 ;t was Vera-Van der Sterren, Lucerne Wcht 1993. Black may be a pawn up after 27 ... l:xe3, but if the weak b-pawn should fall, White will have two connected outside passed pawns. Instead, 22 ...lLle5! 23lLlxe5 fxe5 24 exf4 (24 .txb7 fxe3) holds the balance. 22....t..rS+ 23 ~a1 The safest square. 23 'it>a2 leaves the king too exposed: 23 ... i.e4 24

The New Main Line lbd4 l'Lle5 25 .i.b3 fxe3 26 fxe3 l:tg8 27 l:tn b5 28 l:tf4? l'Lld3 29 .i.c2 (29 .l:lxe4?? l'Llb4+ mates on the back rank) 29 ... .i.d5+ 30 l:txd5 l'Llb4+ winning the exchange, Akopian-Kruppa, St Petersburg 1993. 23.-..i.e4 24l'Lld4l'LleS 2S .i.bS fxe3 26 fxe3 l:tgS Gelfand-Yusupov, Linares 1993.

+

A12) ISl'Lld4 (D)

67

fxe3 .i.f5 19 l'Llxf5 .xf5 20 l:thgl + is to be preferred. 20 l'Llxes ..xes 20 ... l'Llxf5? would be mistimed: 21 l:[d5 .b6 22 ~bl .xe3 (22 ... l'Llxe3? 23 .xh7+ ~xh7 24 l:th5#) 23 l:tg4! l'Llg7 (23 ...lLle7 24 l:th5l'Llg6 25 l:txh7+ ~xh7 26 .i.xf7 .h6 27 .i.xg6+ ~h8 28 'iff2, followed by l:th3 winning . Pelletier) 24 l:td3 .el + 25 ~a2 and Black is in trouble on the kingside, Pelletier-Van der Sterren, Zurich 1995. 21.i.d3 "eS 21.. .•c8? 22 .i.xh7 f5 23 l:td4! Akopian. 22~blfS

White seems to have adequate compensation for the pawn, but perhaps nothing more. After 23 .b3. Akopian-Pigusov. Tilburg 1994 continued 23 ... l:tac8 24 .i.c2 .f6 (24 ...l:tc6? 25 .xf7; 24 ... b5!? 25 l:td7 as ao Akopian) 25 l:td7 ±, although Akopian notes that Black could also have tried 23 ...lbd5. White's next try, after 18 .e4 was found to be prospectless. IS___fxe3 18 ... .i.f5 19 l'Llxf5 .xf5 20.xf5 l'Llxf5 21 .i.d3 ;t Akopian. 19 fxe3 .i.fS Black could also try 19...•e5!? 20 c,i;>bl .i.f5! 21 l'Llxf5 l'Llxf5 22 l:td5 l'Llxe3, much as in Hiibner-Van der Sterren, given in the notes to White's 17th. There, however, White had taken care to avoid l:thgl+, the reason being that after 23 l:txe5 l'Llxc2 24 l:te7 l'Lld4 25 .i.d3 l'Llf3 Black gains an important tempo. This would imply that 17 l:thgl+ c,i;>h8 18 l'Lld4 is in fact an inaccurate move-order, and that 17 l'Lld4 fxe3 18

A13) ISe4!? (D)

B

The latest try. ·slowing down his own initiative. but preventing any black counterplay on the h7-bl diagonal.

68

The Queen' s Gambit Declined: 5 1./41

18... bS The most direct. 18 ... .i.h3 19 ~bl l:tac8 20 "'b3 wins the f7-pawn and gives White a clear positional advantage. Akopian-Egiazarian. Annenian Ch 1996. 19 .i.dS .!DxdS Black has many attempts to sharpen the play still further. but this simple and obvious move seems best. After 19 ... l:tb8?, 20 "'c5 is surprisingly difficult to meet. while the various attempts to 'win' queen for two rooks favour White in view of his control of the open d- and g-files, and the vulnerability of Black's king. On 19 ... .i.e6. 20 .i.xa8 l:txa8 21 ~bl b4 22 "'d2 squashes another attempt to counterattack. 20exdS (D)

B

20••• b4 On quieter moves. White consolidates with ~bl. though 20 ....i.g4 is an intriguing alternative. The critical line runs 21 "'e4 .i.h5 22 "'xf4 .i.g6 23 "'xf6+ ~g8 and now: 1) 24 .!De5 l:tac8+ 25 .!Dc6 b4 26 l:td3 and now 26 ... l:tfe8 loses to 27 l:tc3!. but 26 ... bxa3!? the situation is less clear.

2) 24 l:td3! is a clever idea. using pin and counterpin to clear a route to bl for the king, and this seems to cause serious problems for Black. After 24 ......a4 White continues 25 o!iJd4! (25 .!De5 "'e4 is less clear-cut) 25 ... l:tfc8+ 26 o!tJc6 "'e4 and now the simplest line for White is to force a very favourable ending by 27 "'d4 "'xd4 (27 ......f5 28 ~bl) 28 l:txd4. 21 axb4 The only move to have been tried so far, though there are certainly alternatives to be considered. 1) 21 "'c6?! .i.e6! (21.. ....b6 22 "'xb6 axb6 23 axb4 ;1;) 22 ~bl (22 "'d6? bxa3 23 dxe6? axb2+) 22 ....i.f5+ 23 ~a2 "'b6 24 "'xb6 axb6 =i=. White would be doing very well in the 18 e4 variation, if only he had a safe place for the king! 2) 21 l:td4!? bxa3! 22 b4 a2! 23 bxa5? al"'+ 24 ~d2 "'xa5+ =1= gives a highly practical setting for the Excelsior theme. or as close to it as you are likely to get in an over-the-board middlegame, in which the humble pawn. metaphorical sword in hand. marches boldly from home square to queening square. nonchalantly brushing aside all the diverse obstacles in its path. But in chess, heroics often require cooperation from the opponent in order to succeed. 23 ~b2! l:tb8 24 "'b3 is probably best described as unclear. at least until practical tests emerge. 3) 21 a4?! .i.g4! is too slow. 21 •.:.al+ 2l......xb4 22 l:td4;1; Short. 22 ~d2 'ii'a6 (D) 22 ... 'ii'a2!? 23 "c3 "'xd5+ 24 ~cl "'e6. Short. If 25 .!Dg5. 25 ......e7! defends. but not 25 .....e5?? 26 .!Dxf7+.

The New Main Line

69

It would seem that the main line 11 g4 dxc4 12 .i.xc4 eS is holding up well for Black. We now consider other possibilities for Black, starting with those where Black has already exchanged pawns.

A2)

12•••bS? (D)

23~d4

23 "'c6 is not as strong as it looks, as Black has 23 .. J:ld8! and if 24 'fixa8?, then 24......c4! leaves White's defences totally uncoordinated. The queen cannot return, since 25 'fixa7 fails to 25 ...:xd5+ 26 ~el "'e4+ and Black wins. White's best defensive try is to set up some back-row tricks with 2S :del, but after 2S ......xb4+ 26 ~c1 'ficS+ 27 ~d2 :xd5+ (27 ..."'as+ 28 ~cl "'al+ 29 ~c2! .i.fS+? 30 ~b3! turns the tables) 28 1i'xdS 'fixdS+ 29 ~c 1 1i'cS+ 30 ~d2 .i.e6 Black should win. White avoided this humiliation in Van Wely-Short, Wijk aan Zee 1997, but after 24 ~c3 .i.b7 2S "'xa6 .i.xa6 26 :d4 :ac8+ 27 ~d2 .i.b7, Black won the d-pawn, reaching a drawn endgame in which White later went astray. 23•••:d8 24 bS 1i'b6 2S 1i'e4 .i.b7 26"xf4l:[xdS Both sides determinedly centralize, and both sides just about consolidate. Akopian-Short, Groningen 1996, was agreed drawn, with repetition of position looming, after the further moves 27 ~cl fS 28 "'gS :cS+ 29 ~d2 "'g6 30"'e7"'b6.

w

A reckless pawn sacrifice. White can safely capture either way. 13~xbS

13 .i.xbS .i.b7 14 ~d2 ~b4?! (l4...~S 15 ~4 "'d8 16 e4 .i.gS 17 exdS .i.xf4+ 18 ~bl gives White a distinct advantage - Gurevich) IS axb4 .i.xb4 16 ~4 1i'al+ 17 ~d2 .i.xc3+ 18 ~e2 'fia2 19 :al 1-0 M.Gurevich-A.Sokolov, USSR Ch 1988. 13•••eS 14 ~eS ~eS 15 .i.xeS .i.xg4 16 .i.c7 1-0 Bertholee-Peelen, Wijk aan Zee 1995. If 16....i.xdl, 17 :xdl1i'a6 18 ~d6 'fic6 19 .i.xf7+. If only points came so easily all the time! A3)

12.•.~xg4?! (D) This also sails too close to the wind. 13:hgl'ii'hS Or:

70

The Queen's Gambit Declined: 5 1./41

1) 13 ... e5 14 .i.gS ~f6 (otherwise ~d5, etc.) 15 .i.xf6! .i.xf6 16 ~e4 .i.e7 17 ~gS, and White should win, e.g. 17 ... g6 18 ~xf7 l:txf7 19 l:txg6+ or 17...e4 18 l:td5. 2) 13 ...~ge5 14 ~xe5 ~xe5 15 lDd5 ±. 3) 13 ... ~f6 14 .i.h6 ±. 14 h3 ~f6 15 .i.e2 l:tdS 16 ~e5 'ifh4 17 .i.g5 'ifxf2 IS 00 White soon snared the black queen in G.Georgadze-Kharitonov, Simferopol 1988 (18...clOds 19 .i.h6 g6 20 ~ "xgl 21 l:txgl), but somehow failed to win the game. A4) 12•••l:tdS (D)

w

Not often seen, as the offer to exchange rooks slows White down mther less than the various counter-attacking possibilities. 13 e4!? A little more testing than 13 h3.i.d7 14 e4 .i.e8 1/2- 1/2 Bagaturov-Dzhandzhgava. Tbilisi 1996. 13...~xg4 13 ....i.d7 14 e5 ~e8 15 ~g5 ± Kharitonov. 13 ...l:txdl+ 14 l:txdl ~xg4 15 l:tgl gives White good attacking chances. 14l:tdgl A sane and rational move. 14l:thgl led to crazy complications in Shabalov-Kharitonov, Leningmd 1989. The whole game is worth reprinting for the stunning swindle through which Black knocked back a winning attack: 14... l:txdl+ 15 1I'xdl ~xf2 16 1I'n .i.xa3 17 .i.d2 .i.c5 18 ¢>bl b5?! (18 .....d8 00 Kharitonov) 19lDd5 "d8 20 l:txg7+! ¢>xg7 21 "g2+ ¢>h8 (21...¢>f8 22 .i.h6+ ¢>e8 23 1I'g8+ ¢>d7 24 ~5+! ~xe5 25 .i.xb5+ +Kharitonov) 22 .i.c3+? (22 .i.g5! .i.e7 23 ~5!! +- Shabalov) 22 ... lDd4 23 ~xd4lDdl !! 24 1I'g3 ~xc3+ 25 bxc3 .i.xd4 26 cxd4 f6 0-1. 14••• ~geS 15 ~xe5 ~xe5 16 ~dS exdS 17 .i.xe5 .i.f8 If 17 ... f6, then the reply 18 .i.c3! keeps White in control. Not, however, 18 .i.c??? "xc7 19 .i.xdS+ l:txd5 20 "fIxc7 l:tc5+. ISexd5! Kharitonov's analysis continues 18 .i.xd5 lIxd5! 19 exd5 "xd5 00. But if White can gain a tempo, the assessment of the position will change considerably. IS•••l:txdS

The New Main Line It's difficult to see how Black can avoid the exchange sacrifice; if 18 ... b5, then 19 .i.d3 ±. 19.i.e3! The gain of tempo. 19••:.cS 20 .i.xdS 'iVxdS 21ltdl ± With a sound material advantage for White. 21 .d2?! .i.e6! is more tricky, since 22 WgS :c8! 23 .f6?? loses to 23 .....xhl !. AS)

12...a6 (D)

Quieter than 12 ... eS, and therefore not as popular, but maybe just as reliable. The threat of ... bS-b4 means that White has little time to build up a kingside attack, and so must release some of the tension. 13gS After 13 .i.d3!?: 1) 13 ... ~h8 was answered by 14 gS 112-112 in Ivanchuk-Ehlvesl, Novgorod 1995. Presumably 14 ... lDg4 IS .*.xh7 (IS ':hgl lDgeS 00) Is ...lDxf2 holds no joy for White. Even so, there is plenty of unexplored territory here, for example 14 .*.g3lDxg4 IS .i.xh7. 2) A recent try for Black, reached via 1O...dxc4 11 .i.xc4 a6 12 .*.d3 .i.e7

71

13 g4, is 13 ... g6, for example 14 gS lDh5 IS h4 eS 16 .i.h2 .i.e6 17 .i.e4

':ac8 18 ~bl .i.xa3 19 .a4 .*.b4 20 .xaS lDxaS 21 .i.xeS ~4 22 .i.d4 lDd6 ; R.Kempinski-R.EkstrOm, Cappelle la Grande 1997. 13...lDhS The thematic square, blocking any pawn-roller, and preserving options on the bishop. 13 ... lDd7? is a mistake. Mter 14 ~4 ~S ISlDf6+! White has a winning attack, Kaidanov-KonsaIa, Poland 1988. Kaidanov gives the main line as IS ... gxf6 16 gxf6 .*.xf6 17 ':hgl+ ~h8 18lDgS .i.xgS 19 .i.xgS f6 20 .i.h6 ltf7 21 f4! lDe7 22 .g2 lDg6 23 fS +-. 14.i.d6 Or 14 .i.d3 g6 (14...~h8!? Beliavsky) IS .*.e4 (IS cttbl!? Vladimirov) Is ... lDxf4 (1S ... eS?? 16lDxeS wins a pawn, Styrenkov-Gasthofer, Podolsk 1993) 16 exf4 .c7 17 h4 "xf4+ 18 ~bl hS (18 ...f5? 19 gxf6 .*.xf6 20 hS ± Vladimirov) 19 ~2 "c7 20.i.xg6 fxg6 21 .xg6+ ~h8 E.VladimirovBass, Matalascaiias 1989. Obviously White has a draw whenever he wants it, but it is unlikely he can do better. 14....i.xd6 15 .J:xd6 tOes 16 .*.e2 lDxf3 17 .i.xf3 "xgS 18 ~ ..rs 19 .i.xhS 1i'xhS 20 ':gl Agreed drawn in Gelfand-Beliavsky, USSR Ch 1989. A few years later this position was played out: 20...fS (20 ...•h6!? Beliavsky) 21lDf6+':xf6 22 .c7 .h6 (22 ... g6? 23 ':d8+ ':f8 24 ':xf8+ ~xf8 2S ':dl +-) 23 ':d8+ ':f8 24 ':xf8+ ~xf8 2S ':dl .i.d726 ':xd7 bS, led eventually to a draw in the game Akopian-Beliavsky, Erevan OL 1996.

72

The Queen's Gambit Declined: 5 J.f4!

B)

1l...:ct8 (D)

w

In some of the earlier games with 11 g4, Black tried to hold firm in the centre. However, experience has shown that in this particular variation White's attack is better met by active pieceplay than by trying to keep the central pawn tension. 12h3 Consolidating his gains. 12 li)d2!? leads to simplifications after 12... dxc4 13 li)xc4 :xdl+ 14 Wxdl Wd8. In comparison with corresponding lines after 11 h4, the vulnerability of the g-pawn costs White a tempo, but he still has chances of keeping an edge. After 15 Wxd8+ .i.xd8 (15 ...li)xd8 16 .i.e2 li)d5 17 :dl!? li)xf4 18 exf4 f6 19 .i.f3 ~f8 ;t Agrest-Ambartsumian, Podolsk 1989) 16 h3lai5 17li)xd5 exd5 18lai6 .i.c7 19 .i.g2 .i.xd6 20 .i.xd6 .i.e6 White had fractionally the better of the draw in Malaniuk-Kharitonov, Simferopol 1989. 12...a6?! It is surprising that Black has not tried 12...dxc4 here. Speelman gives 13 :xd8+ Wxd8 14 .i.xc4 =, while 13

=

.txc4 .i.d7 leads to a quiet sidevariation of Line A4 above. 12....td7 13 ll)d2 a6 transposes; Black must be careful not to let his queen get snagged by li)b3. 13lOd2 Black has tried a number of moves here, without challenging the verdict that White is better. 1) 13 ....td7 14 .te2 :ac8 15 g5 ;t (now that h5 is no longer available for the knight) 15... li)e4 16 li)dxe4 dxe4 17 li)xe4 li)e5 18 :d4 .i.c6 19 :hdl li)g6, and after 20 :xd8+?! :xd8 21 :xd8+ J.xd8 Black had excellent compensation for the pawn in M.Gurevich-Kharitonov, USSR Ch 1988. Gurevich noted as better 20 .td6, with the idea 20....i.xe4 21 Wxe4 J.xd6 22 :xd6 :xd6 23 :xd6 WeI + 24 .i.d 1 Wxf2 25 :xe6. 2) 13 ...e5? 14 g5! li)e8 (14 ...li)h5 15 .th2 ±; 14...li)e4 15lbb3 Wxc3 16 bxc3 .txa3+ 17 ~bl .tf5 18 ~a2! .tb4 19 .txe5 ± Speelman) 15 lbb3 "b6 16li)xd5 :xd5 17 cxd5 ± Speelman-Short, London Ct (3) 1988. The rest of the game is given in the introductory chapter. 3) 13 ... b5 14 cxd5li)xd5 15li)xd5 exd5 (15 ...:xd5 16 ~bl Wb6 17 .tg2 :c5 18 "d3;t {Nunn} is safer but insufficient to equalize) 16 Wxc6 .tf5 17 li)b3! Wa4 18 gxf5 :ac8 19 .tc7 Wxb3 (19 ....td6? 20 ll)c5! - Nunn; if then 20....txc5, 21 "xc5 :d7 22 'ifd4!) and now Nunn in his original analysis cited in Informator gives 20 .tg2 '! ±'. This is not correct, since 20...:d6! is fully satisfactory for Black after 21 Wc2 (certainly not 21 "b7? Wc4+, nor 21 Wc5? :d7 22 .txd5 "xdl+) 21...Wxc2+ 22 ~xc2 :xc7+

The New Main Line

.a2

with equality. However, 20 :d3! 21 :c3 seems to preserve a substantial advantage for White. Then 21...-*.f6 22 -*.xd8 :xc6 23 -*.xf6! is an attractive simplification, while 21.. .•al+ 22 ~c2 -*.f6 (22 ...•el 23 -*.xd8!) 23 -*.g2 24 -*.xd8! :xc6 2S ~xf6, Nunn, wins in similar fashion. 4) 13 ...•b6 14 -*.g2 d4 IS lOa4 'fIa7 16 ~xc6 bxc617 exd4 .xd4 18 -*.e3.d3 19 .xd3 :xd3 20 gSlOe8 21 ~c2 :d8 221Ob3 eS 2310aS -*.fS+ 24 ~c3 ± E. Vladimirov-Campora, Moscow 1989.

"a2

3.4 10...i.e7: White's 11th move alternatives

.c2

1 d4 dS 2 c4 e6 31Oc31Of6 4 00 -*.e7 5 -*.f4 0-0 6 e3 c5 7 dxc5 ~xcS 8 1Oc69 a3 .a510 0-0-0 ~e7 (D)

A: 11lOb5? B: 11 adS

c: 11~g5

D: 11lOd2

E: 11 ~bl

73 73 73 73 74 7S

A) 11 lOb5? is pointless, and just makes it easier for Black to open the c-file: ll...dxc412-*.xc4~d713~c7 b614:xd71Oxd71S.e4a616.xc6 axbS 17 .xd7 -*.xa3 18 .xbS :fc8 19 bxa3 :xc7 20 ~b1 .xa3 21 :c1 :ac8 22lOd2 .d6 23 lOb3 :xc4 0-1 Deak-Bauer, Hungary 1993. Or, with fewer complications, 11...~d7 12 ~c7 b6 13 lOd6 (13 b4?! -*.xb4 14 axb4 lOxb4 followed by ...~xbS, etc.) 13 ... -*.xd6 14 ~xd6 l:tfcS IslOd2 bS 16 cS b4 + Vescovi-Arlandi, Catania 1995. White is wasting too much time. B)

11 lOd2 and the prophylactic 11 ~bl have occasionally been tried in grandmaster play, and indeed were to become, for some obscure reason, relatively fashionable in 1997. White is at best aiming for a slight edge. The first three of our lines are strictly minor alternatives:

11 cxd5 lOxdS 12 lOxdS exdS 13 -*.d3 would be a good idea if it weren't for the placement of the white king! N.Saleh-I.Seitaj, Thessaloniki OL 1998 continued 13 ... h6 14 lOeS ~e6 IslOxc6 bxc6 16 ~eS cS 17 g4 d4 IS .i.c4.i.xc4 19 'ii'xc4 :re8 =t. But Black can do better still with 13 ...~e6! followed by quickly playing a rook to cS, when White is in difficulties, for example 14 -*.xh7+ ~hS IS -*.d3 l:tacS (ls ...lOd4? 16 -*.c7!) 16 .d2 lOeS+ 17 -*.c2 (17 'ittbl 'ii'xd2 IS l:txd2 lOxd3 19 l:txd3 .i.fS) 17 ....i.fS! IS 'ii'xaS l:txc2+ 19 ~b1 l:tcS+ and Black wins.

C) 11.i.g5 :d8 12 -*.d3 dxc4 13 .i.xc4 .i.d7 14 ~b1 kS =Kveinys-Klovans,

74

The Queen' s Gambit Declined: 5 ~4!

Groningen 1991, is too slow to cause Black problems. Back now to the more serious variations.

0) 1l~2(D)

By analogy with the 10 l:1dl line. 1l... eS

Preserving the analogy. I) 11...dxc4? forces Black's queen into adrenaline-sapping exertion: 12 lbxc4111cs 13 b4 1IIh514.i.e2 1IIg61S .i.d3111hS 16 h3lbdS 17 g4111h4 18 .i.g3 'iWh6 19 J.e4 ± Avrukh-Gild.Garcia, Linares 1997. 2) 11 ....td7 is a little inflexible. As we shall see later (section 3.6 below), Black often prefers to retreat his bishop to f8 rather than to e7 in the 1O....td7 line. 2a) White tried too hard in AgrestG.Ginsburg, Pinsk 1993: 12lbb3 1IIb6 13 e4?! eS 14 cS .d8 IS exdS exf4 16 dxc6 bxc6. All he had succeeded in doing was in giving Black the bishoppair in an open position. 2b) R.Kempinski-Lutz, Groningen 1996 continued 12 g4 :fc8 13 .te2!

(White's timing was less successful in G.Georgadze-Vaganian, Pula Echt 1997: 13 gS lbhS 14 lbb3 .b6 IS cxdS lbxf4 16 exf4 lbaS 17 lbxaS WxaS 18 :d3 .td619 .th3 .txf4+ 20 ~bl .txgS 21 'iWe2 'iWa6 =1=) 13 ...eS 14 gSlbe8 IS .td3 d4 16lbb3 Wd8 17 exd4±. 3) 11. .. a6 is discussed, under 1O... a611 ~2 .te7, in section 3.7 below. White is doing well. 4) 11 ... Wb6 !? is the latest try, this time by analogy with II ~bl a6 12 ~2 111b6. Line E below: 4a) 12 lbb3 lbaS 13 lbxaS .xaS 14 e4 (14 cxdS exdS =) 14... dxe4 IS lbxe4lbxe4 16 1IIxe4 .i.xa3 17 bxa3 Wc3+ 18 Wc2 Wa1+ 19 1IIbl (not 19 ~d2?? 'iWd4+) 19 ....c3+ 20 .c2 1/2- 1/2 Gabriel-Lutz, Bad Homburg 1997. 4b) 12 .td3 d4 13 cS WxcS 14 lbb3 Wb6 IS exd4 ~aS (IS ... .td7!?) 16 ~S .txcS 17 dxcS ~b3+ 18 ~bl lbxcs 19 .teS (19 .te3!? .c6 20 .tbS Wc7 21 lbdS ~xd5 22 1IIxcS with good compensation for the pawn; there are probably other reasonable tries) 19... clDxd3 20 :xd3 lbg4 21 .txg7 ~xg7 22 :g3 eS 23 h3 1IIg6 24 hxg4 .te6 2S .xg6+ hxg6 26 :tel 1/2- 1/2 Hebert-Fahrbach, corr. Internet 1997. 12.tg3 12 .tgS!? d4 13 ~b3 Wb6 14 exd4 ~xd4 IS ~xd4 exd4 16 .txf6 .txf6 17lbdS Wd8 is a formation more commonly seen after 10 :tdl: I) 18 .td3 should be met not by 18....tgS+? 19 ~bl ~h8 20 f4 .tf6 21 :hel ± Osterman-Krudde, Dieren 1990, but by 18 ... ~h8 =. 2) 18 Wd2!? .tg4 19 J.e2 (19 f3?? loses the exchange to 19 ....tgS; 19

The New Main Line lLlxf6+ "xf6 20 f3 is probably critical - White wins a pawn, but Black is well ahead in development) 19 ... .1xe2 20 "xe2 l:r.c8 21 ~bl g6 22 l:r.hel ~h8 23 "d3 .1g7 =Lima-Pelikian, Brasilia 1997. 12•••d4 12 ... dxc4?! 13lLlxc4 "c7 14 ~b5 'it'b8 15 lLld4 ± Ftaenik. 13lLlb3 'irb6 After 13 .....d8, Fta~nik gives both 14 exd4lLlxd4 15lLlxd4 exd4 16lLlb5 .1d7 17lLlxd4 l:r.c8 and 14lLle2 g6 15 exd4 .1f5 16 "c3 lLle4 17 "e3 as leading to a slight edge for White. The text-move looks more natural. 14 cS 1Fd8 Fta~nik points out that the extravagant-looking 14....1f5 is also possible. 15 e4 and now: 1) 15....1xc516lLla4.1xe417.1d3 .1xd3 followed by ..."c7 with compensation for the piece. Even so, 18 'irxd3 'irc7 19 lLlaxc5 l:r.fc8, designated as unclear by Fta~nik, looks good for White after the simple 20 ~bl. White then has excellent chances of keeping Black's extra central pawns blockaded, enabling him to use his extra piece effectively. 2) The even more extravagant 15 ... lLlxe4?! is best met by 16lLlxe4! .1xe4 17 cxb6 .1xc2 18 ~xc2 ±. Instead 16 cxb6?! lLlxg3 17 .1d3 .1xd3 18 .xd3 lLlxhl 19 l:r.xhl dxc3, assessed as ± by Fta~nik, seems thoroughly unclear. 15 exd4 lOxd4 After 15 ...exd4 16 lLlb5 .1d7, 17 lLl3xd4 lLlxd4 18 l:r.xd4 transposes, while Fta~nik also gives a massive tactical liquidation into a level ending: 17 lLl5xd4 lLlxd4 18 l:r.xd4 l:r.c8 19 .1b5

75

"e8 20l:r.xd7lLlxd7 21 l:r.dllLlxc5! 22 .1xe8 lLlxb3+ 23 ~bl l:r.xc2 24 ~xc2 l:r.xe8 25 ~xb3. 16 lLlxd4 exd4 17 lLlbS .1d7 18 l:r.xd4 18 lLlxd4 .1xc5; 18 b4 .1xb5 19 .1xb5 a5 - Fta~nik 18....aS! Improving on the somewhat mechanicaI18 ...l:r.c8?! 19lLld6 .1xd6 20 .txd6 l:r.e8 21 .1c4 b6 22 b4 ± Fta~­ nik-Klovans, Erfurt 1993 . After the text-move, 19 b4?! "a6 gets White nowhere in view of the weakness of the a-pawn. Other lines lead to equality, for example 19 .1d6 .txb5 20 .1xe7 "el+ 21 l:r.dl "xe7 22 .1xb5 l:r.ac8 or 19lLlci6 b6! 20 lLlb7 .el+ 21 l:r.dl "e6 22 .tc4 .c6 23 .td6 .1xd6 24lLlxd6 "xe5. E)

11 ~bl (D)

B

Though fairly safe, this is a bit too slow to promise much. 11•••a6 1) 11.. ..td7?! seems a little awkward when White's king has already left the c-file, and vigorous play with 12 g4 or 12 h4 would be promising.

76

The Queen's Gambit Declined: 5 ~f4!

White played too slowly with 12 h3?! in Eslon-A.Hoffman. Javea 1992. and after 12...lIfcS 13 g4 bS 14 cxbSlDds IS lDd2 a6 Black already held the initiative. 2) H ... lIdS!? 12 lDd2 dxc4 13 .ixc4 (13ll)xc4l1xdl+ 14 Wxdl WdS =) 13 ...WfS 14lDde4ll)xe4 IS lIxdS+ .lxdS 16 ll)xe4 eS 17 .ig3 Wg6 IS .id3 'i!i'h6 19 lIdl .ie7 Van WelyJinrong Liang, Beijing 1997. A commendably straightforward way of taking advantage of the lack of pace of 11

=

~b1.

12 ll)d2 'ikb6 12 ... bS 13ll)b3 Wb6 14 cxdS exdS IS .igS .ie6 16 .ixf6 .ixf6 17ll)xdS .ixdS IS lIxdS lIac8 gives Black good attacking chances in return for the pawn, Van Wely-Khuzman, Amsterdam 1995. Sometimes the best way of 'defending' an isolated pawn is to sacrifice it. 13 ll)b3 ll)aS 14 ll)xaS "'xaS IS exdS exdS 16 .ieS .ie6 17 .id3 lIae8 18"'d2~h8!

At first it is difficult to see what this move is aiming at, but the Van WelyPigusov game will clarify matters. Black envisages a defensive formation with ... lDd7, .. .f6 and ....igS, and with the kingside thus secured it will be possible to try for something on the queenside. The less subtle 18 ... lLld7 19 .id4 ll)bS allows White to pressurize the isolated d-pawn: 20 ll)e2 Wxd2 21 lIxd2 ll)c6 22 .ic3 lIfd8 23 lIhd I bS 24 ll)f4 .igS 2S .ic2 ;t Van Wely-Van der Sterren, Antwerp 1997. 19f3 Black's pieces are much better placed to work with the isolani after

19l1)e4 Wxd2 20 ll)xd2ll)d7 21 .if4 f6 22 ll)f3 ll)cs = Gelfand-Short, Novgorod 1996. 19•••ll)d7 20 .id4 f6 21 "e2 .ig8 22 .ifS lIe7 23 'iVctlDeS 24 b4 Otherwise Black is clearly better. Now after 24 ... .ixb4 2S axb4 Wxb4+ 26 Wb2 We7 27 ll)a2 as 28 ~a11Dc4 29 WbSll)d6 30 Wbl bS Black had excellent compensation for the sacrificed piece in Van Wely-Pigusov, Beijing 1997. The recent theoretical battle with 11 ~bl has not gone in White's favour. Concluding our survey of 10....ie7, the recommendation for White is definitely H h4!, which, in conjunction with some of the innovations suggested in the text, promises a clear advantage for White. The older move II g4 leads to equally complicated positions, with long forcing lines, but Black emerges with a satisfactory position. Of the quieter moves, II ~bl is only equal, while 11 ll)d2 might perhaps offer the chance for a small edge.

3.5 10... dxc4 11 ..txc4 without 11 ....*.e7 1 d4 dS 2 e4 e6 3ll)c3ll)f6 4ll)f3 .ie7 S .if4 0-0 6 e3 eS 7 dxeS .ixcS 8 "'e2 lLle6 9 a3 "'as 100·0·0 dxe4 11 .txc4 (D)

Taken next as a matter of convenience, since 11.. ..te7 transposes to lines we have just looked at, after 12 h4! or 12 g4. 11•••a6 Again with transpositional possibilities.

The New Main Une

11...l:ldS 12 ~2 i.e7 13 lLlb3 l:lxdl+ 14l:lxdl Wh51Sf3lLla5 ought to be good for White, but he lost the thread, and later the game, in BleesPliester, Dutch Ch 1990, after 16 g4?! lLlxb3+ 17 i.xb3 WcS GO. Instead, 16 lLlxaS Wxa5 17 e4 gives White a menacing lead in development. 12lLlgS! Aiming for yet another transposition to the lO...i.e7line. Instead: I) 12 g4 and then: la) 12...i.e7 13 gSlLlh5, etc., leads back to lines already considered under lO...J.e7 11 g4 dxc4 12 i.xc4 a6 (section 3.3, Line AS). Ib) 12 ... eS?! 13 J.gS! is strong for White, since 13 ... ~xg4 is met by 14 h3 ±. Instead, Azmaiparashvili-A.Davidovic, Sydney 1990 finished 13 ...lLld7 14lLle4 i.e7 IS i.xe7 ~xe7 16lLlfgS lLlg6 17 ~6 1-0, illustrating perfectly the dangers of a mistimed ... eS. lc) Black offered a pawn race and lost in Yepez-Pau, Catalonia 1996: 12... bS 13 J.d3 b4 14 gS bxc3 IS gxf6 cxb2+ 16 bl Wxa3 17 i.xh7+ hS and it was White's attack that prevailed. 2) 12 J.d3 J.e7 13 g4 is another transposition to the lO...J.e7 line (3.3,

77

Line AS, 13 J.d3). Black's position seems, on the limited evidence so far, to be satisfactory. 12.••i.e7 12... bS?? 13lLlce4 +- Ruban. 13h4 13 J.d3 worked well for White in the game Alterman-Jinrong Liang, Beijing 1997: 13 ... g6?! 14 h4lLleS IS hS (tempting, but having provoked ... g6, White might want to spend a move preserving the bishop with IS J.e2!?) Is ... lLlxd3+ 16 l:lxd3 eS 17 lLlxh7! exf4 IS lLlxfS J.xfS 19 hxg6 fxg6 20 l:ldS Wc7 (20...lLlxdS 21 Wxg6+ i.g7 22 WeS+ i.fS and now 23 l:lh5!, as pointed out by Burgess, is an elegant way to conclude the attack) 21 l:lgS J.g7 22 Wxg6 Wf7? (22.....e7 offers longer resistance) 23 l:lhS+ 1-0. However, Black has no need to set up the target on g6; 13 ...h6 14 h4 lLleS should be OK for Black, as White's queen and bishop are the wrong way round on the bl-h7 diagonal. 13•••h6 Again Black cannot get on with his queenside attack: 13 ... bS 14 ~ce4 g6 IS i.b3 ± Ruban. 14 J.a2! We have now transposed back to the 10...J.e7 11 h4 dxc4 12 J.xc4 a6 13lLlgS h6 variation (3.2, Line A23), so let us be consistent and recommend the same move. White stands well.

3.610....id7 1 d4 dS 2 c4 e6 3lLlc3ll)f6 4 W i.e7 S J.f4 0-06 e3 cS 7 dxcS .i.xcS 8 lLlc6 9 a3 10 0-0-0 J.d7 (D) A straightforward developing move, which allows Black to place his king's

"as

"c2

78

The Queen's Gambit Declined: 5 J..f4!

B

rook on c8, which in tum vacates f8 for the bishop, where it is better placed defensively than on e7. Black contemplates defending on the kingside with minor pieces (... R.f8, ... R.e8, ... iLle7) and attacking on the queenside with pawns and major pieces. The one real drawback to this plan is that Black is voluntarily placing himself under pressure on the d-file; maybe White can make this work in his favour. White has several options here; we consider: A: 11 g4 78 B: 11 ~b1! 81 C: 11 adS 83 D: 11 R.gS 84 E: lliLld2! 8S F: 11 h4!? 8S

A) 11 g4 (D) This was played many times in the late 1980s and early 1990s, before White seriously questioned whether the g4 push was really achieving anything. Since the threat of gS and taking on dS is easily met by placing a rook on the c-file, other plans are now preferred, of which 11 ~bl is probably White's best choice.

11...:Cc8 11...dxc4!? may also be considered: 1) If 12 R.xc4, then not 12 ...l:.fd8?! 13iLld2iLle7 14iLlb3""6 IS gSiLlfdS 16 iLlxdS iLlxdS 17 iLlxcs 'ii'xcs 18 R.d3 ;!; Gulko-Portisch, Reggio Emilia 1990/1, but rather 12 ... l:.fc8! (Gulko). Black is doing well;:for White to win the pawn by 13 gSiLldS would be foolhardy. 2) After 12 iLld2 bS, 13 iLlde4 b4 14iLlxf6+ gxf6 ISiLle4 R.e7 16 l:.xd7 bxa3 17 bxa3 l:.ad8 leads to complications that tum out well for Black (Gulko), but 13lLlce4!? is worth a shot. e.g. 13 ... R.e7 14iLlxf6+ R.xf61SiLle4. White is much better after 13 ... R.xa3 14 iLlxf6+ gxf6 IS bxa3 "xa3+ 16 'ii'b2 'ii'xb2+ 17 ~xb2, since 17 ... eS is met by 18 iLle4. 12~bl

1) 12 h3?! is too slow. TimmanIvanchuk, Hilversum (1) 1993 showed to good effect Black's regrouping plan after 12 ... R.e8 13 iLld2 R.f8 14 R.e2 iLle7 IS h4?! (after IS gS iLld7 16 h4 bS Ivanchuk cited 17 iLlb3 "d8 18 cxbS iLlb6, with Black having good attacking chances as he uncoils; he also gave 17 cxbS l:.xc3!? 18 bxc3iLlg6 - it

The New Main Line used to be Petrosian who had the reputation of preparing for a decisive attack against the king by retreating everything to the back rank!) IS ... bS 16 gS li)e4! and the white king soon came under serious attack. 2) 12 h4?! looks ferocious, but White's attack failed to connect after 12 ... bS 13 cxbSl:iJe7 14 ~bl (14 gS? J..xa3 IS gxf6? lhc3) 14... J..xa3 IS 'ii'a4 J..b4 16 'ifxaS J..xaS 17 J..d6 J..xc3! 18 J..xe7 (18 bxc3 li)e4 19 J..xe7li)xc3+ 20 ~c2li)e4+ 21 ~b2 li)xf2 followed by ... li)xhl and ...l:ab8 +) 18 ... li)xg4 19 bxc3 li)xf2 + in Shabalov-Tisdall, Oslo Cup 1991. 3) 12lt)d2 has also been tried, but it is not clear that the knight should abandon the kingside. 12 ....tf8 13 lLlb3 'irb6 14 gS lLlhS IS .tg3 g6 16 J..e2li)aS 17li)xaS 'irxaS 18 e4 d4 19 l:xd4 'irxgS+!1= S.Farago-R.Stem, Budapest 1991. After the text-move, play branches into: AI: 12••. J..f8 79 A2: 12... bS!? 80 A3: 12•••dxc4!? 81

A1) 12.••.tfS (D)

79

13gS 13li)gS is also promising, although again it raises the question of why White ever played g4. I) 13 ...1:iJe714h4g6IShSJ..g716 .td6 'ird8 (16...lt)c6 17 hxg6 hxg618 cxdS exdS 19 J..d3 ± Khalifman) 17 hxg6 hxg6 18 .teSlt)c6 19 J..xf6 'irxf6 20 f4 li)aS, and instead of the miscalculated 21li)ce4?? 'ire7 22 l:h7 dxe4 23 'irh2 f6 -+, Khalifman-Ki.Georgiev, Manila IZ 1990, White could have played 21 cxdS .ta4 22 'irxa4 l:xc3 23 'ird4!; (Khalifman). 2) 13 ... g614h4.tg7IShSJ..e816 hxg6 hxg6 17 0, and now: 2a) Black should try 17 ...eS 18 J..g3 d4 19 exd4 li)xd4 20 'iWh2 .tc6 21.txeSli)xf3 22li)xf3 .txO 23 ~ .txdS 24 cxdS ;1;, or maybe ±. 2b) The game Vera-Gild.Garcia, Matanzas 1992, finished spectacularly after 17 ... bS? (too late) 18 cxbSl:iJe7 19 J..eS d4 20 .txd4lt)edS 21li)xdS! l:xc2 22lLlxf6+ ~f8 23 li)gh7+ ~e7 24 J..cS+ 1-0, in view of 24 ...l:xcS 2S li)g8#. A checkmating position to dream of. 2c) 17...dxc4?! is a more recent try, but probably no improvement. S.Pedersen-Parker, London 1997 (reached via II h4; see line F below) continued 18 'irh2 eS 19 J..g3 l:d8, with unclear complications after 20 l:c 1 WcS 21 li)ce4lLlxe4 22 ~e4 'irxe3. However, Black's kingside would have been highly vulnerable after 20 l:xd8! followed by .txc4. 13•••li)h5 14 .tg3lt)e7 14... g6 prepares an imaginative piece sacrifice: 1) IS cxdS .tg7 16 dxc6 .i.xc6 17 J..e2 .i.xc3 18 bxc3 li)xg3 19 hxg3

80

The Queen's Gambit Declined: 5 j"f4!

.i.a4 20 'iWb2 and now in DinstuhlLengyel, Budapest 1994 Black played 20....i.xdl? 21 .i.xdl :xc3 22 lDd2 l:acs 23 lDe4! l:cl + 24 Wa2 ±. Better would have been 20...:xc3 21 lDd2 (21 :d3 .i.c2+ 22 'iWxc2 :xc2 23 Wxc2 :cS+ +) 21.. ..i.c2+! 22 Wa2 .i.xdl 23 :xdl (23 .i.xdl 'iWdS+!, explaining why Black wanted to force the king to a2) 23 ...:acS 24 lDc4 :3xc4 2S .i.xc4 :xc4 26 'ilxb7 =, but not 26 'iff6?! :c8 27 :d7 :c2+ 2S Wbl'iWbS++. 2) However, White does not have to accept the bait. IS .i.h3! ± places the black centre under enormous pressure, since on IS ...lDe716 cxdslDxdS there follows 17 :xdS!. 15 .i.e2 .i.e8 16 .i.d6 16 It)eS f6 17 gxf6 gxf6 IS lDf3 .i.g6 (IS ...dxc4 19lDd4 .i.g6 20 e4 eS 21 lDfS lDxfS 22 exfS .i.f7 23 :d7 with an attack - Beliavsky) 19 e4 Gelfand-Beliavsky, Linares 1990, and now 19... fS! looks good for Black. White inverted moves IS and 16 in this game. 16 lDd4 lDxg3 17 hxg3 g6 18 :h4 00 (Georgadze and AzmaiparashviIi). 16•••g6 16...dxc4 17lDeS g6 18 .i.xhS gxbS 19 lDe4 lDg6 20 lDf6+ WhS 21lDxeS! lDxeS 22 .i.xeS+ 'iWxeS 23lDf6 'iWfS 24 'fi'xfS exfS 2S l:d7 gives White a clear advantage. 17lDeS "d8 17 ...lt)g7 18lDg4 ±. 18 .i.xe7 "xe7 19 cxdS 19 .i.xh5? 'iWxg5. 19...exdS 20 f4 ± G.GeorgadzeSturua, TbiIisi 1981 (notes from move 16 by Georgadze and AzmaiparashviIi). Here Black's defensive structure proved too passive.

A2) 12...bS!? (D)

Enthusiastically awarded a double exclamation mark by Mikhalchishin in Informator 51, but we make a more modest assessment here. The kingside barricades can wait until White has started attacking there; meanwhile there is work to be done on the queenside. 13 cxbS! 1) 13 cxdS b4! 14 lDa4 (14 dxc6 .i.xc6 IS axb4 .i.xb4 16 .i.e2 .i.xc3 17 bxc3 .i.e4 -+ Ftacnik) 14... .i.f8 15 dxc6 .i.xc6 -+ Mikhalchishin. 2) 13 lDxb5 is well met by either 13 ... lDe7 + Ftaenik, or 13 ... a6 141Dc3 .i.xa3 IS bxa3 'iWxa3 16 "b2 :ab8 17 .i.xb8 :xb8 18 lDb5 'iWaS + Mikhalchishin. 3) 13 g5 It)bs 14 cxb5 lDxf4 15 exf4 (IS 'iWa4 d4!; 15 bxc6 .i.xc6 16 exf4 d4 17 lDe4 .i.xa3) 15 ...lDe7 16 lDe5 .i.e8 + Mikhalchishin. 13•••lDe7 14 .i.e5 1) 14 lDd2 'iWd8 (14 ....i.xa3?? 15 lDb3) IslDb3 (15 gSlDhS followed by ...lDg6) 15 ...lDe4! 16lDxc5 (16lDxe4 .i.xe3 +; 16 .i.e5 .i.d6! 17 .i.xd6lDxd6 ~ Mikhalchishin) 16...:xcS 17 .i.e5

The New Main Line ~xc3+ 18 j,xc3 j,xbS ; GelfandBeliavsky, Linares 1991. Black has regained his pawn, and has achieved more on the queenside than White has on the kingside. 2) 14 Wd2 j,xa3 15 ~xdS "a4 16 ~xf6+ gxf617 ~d4 j,b4 18 b3"aS 19 "a2 "xa2+ 20 ~xa2 eS + VeraGild.Garcia, Matanzas 1994. 3) 14 :cl, Mikhalchishin, awaits exploration.

14...~xg4 15 j,xg7 ~e3?! Following IS ...~xg7 16 :gl, 16.. .fS 17 h3 hS 18 hxg4 hxg4 19 ~S j,e8 20 ~xg4!? gives White a promising attack according to Kishnev. However, this is far from exhaustive, and Black may well be able to seek improvements, for example 16 ...hS 17 h3 a6!? 18 hxg4 h4, when Black's counterattacking possibilities should not be dismissed lightly. 16fxe3~xg7

Pawns are now level, and Black's king is exposed. However, Black's bishop-pair and much the better pawn structure give him hope. Now, Kishnev-Barsov, Zwolle 1993 continued 17 h4j,xe3 18 :d3 (18 "g2+~fSI9 ~gS CD Kishnev) 18 .....b6 19 "g2+ 'ifi!fS 20 ~S?! (20 ~gS is better Kishnev) 20 ... j,e8 21 ~g4 :xc3! 22 ':xc3 j,d4 ;. The exchange sacrifice kills off White's attack, and leaves Black the chance to make something of his bishop-pair and three connected passed pawns. White can improve upon this though.

17e4! Not mentioned by Kishnev, this breaks up Black's pawn centre, making it easier for White to use the central squares as staging posts for the

81

kingside attack. After, for example, 17... j,xa3 18 exdS ~xdS? (18 ... exdS ±) 19 .g2+ WfS (19...~h8?? 20 :gl ~xc3+ 21 bxc3 +-) 20 ~xdS exdS 21 ~gS both kings are exposed, but White's attack is faster. White is better here, but earlier on IS ...~xg7 seems fine for Black. A3)

12.••dxc4!? 13 lOcl2 liJe7 13 ... bS may be compared with the position reached after 11. .. dxc4 12 lbd2 bS; the inserted ~bl and ...:fc8 slightly surprisingly help White. Thus 14 ~e4 b4 IS ~xf6+ gxf6 16 :xd7 bxc3 17 j,xc4;t cxb2? 18 :xf7. 14~xc4.d8

Black's play might give a slightly passive impression, but his position is resilient, and he will soon be able to get moving on the queenside. Khalifman-Kotronias, Sochi 1989 continued 15 e4 ~g6 16 j,g3 .e7 17 h4 hS 18 gS~81geS.i.c620:g1 bS21lM6,

and now 21.. ..i.xd6 22 exd6 "b7!? might be worth considering. B)

11 ~bl! (D)

B

82

The Queen's Gambit Declined: 5 i,.f4!

It is questionable whether White actually achieves anything after an early g4, while ~bl, to remove the king from the perilous c-file, is usually necessary sooner or later. It is more flexible to play it immediately, and then play according to Black's reply. 1l...dxc4 1) 11...l:tfc812cxdS ~713~S! lDexdS (13 ... exdS 14lDxd7lDxd7 IS 'iVa4 ± Akopian-Gild.Garcia, New York 1994) 14lDxd7lDxd7 IslDxdS exdS 16 .b3 a6 17 .ie2lDb6 18 l:td3 .ie7 19 .ieS ':'c6 20 .io .if6 21 .ixf6 ':'xf6 22 l:thdl g6 23 1!fc3, Delemarre-Van der Sterren, Antwerp 1995, and the draw agreement had more to do with the 300-point rating difference between the players than with the position, which clearly favours White. 2) 11.. ..ie7 screams out for a hack attack with 12 lDgS, one point being that tactics based on l:txd7 make it very difficult for Black to capture with his d-pawn. After 12 ...h6 13 h4, White is clearly better. Another possibility is 12 g4 dxc4 (12...eS?! leaves Black with an inferior version of the to....ie7 11 g4 line, for example 13 gS .ig4 14 .ie2 lDe4 IS lDxe4 dxe4 16 lDxeS .ixe2 17 "xe2 l:tfd8 {17 ... lDxeS? 18 l:tdS} 18 lDxc6 bxc6 19 h4 "fS, with White a clear pawn up, but facing considerable technical difficulties, KallaiBorocz, Budapest 1995) 13 .ixc4 l:tfd8 14 gS lDh5 IS .id6 .ie8 16 .ixe7 lDxe7 17 .id3 g6 18 .ie4 ~ Beliavsky-Stohl, Bled 1996. 12 .ixc4 .ie' 12...l:tfc8 13 ~4! lDxe4 14 l:txd7 ±Akopian. 13lDgS l:tfd8 14 h4

14lDds exdS IS l:txdS "b6 16 l:tbS lDd4 17 l:txb6lDxc2 18 l:txb7 .ifS 19 ~a2 ~f8 00 Akopian. However, the alert reader will have noted that the positions reached are the same as in Korotylev-Kharitonov, Moscow 1996 (section 3.2, Line A31), except that h4 and ... h6 have been omitted. By analogy, this would suggest for White 16 ~xf7 (instead of 16 l:tbS) 16... lDxdS (16...~f8 17 l:tbslDd4 18 l:txb6lDxc2 19 lDxd8! axb6 20 lDe6+ .ixe6 21 .ixe6 ±) 17 .ixdS ~f8. Now 18 ~gS, as analysed in the Korotylev-Kharitonov game, is harmless with no pawn on h4 to back up the knight. Similarly, 18 "xh7 .ifS+!, Burgess, favours Black. There is a third way, 18 lDd6!?, but even here Black can defend: 18....if6! (18 ....ifS? 19 "xfS+ .if6 is less accurate due to 20 b3! +-) 19lDc4 'iVbs 20 l:tdl .ig4 and the attack collapses. So probably 14lDds has to be abandoned, although there may still be room to look for new ideas - maybe 18 l:tdl!? 14•••lDeS! 14...l:tac8? ISlDce4lDxe4 16.xe4 g6 17 l:txd7 l:txd7 18 .ixe6! +- Akopian . 14... h6 directly transposes to the Korotylev-Kharitonov game. IS.ixeS! Black has covered all the tactics on dS. Akopian notes ISlDdS? exdS 16 .ixeS (16 l:txdS? .ia4) 16...dxc4 17 .ixf6 .ifS 18 e4 l:txdl+ 19 l:txdl gxf6. Also, IS ~ce4? lDxc4 16lDxf6+ .ixf6. 15.•.•xeS 16 f4! This pawn sacrifice is necessary to maintain the initiative. Again the immediate tactics don't work: 16lDd5?

The New Main Line exd5 17 ':'xd5 'fIc7 IS lDxf7 b5! 19 lDxdS 'fIxc4 -+ (Akopian). 16.....CS!

The sacrifice is best declined. Black got poleaxed in Akopian-Gild.Garcia, Philadelphia 1994, after 16 ...'ii'xe3?! 17 f5 ':'acS IS ':'hel 'fIg3 19 fxe6 ':'xc4 20 exf7+ ~fS 21 'fIe2 +-. 17.i.b3(D)

As recommended by Akopian, who notes that after 17 lDce4 lDxe4 IS 'ifxe4 .i.xg5 19 hxg5 g6 20 ':'c I 'fIf5! Black is holding comfortably.

17.••.i.e8! 17 ...h6 IS e4 hxg5 19 hxg5 and now: I) 19...lDeS 20 e5 g6 21 ':'xd7 ':'xd7 22 .i.xe6 fxe6 23 'fIxg6+ lDg7 24 'fIh7+ ~f7 25 g6+ ~eS 26 'fIxg7 .i.fS 27 'fIf6 'fIc4 2S ':'hS 'fin + 29 ~c2, and White's king soon escapes the checks, leaving Black in a desperate position. Akopian gives 19 ... lDeS as the "only move"; as we shall see, there are alternatives, but nothing effective. 2) 19... g6 20 e5! (20 gxf6 .i.xf6 00) 20... lDh7 21 ':'xd7 l:r.xd7 22.i.xe6lDfS 23 .i.xd7lDxd7 24 'fId3!!. The sting in the tail of the combination; the threat

83

of'flh3 gives Black no time to save his knight. 3) 19 ...lDh7 20 ':'xh7 ~xh7 21 ':'hl+ ~gS (21...~g6 22 e5+ f5 23 exf6+ 'fIf5 {23 ...~f7 24 g6+J 24 g4! 'fIxc2+ 25 .i.xc2+ ~f7 26 g6+ ~f6 27 g5#) 22 e5 g6 23 "d3! ~fS (23 ....i.fS 24 lDe4 +-) 24 'fIh3 ~eS 25 lDe4 'iVb6 26 'fIhS+.i.fS 27&Dd6+ 'fIxd6 2S exd6 and White wins, e.g. 2S ....i.c6 29 'figS ':'xd6 30 ':'hS ~e7 31 ':'h7 with a decisive advantage. 18 e4 eS 19 ':'dO! Akopian suggests 19 g3 with a slight edge, but a resolute exchanging policy should keep Black in the game, for example 19...':'xdl+20':'xdl ':'dS 21 ':'xdS .i.xdS 22 'fId3 'fId4 23 'fIxd4 exd4 24 &Dd5 h6 25 lDxf6+ .i.xf6 26 lDf3 .i.c6 27 .i.d5 .i.xd5 2S exd5 ~fS =. Black's bishop-pair provides good insurance for an endgame. With 19 ':'dn!, White argues that his pressure is greatest if he can keep pieces on, and that it does not matter if the d-file is ceded, so long as White can build up pressure elsewhere. White's plan is to roll up the centre with fxe5, ':'f5 and e5, and it is difficult to see how Black defends, since 19... h6 is met simply by 20 fxe5. Given that the tactics with 14lDd5 do not quite work out, Akopian's 14 h4 seems critical.

C, 11 adS (D) This is a standard plan against ....i.d7 in positions where White has not castled queenside, but the exposed position of the white king makes it more hazardous here. 11•••lDxdS 12lDgS?!

84

The Queen's Gambit Declined: 5 i..f4! gathering of Black's pieces in front of the king.

B

0) ll.i.gS (D)

B

12 ~xd5 exd5 13 ~bl might still leave White better. 12•••g6 13 ~ge4 Lputian gives 13 h4?! :ac8, when Black's attack is favourite to come in fIrst. 13•••.i.e7 Piketdescribes 13 ... ~xc3 14.xc3 .xc3+ 15 ~xc3 e5 as unclear, but the chances must be with White after 16 :Xd7 exf4 17 ~4. 14.i.c4 14 ~xd5?! exd5 15 ~c3 .i.f5 16 .a4:acS~; 14~bl :fdSaoLputian. 14•••~b6! 14...~xf4? 15 :xd7 ~xg2 16.i.b5 ± Piket-Timman, Amsterdam 1995. 14...~c315.xc3 'ii'xc3+ 16~c3 e5 17 .i.h6 :fdS IS ~ ;I;; Piket. IS.i.b3 15 .i.c?? ~5 16 .i.b3 :fcS 17 .i.d6 (17 .i.xb6 axb6 IS f4 .i.c6 19 fxe5 .i.xe4 20 .xe4 .i.xa3 wins for Black, Lputian) 17....i.xd6 IS ~xd6 ~a4! =1= Azmaiparashvili-Lputian, USSR 1991. White's bishop excursion used up too much time. IS •••~ This position is assessed as unclear by Lputian, but if I were White, I would be deeply worried about the

The most direct attempt to win the battle of the d-file, but it uses up a tempo which might otherwise have been spent securing the king's position. 1l•.•.i.e7 12 adS 12 .i.xf6? .i.xf6 13 cxd5 ~e7! 14 .i.c4 :acS 15 dxe6 .i.xe6 16 .i.xe6 fxe6 17 :d3 ~5 =1= Klinger-Bonsch, Linz 1990. A memorable pawn sacrifIce justifIed by White's front-of-king weaknesses. 12 .i.d3 h6 13 .i.h4 :acS 14 ~bl a6?! (mistimed; 14...:tdS! =) 15 cxd5 ~xd5 16 ~xdS .i.xh4?? (16...exd5 ;1;;) 17 ~c3 .i.f6 IS .i.h7+ ~hS 19 :xd7 +- Kobrin-Kunin, Petach Kivka 1997. 12...~xdS 13 ~xdS exdS 14 ~bl 14 .i.d3 :fcS!. 14•••l:ac8 15 .i.xe7 ~e7 16 Wd2 Given as equal by Bonsch, but White can make Black suffer a long time before agreeing to the draw. White has defInite chances of pressure against the isolated queen's pawn.

The New Main Line This line deserves further testing, but our next line suggests that .i.g5 might profitably be delayed a move.

85

B

E)

11lLld2! Not often played, but probably strong. Black is immediately forced to work out how to disentangle his bishops. 11...lLle7 11.. ..i.e7 leads to positions discussed under 1O... .i.e7 11 lLld2 .i.d7 (3.4, Line D). There the view was expressed that 11.. ..i.d7 was not the most accurate, the other bishop belonging on fS rather than e7. The recommended line was 12 g4 l:tfcS 13 .i.e2 ~. 12.i.g5! Damaging Black's pawn structure. 12•••:t'ca 12 .....c7 13 .i.xf6 gxf6 14 cxd5 exd5 (14 ...l:tacS 15 dxe6! fxe6 16 lLlde4 ±) 15lLlb3 .i.b6 16lLld4 ~. 13 .i.xf6 gxf6 14 lLlde4! dxe4 15 l:txd7 f5 ± Alterman-Jinrong Liang, Beijing 1995. White is better on the kingside, has taken control of the open central file, and has an extra queenside pawn, having already neutralized Black's play there. F)

11 h4!? (D) Yet another promising try, but with only a solitary appearance on the database. There are several transpositional possibilities. 11•••:t'ca 11...dxc4?! 12 lLld2 b5 13 lLlde4 leaves both Black's bishops hanging.

12 «i>bl .i.e8 12...lLle7 13 b4 ±. 13lLlg5 g6 13 ... h6 14 cxd5 exd5 15 lLlxd5 hxg5 16 hxg5 lLlxdS 17 l:txdS lLle7 IS 1Ifh7+ «i>f8 19 1IfhS+ lLlg8 20 l:th7 wins for White. 14 hS .i.f8 15 bxg6 bxg6 16 g4 .i.g717 f3 S.Pedersen-Parker, London 1997. We have reached positions discussed earlier (section 3.6, Line AI) under the move-order 10....i.d7 11 g4 l:tfc8 12 «i>bl .i.f8 13 lLlg5 g6 14 h4 .i.g7 15 h5 .i.e8 16 hxg6 hxg6 17 f3, where the next few moves are discussed. White has a promising attack. In conclusion, 1O... .i.d7 is no problem for White, who has a wide choice of advantageous replies. 11 «i>bl! and IllLld2! both seem good for White, while 11 h4!? is promising. and even 11 .i.g5 keeps the initiative on a modest scale.

3.710... a6 1 d4 dS 2 c4 e6 3lLlc3lLlf6 4lLlf3 J.e7 5 J.f4 0-0 6 e3 eS 7 dxeS .i.xeS 8 .. lLlc6 9 a3"aS 100-0-0 a6 (D)

a

The Queen's Gambit Declined: 5 i.,f4!

86

better for White in M.Gurevich-Silva, Bern 1989. Gurevich assesses 14...h6 as unclear, but one feels that White ought to be slightly better, maybe by IS i.e2!?, planning to meet IS ...i.g4? with 16lilgs.

w

w Black has ideas of getting on with his queenside counterplay without, as in the 1O... £d7line, leaving a hanging bishop. The problem is that this move does not actually help get his pieces out, so if White can neutralize the plan of ... bS he stands well. Again an immediate response on the queenside is indicated: A: 11 adS 86 B: 11 /t)d2 87 First we should note that 11 £gS?! gave Black no problems in the game Pardoen-Solomon, Brisbane 1995. After 11...dxc4 12 £xf6 gxf6 13 £xc4 bS 14 £dS (a flashy but ineffective move) 14 ... £b7! IS 'iFe4 £xa3 16 'iFg4+ ~h8 17 tbe4 exdS 18 'iffS £e7 White's desperate attack had burnt itself out; if 19lilxf6 then 19 ...'ifal + 20 ~c2lilb4+

-+.

A) 11 adS So that ... bS will attack precisely nothing! 11 •••exdS (D) After 11 ...lilxdS 12 lilxdS exdS 13 ~bl £e7 14 b4, 14...£g4?! IS lilgS g6 16 f3 £fS 17 £d3 was clearly

12i.gS There is no real consensus on this position, but it makes sense to try to damage Black's pawns. Alternatively: 1) 12 ~bl i.e6 13 i.d3?! (13lilgS transposes to '2'; 13 b4? lilxb4 14 axb4 i.xb4 IS lila2 :ac8 16 "'b2 i.a3 17 'ifd2 i.fS+ 18 ~al :c2! 19 "'xaS i.b2+ 20 ~bl :d2+ and mate follows, Ubilava; 13 'iFa4 "'xa4 14 lilxa4 i.a7 IS i.d3 d4 16:c 1 dxe3 17 fxe3, Ubilava) 13 ...b8 14lilgS h6 IS lilxe6 fxe6 16 'iFb3 bS + ConquestUbilava, Tbilisi 1988. 2) 12 lilgS i.e6 13 ~bl :ac8 14 :xdS!? (14 i.d3?! transposes into Conquest-Ubilava) 14 ...lile7 (but not 14... i.xdS? IS lilxdS "'el+ 16 ~a2 +-) IS :eS lilg6 16 :xe6 fxe6 17 i.c4lilxf4 18 exf4 ~h8 19lilce4 "'c7 20 i.xe6 i.xa3 21 "'xc7 :xc7 22 bxa3 lilxe4 23 lilxe4 :xf4 ~ Tisdall0gaard, Norwegian Ch 1995. White has a slight material advantage, but

The New Main Line not enough to create anything workable, given Black's active rooks. 12.•• j.,e613 j.,xf6 gxf614 ~bl An unnecessary precaution, with the black rook not yet on cS. 14llXl4!? ;1;. 14•• JUc8 15 'iWa4 'iWxa4 16 lLlxa4 j.,a7 17 lLlc3 :d8 18 lLld4 lLlxd4 19 exd4 ~f8 Malaniuk-Ziatdinov. Kusadasi 1990.

=

B) 11~2

Thematic, and again quite promising. 11••• j.,e7 12lLlb3 I) 12 g4!? also looks strong: 12...eS 13 gS exf4 14 gxf6 (14lLlb3?! "d8 IS gxf6 j.,xf6 16lDxdS fxe3 :j: ErikssonSanden, Helsingborg 1991) 14... ..ixf6 IS lDxdS ..ieS (IS ... fxe3 16 lDxf6+ gxf6 17lDe4 ±; IS ... j.,dS?? 16lDb3 'ifa4 17lDc3 +-) 16lLlb3 "dS 17 exf4 j.,bS IS ~bl ± Hjartarson-Thorsteinsson, Reykjavik 1997. Black could consider instead 12... bS!? 13 gSlDhS 14 cxdS lLlxf4, but IS dxc6! will force him to spend a lot of time regaining the pawn. 2) 12 ..ie2 e5?! (12 ... b5! is more thematic) 13 j.,g5 d4 14lDb3 "b6 IS exd4 exd4 16 ..ixf6 dxc3 17 j.,xc3 "xf2 IS lIhfl ± Astrom-Nordstrom, Stockholm 1995. 12•••"b613 c5 'ii'd8 14 e4 d4 15 e5 lDh5 16 ..ig3 "c7 17 lLla4lLlxe5 18 l:[xd4 lDxg3 19 bxg3 h6 20 j.,e2 ± Tabatadze-Kuzmina, Moscow 1991. White has the upper hand on the queenside, and is about to roll pawns on the kingside. So 10... a6 11lDd2 is looking good for White.

87

3.8 10.. Jld871 1 d4 d5 2 c4 e6 3 M till'6 4lDf3 i.e7 5 j.,r4 0-0 6 e3 c5 7 dxc5 j.,xc5 8 'iWc2 lDc6 9 a3 "a510 0-0-0 lId8?! This soon gets Black in a tangle. 11lLlb5! I1lDd2 dxc4 12lDxc4 lIxdl+ 13 "xdl "dS 14 "xdS+ lDxdS IS i.e2;1; Gelfand-Ki.Georgiev, Novi Sad OL 1990. 11••• b6 l1...dxc4 12 ..ixc4 ±. What is Black going to do with her queen? 12lDes j.,b7 13lDxc6 i.xc6 14 b4 ± Miles-Amura, Andorra 1994.

3.910...lDe4 1 d4 d5 2 c4 e6 3lLlc3lDr6 4lLlf3 ..ie7 5 j.,f4 0-0 6 e3 c5 7 dxc5 j.,xc5 8 'iWc2 lLlc69 a3 "a510 0-0-0 lDe4 (D)

The sharpest move on the board, but more or less abandoned by Black since Kasparov won a miniature against Vaganian in 1992, playing the equally sharp 11lLlb5. The good news for Black is that Kasparov's published analysis on 11lLlbS, suggesting a big advantage for White, is unconvincing;

88

The Queen's Gambit Declined: 5 ij4!

at one point he even misses a simple back-rank check. The bad news for Black, and thus the good news for White, is that White can safely snatch the e4-pawn. We examine the simpler and stronger move first: A: 11~e4! 88 B: 11 ~bS 89

Play may continue: 14 ~eS ~eS IS .txeS .txe3+ 15....txa3 16 J:[d5 'ifel+ 17 W'dl. 16 fxe3 "xeS 17 "d2:es 18.tel Black could get a genuine initiative if White tried to hold on to the pawn with 18 J:[el. 18.....xe3 19 "xe3 J:[xe3 20 J:[d8+ ~f721.tf3±

A)

11 ~xe4! dxe4 Now: AI: 12 'ifxe4! 88 88 A2: 12 .!Dd2 A1)

12 'ifxe4! (D)

B

Black can offer the exchange of rooks with 21...J:[e8, but his queenside will still be very troublesome to disentangle. I find it difficult to see how Black is supposed to improve in all this. The fact that grandmaster theory has tended to concentrate on the complicated and obscure II ~b5, rather than the simple and strong II ~xe4, is perhaps indicative that the spirit of oldfashioned romanticism is not yet dead. Sometimes though the Steinitzian accumulation of small advantages is more to the point; and a whole pawn is quite large in the scale of small advantages. A2) 12~(D)

Less direct and less menacing. 12•••fS A suggestion in Informator 54 (allegedly by Yusupov) gives 12 ... .txa3 "! with an attack". Yes, but after 13 bxa3 'ii'xa3+ 14 ~d2, White has an extra piece and the better development, and it is questionable whether Black's attack has any legs after, e.g., 14...:<18+ 15 .td3, 14...'iVh4+ 15 ~e2, 14..:ifb2+ 15 ~el! or 14...e5 15 .txcS. 13 'ifc2 eS With compensation for the pawn, according to Valdes. I just don't see it.

B

The New Main Line 12.••eS 12 ...fS? 13 lOb3 Wb6 14 lOxcs "xcS 15 .ld6 wins the exchange, so Black is obliged to weaken dS. 13.lg3 An Informator reference gives 13 lOb3 Wb6 00 S.Spasov-Sergiev, COIT. 1991, but with no follow-up. It seems sensible to delay the knight move, to keep the pressure on e4. 13•••.lg4! An important zwischenzug. 13 ...fS?! 14 ~b3 Wb61S :td5! .le716 .lxe5;t 14 :tel rs 15 f3 exf3 16 gxf3 .lhS If anyone is better here, it is Black. B)

11 lObS (D)

89

12...:ta7? 13 cxdS ± Valdes-Otano, Cuba 1991, the point being that for once 13 ...exdS 14 :txdS is possible for White, without having to worry about ....lxe3+. Now we have an important branching point, with White having the choice of: Bl: 13lbdS 89 B2: 13lbxas 91 B3: 13tOxdS 91 Three minor alternatives: 1) 13 cxdS? Wxc7 14 Wxe4 exf4 15 dxc6 fxe3 =+= Yusupov. 2) 13 .lxeS lOxeS 14 lOxaSlOg4 =+= Ftacnik. 3) 13lOxeS Wxc7 14lOxc6 Wxc6 15 cxdS We8 16 f3 .lfS is unclear according to Yusupov, but Black is surely much better after 17 fxe4 :tc8.

8

Let us now take a look at Kasparov's attempted refutation. B1)

13lbdS (D)

If this is genuinely strong, then fine; if not, 11 ~xe4 should be preferred. 11•••a6 11...eS 12 cxdS exf4 (12 ....lfS 13 ~h4) 13 Wxe4 tOe7 14 b4! 'ifa4 15 Wc2 +- Valdes. 12lOc7 12 .lc??! b6 13 ~bd4 (Yusupov) 13 ....lxd4, followed by ...:ta7 at the fIrst opportunity, is fine for Black. 12..•eS!

Four pieces are en prise, but they can't all be taken at once! As might be expected, there is a bewildering array

90

The Queen's Gambit Declined: 5 /£.f4!

of alternatives, some of which leave various pieces hanging for a few moves. The complicated lines are the most fun, and also the most difficult to analyse, but Black's best shot may well be a simple retreat. Maybe there is a moral in all this? 13••• t?Jf6! This move retreats one of the attacked pieces to a safe square, and opens up an attack on another piece. As such, it is certainly a logical try. Alternatives: 1) I3 ... f5?! 14 ':xe5! ± t?Jxe5 15 i.xeS ':a7 (lS ...':b8 is better - Kasparov) 16 /DdS! (so that if 16... i.d6, 17 b4! "'xa3+ 18 i.b2 and the queen soon gets trapped) 16... b6 17 i.d3 i.d7 18 b4 "'xa3+ 19 i.b2 "'a4 (19 ......a2 20 t?Jc3! +- Kasparov) 20 bxcS bxc5 21 t?Je5 Wxc2+ 22 i.xc2 i.e6 23 /Df4 1-0 Kasparov-Vaganian, Debrecen Echt 1992. 2) I3 ...'ii'xc7 14 'ii'xe4 and now afterl4...exf4 15 ':xc5 fxe3 16 fxe3!? ±, one doubts Yusupov's claim that Black has compensation for the pawn. Black may consider instead 14...f5, but after 15 i.xe5 (15 'iVc2?! exf416 ':xc5 fxe3 17 fxe3 'ii'e7) 15......e7 16 'ii'h4! Black's hopes of effective gambit play will be disappointed. 3) 13 ...i.e6?! 14 ':xe5 (an original method of destroying a pawn centre!) 14... t?Jxe5 15 i.xe5 i.f5 16 i.d3 +Kasparov. 4) 13 ...t?Jxf2 and now: 4a) Kasparov recommends the continuation 14 t?Jg5(??) i.f5 IS 'ii'xf5 in lnformator 56, totally missing the winning 14......el+. Even world champions can sometimes miss simple back-rankers!

4b) 14 1i'xt2 "'xc7 15 t?Jxe5 t?Jxe5 16 ':xe5 (16 i.xe5 Wc6 17 "'g3 f6 18 lIxc5 'ii'xc5 19 i.d6 Wc6 20 i.xfS 'itxfS 21 i.d3 i.e6 22 "'f3, Burgess, may also be good) 16...i.e6 (16...i.d6?! 17 ':d5) and now: 4bl) White has the opportunity to over-combine with 17 ':gS "'c6 18 ':xg7+? 'itxg7 19 i.h6+ 'itg6!! 20 'ii'g3+ 'itxh6 21 "'h4+ 'itg6 22 'ii'g3+ i.g4!. 4b2) 17 "'g3!? is another try. After 17...i.d6 Nunn points out 18 i.d3 !? with the spectacular possibility 18... i.xc419 i.xh7+'ith8 20b4!!, the main point being 20 ...i.d3+ 21 lIc5. 18 ... ':fd8! keeps Black in the game though. Another approach is that mentioned by Burgess: 18 ':g5 i.xf4 19 ':xg7+ 'ith8 20 exf4. However, after 20... i.xc4 Black's position is unexpectedly resilient: 21 "'c3 "'xf4+ 22 'itbl f6 is unclear, while after 21 ~bl ':ad8 22 i.xc4 'ii'xc4 the tangle of white pieces on the kingside gives Black the chance to create counterplay. 5) 13 ...i.f5 14 t?Jxa8 t?Jg3 appears vastly complicated at first, but 15 e4!! (Burgess; Kasparov's 15 'iVd2 is less clear-cut) 15 ... i.xe416i.d2! traps the black queen and forces liquidation into a queenless position where White is material up. Some of the key positions look as though they have been reached by enthusiastic primary school children, the pieces being apparently so randomly placed. But evidence of the design of strong players is shown by the way that the queen is snared.

14':xe5 14 t?Jxa8? exf4leaves Black with a clear advantage.

The New Main Line Kasparov's Informator analysis stops here, with the strong implication that White is better. This can be disputed. 14•••lLlxeS IS.ixeS IS lLlxa8 lLlxf3 16 gxf3 "el + 17 "dl "xf2 is at the very least comfortable for Black, for example 18 "e2 "xe2 19 .ixe2 lLlhS. IS...':a7 16.ixf6 16lLlgS? .ixe3+. 16.....xc7 16... gxf6 17lLlds ;to 17 lLlgS g6 18 lLlxh7?! This combinative try is the only attempt to keep the initiative going; on other moves Black is fine. The idea is that now 18 ...~xh7? is met by 19 Black can do better though. 18....ifS! 19 .id3 ~xb7 20 .ixfS "c6! Winning a piece. Kasparov's 13 ':xdS gives rise to massive complications, but does not seem to refute Black's play. So we tum now to other tries.

"e4.

82) 13lLlxa8 On the 'grab the material and run' principle. 13...exf4 14 adS lLlxf2 As given by Ftacnik. Black is doing well after IS dxc6 .ixe3+ 16 ~bl lLlxhl. Also to be considered is 14...lLle7 (Valdes) IS "xe4 .ifS 16 "xf4 .ixa3 17 :d2 ':c8+ 18 ~dl .ixb2, Yusupov, with a strong attack. 83) 13lLlxdS (D) The safest option.

91

13...lLlxf2 14lLlgS After 14 "xf2 exf4 IS lLlxf4 ':e8 16lLld4.ig4 17 .ie2 .ixe2 18lLlfxe2 lLleS (Yusupov) White's extra pawn probably counts for less than Black's extra piece-activity. 14....ifS! 15 "xf2 IS "xfS g6 (the knight blocks off h3!) 16lLlf6+ ~h8 17 W'd7, and now either 17 ....ixa3 or 17 ....ie7 - Yusupov. IS...exf4 16 "xf4liJe717lLlxe7+ 17 .id3 lLlxdS 18 "xfS g6 19 'ii'xdS .ixe3+ :;: Yusupov. 17•••.ixe7 18 ':dS! "el+ 19 ':dl 20 ':dS 9e1 + 21 ':dl lh· 1h Gelfand-Yusupov, Linares 1992.

"as

This was an exhausting section to write, but the conclusion is clear enough. 1O... lLle4 is objectively best met with 11lLlxe4 dxe4 12 "xe4.

3.10 General conclusion on 100-0-0 White is better. 10 O-O-O!, suggested by Kaidanov, must be regarded as one of the most important opening innovations of the 1980s, opening up a whole

92

The Queen's Gambit Declined: 5 J.f4!

new field of research for opening theory, in a system which had previously been assumed to be veering towards equality. To summarize the key variations after 10 0-0-0: 1O... ~e7 11 h4!;t (11 g4 00; 11lLld2 00); 1O...dxc4 11 ~xc4 a6 12lLlg5 ;t; 10... ~d7 11 ~bl ;t (11lLld2 ;t; 11 g4 00); 10... a6 II lLld2 ;t or 11 cxd5 exd5 12 ~g5;t; 1O... lLle4 I1lLlxe4 dxe4 12 Wxe4;t. Most of the ;t's tend much more towards ± than towards Unless Black can challenge at least one of these verdicts, then either it is necessary for Black to vary much earlier, or the ~f4 system is at least as good for White as the ~g5 system. Meanwhile, here are·some practical tips concerning the variations we have just seen. Yes, 10 0-0-0 is potentially an aggressive and violent system

=.

involving a lot of sharp tactical play, but remember that the basic positional advantages lie with White. He has the better development and the freer piece disposition, and the pressure against the opposing centre. His one real cause for concern in his own position lies in the placement of his king. The important thing for White to remember is to keep his positional trumps. This sometimes involves a kingside attack, but Black's king is better shielded by pawns than White's, and it is unwise to deal with the position purely in terms of attacks against kings. White should think first about the centre, and only secondly about the kingside. And if the kingside is the correct theatre of operations, White should lead with pieces (lLlg5, lLlce4, etc.) rather than with pawns. White must not forget either that it is not just the black king that is vulnerable; the queen may still be chased with iOd2b3.

4 Alternatives to the Main Lines for White 4.1 Introduction If this were a repertoire book ('White to play and win with the 5 ~f4 Queen's Gambit') this chapter would be unnecessary; after having completed our survey of 10 0-0-0 we would shoot straight on to what to do against Black's attempts to avoid the main line. Indeed, it will be noted that few of the games in this chapter post-date 1995, the year that Kasparov introduced 11 h4 into grandmaster play. This indicates the consensus of grandmaster practitioners that 10 0-0-0 is indeed the most dangerous try against ...c5 systems, although it is noticeable that Black has been a little more willing to defend this type of position in 1997 than in 1996. Why then include this chapter? The first reason is completeness, which is a good enough reason in itself, from both theoretical and historical points of view. The second reason is that players with the black pieces in particular will want to know about deviations by White from the main line. The third reason is that White too might well want to know about alternative plans, whether through reasons of personal style, or because of the possibility that Black might yet find something strong against the New Main Line.

In general, the lines presented in this chapter aim for a modest edge for White. The extent to which this is achieved is variable.

4.2 White's alternatives on move 10 After 1 d4 d5 2 c4 e6 3 ~c3 ~f6 4 ~f3 i.e7 5 ~f4 0-0 6 e3 c5 7 dxc5

~xc5 8 'ii'c2 ~c6 9 a3 "as, we have already considered 10 l:ld 1 and 10 00-0. White has also tried 10 l:lcl and, more importantly, 10 ~d2. We also mention alternative ideas at move 10 in the sequence 9 l:ldl "as, by which White avoids a transposition to the Old Main Line by 10 a3. Our sections are: A: 9 a3 10 l:ld 93 B: 9 a3 10 ~d2 94 C: 9 l:ldl 10 ~e2 100 D: 9 l:ldl 10 ~d2 101

"as "as "as "as

A) 9 a3 ••510 l:ld (D) Strictly a minor alternative. White renews the threat of b4, and provides extra cover for his knight, but does nothing to develop the kingside or create tension in the centre. Black equalizes comfortably enough. 10...d4

94

The Queen's Gambit Declined: 5 1;.f41 B)

9 a3.aS 10 tDd2 (D)

Or 1O... .i.e7 II .i.e2 and now: 1) 11...dxc4 12 .i.xc4 still gives White a slight pull, e.g. 12 ... ~5 13 .i.xd5 exd5 14 "b3 ;t Taimanov-Larsen, Vinkovci 1970; or 12...e5 13 .i.g3 .i.g4 140-0;t Denk-Foessmeier, Austria 1989. 2) 11...lOe4! 12 cxd5 .!Dxc3 13 "xc3 "xc3+ 14 :xc3 exd5 150-0 .tf6 16 :b3 h6! =Taimanov-T.Georgadze, USSR 1983 seems to be a safe enough option. 11 exd4 11 b4? "xa3 12 bxc5 dxc3 13 "xc3 "xc3+ 14 :xc3 .!De4 + T.Georgadze. 11•••tDxd4 Simple and steady. If Black is out for blood, he may try 11.. ..i.xd4!? 12 .te2 e5 13 .td2 "d8 14 .tg5 h6 15 .th4 g5 16 .tg3 :e8 GO GIek-Lputian, Tashkent 1984. 12 tDxd4 .txd4 13 .i.d3 b6 13 ...e5 14 .td2 "d8 = T.Georgadze. If 15 .tg5? then 15 ....txf2+. 14 0-0 .txc3 15 .xc3 .xc3 16 :xc3 .i.b7 1713 :ac8 Tomaszewski-Staniszewski, Naleczow 1995. White's bishop-pair promises little, as Black's pieces are healthily placed and he has an extra central pawn.

=

A serious alternative to the main lines, which was explored a lot in 1993 and 1994, just before Kasparov's 10 0-0-0.te7 11 h4line altered the terms of the debate. Of Black's choices, the second is more critical: B1: 10....te7 94 B2: 10....tb4 97

1O.....d8?! 11 :d1 "e7, ZaradicBitman, Bucharest 1967, draws obvious comparisons with the 9 .....e7line (Chapter 5.1, line F), with White being a tempo up if he can find something more constructive than 12lbf3. 12 .tg5 :d8 13 .!Db3 dxc4, as in the game cited, is unconvincing, but 12 b4!? .td6 13 .txd6 .xd6 14 cxd5 looks promising. B1)

10....i.e7 (D) Inviting transposition to the Old Main Line with 11 :d 1. Also 11 0-0-0 leads to one of the lesser offshoots of the New Main Line (Chapter 3.4, Line D), with 11 ... e5 12 .tg3 d4 13lbb3

Alternatives to the Main Lines for White

'iFb6 tending towards equality. White does not have to touch the rook though; the following moves are all more constructive: Bll: 11 .te2 95 B12: 11 .tg3 95 B13: 11 ~b3 96 811) 11.te2 Why not? White's main problem in the 10 l:tdl line is that his king gets stuck in the centre inconveniently long, so he prepares to castle before taking action in the centre. 11•••.td7 11...e5, as the previous note hinted, leaves White with an improved version of the Old Main Line. 12 .tg3 d4 13 ~b3 Wb6 14 exd4 (after 14 ~d5, l4 ... 'fId8 15 ~xe7+ 'ilxe7 16 exd4 exd4 17 0-0 ;!; was the game Ornstein-C.Hartman, Swedish Ch 1987, while 14... ~xd5 15 cxd5 ~a5? 16 ~d2! dxe3 17 fxe3 leaves the black knight stranded) 14 ...exd4 15 ~b5 .tg4 16 .txg4 ~xg4 17 0-0 :lacS 18 :adl :lfeS 19 :lfel ± Suba-Margolin, Ubeda 1996. 120·0 Insufficiently forceful.

95

12 O-O-O!? is much more testing, returning to the New Main Line. Although this position is not directly discussed in Chapter 3, White usually preferring against ....td7 to play g4 in advance of .te2, the position after 12 ...:lfc8, and now 13 g4 e5?! 14 g5, R.Kempinski-Lutz, Groningen 1996, was assessed as very good for White in Chapter 3.4, Line D. 12..•:lac8 After 12...:lfc8 13 :lfdl Wd8, forcing an isolated pawn did not help White much in Bern-Heine Nielsen, Gausdal 1993: 14 cxd5 ~xd5 15 ~xd5 exd5 16 Wb3?! ~a5! 17 Wxd5 .ta4! :f. 13 :lfdl Wb6 14 b4 as 15 cS Wd8 16 Wb2 e5 17 .tg3 d4 18 exd4 exd4 19 ~a4 axb4 20 axb4 ~d5 21 b5 ~cb4 22 ~f3 .txc5 23 ~xcS :lxc5 24 ~xd4 'irf6 25 'ira3 Ib.. 1/z YrjolaT.Upton, Pula Echt 1997. But 120-0-0 is critical. 812) 11 J.g3(D)

Prophylaxis, directed against the advance ...e5. 11•••.td7

96

The Queen's Gambit Declined: 5 ~f4!

1l .. :.d8!? and 11...1Wb6!? were, effectively, both tried in the same game! Petrosian-Liberzon, Biel IZ 1976 continued 11..:.d8 12 M (12 :dl .i.d7 13 cxd5lOxd5 14lOxd5 exd5 = Liberzon) 12.. :.a5 13 lOd2 'lrb6 14 .i.e2 d4 15 lOa4 "d8 16 e4lOd7 17 :dl e5?! (17 ... f5 18 exf5 exf5 19 f4 lOf6 co Liberzon) 18 b4 with a slight advantage for White. The text-move is based on the generally quite reasonable belief that it is better to develop than to undevelop, but it leaves Black less solid in the centre. 12 .i.e2 12 :dl :ac8 13 .i.e2, as played in the Costescu-Ionescu game below, transposes.

Black is looking solid in this family of variations.

13cxdS 13 O-O? is carelessly mistimed: 13 ...d4 14 lOa4 "dS :; L.HansenDeep Blue, Copenhagen 1993 . 13... exd5 14 0-0 :ac8 15 :Idl .i.e6 The position is balanced. White's pieces are not placed actively enough to take any advantage of the isolation of Black's d-pawn. Van Wely-San Segundo, Buenos Aires 1995 continued 16lOb3 a6 17 l:tacllOe4!? 18lOxd5 .i.xd5 19 :xd5 lOd4 20 lOxd4 :xc2 21 :xc2=. Possibly 11 .i.g3 is a little too elaborate to cause Black's defences serious problems.

12..•""6 One of several moves to have been tried. 1) 12...:tdS?! 130-0 a6 14 b4 "b6 15 c5 "a7 16lOa4 ± L.Hansen-P.Andersen, Denmark 1993 shows the old problem of the blocked retreat. 2) 12 ...:fc8 13 0-0 "dS 14 :adl d4?! (14 ... lOa5! looks fine for Black) 15 lObS dxe3 16 fxe3 17 "d3 .i.e8 18 b4 :d8 19 1i'b3 ± Tukmakov-S.Jakob, Bern 1991. 3) 12...:ac8 13 0-0 (13 :dl "b6 should be met by 14 cxd5 !? rather than 14 O-O? d4 15 lOa4 "as:; CostescuO.Ionescu, Bucharest 1993) 13 .....dS (intending to play ...d4) 14 cxdS exd5 (14 ...lOxd5!?) 15 lOn ;t L.HansenCifuentes, Wijk aan Zee 1994. 4) 12 .....d8!? 13 cxd5 lOxd5 14 lOxd5 exd5 15 0-0 d4!? 16 e4:C8 17 'lrd3 f5 IS n ~hS and Black was starting to take the initiative in Komarov-Beliavsky, Ni~ 1996.

813) 11lOb3(D)

B

"b6

11•••""6 11.. ...dS 12 :dl a5 13 lOa4 (13 cxd5!?) 13 ... e5, and now 14 .i.g5 d4 15 .i.e2 would transpose into the Old Main Line (Chapter 2.5, Line 04). However, 14 .i.g3! d4 15 .i.e2 g6 16 exd4 exd4 17 0-0 .i.f5 18 .i.d3 was better for White in Tisdall-T.Upton,

Alternatives to the Main Lines for White Gausdal 1993. The g3-bishop exerts pressure on the queenside. 12cxdS This exchange is rather more likely to provide prospects for White, without him having wasted a tempo on the move .i.g3, as in Line B 12 above. Other tries: 1) 12 l:ldl e5 13 .i.g5 again transposes to the Old Main Line (Chapter 2.6 above). One feels that White ought to be aiming for more. 2) 12 .i.e2 and then: 2a) After 12... e5, 13 .i.g3 transposes into Line B 11 above, which is favourable to White. A less successful idea for White is 13 .i.g5 d4 14 .i.xf6 dxc3! 15 .i..xe7 cxb2 16 tWxb2lL1xe7 = I.Garcia-A.Hemando, Ibercaja 1993, a tactical possibility not available in positions where l:ldl has been played. 2b) Black might well prefer to try 12... lLIaS!? 13 lLIxaS "xaS 14 0-0 dxc4 15 .i.xc4 .i.d7, much as in line '3' below. 3) 12 .i.d3 h613 .i.g3lL1a5 (13 ...d4? 14 lLIa4 "a6 15 lLIxd4) 14 lLIxa5 It'xa5 150-0 dxc4 16 .i.xc4 .i.d7 17 b4 "b6 18 "b3 l:lfc8?! (18 ...lLIh5!? 00) 19 :Udl ;t Tomaszewski-Slezka, Sumperk 1990. 12••.lLIxdS 13lL1a4 13lL1xd5 exd5 14 l:dl.i..e6 = Przewoznik. 13..:ii'd8140-0-0! This, rather than the obvious 14 .l:1d 1, beeause of the unlikely tactical point that White does not want Black to play ...lLIxc2 with check. Thus if 14.....e8: 1) 15 .i..b5 .i.d7 (15 ... lLIcb4?! 16 'ii'e2 lLIc2+? 17 ~f1 .i.d7 18 .i.xd7 'ii'xd7 19lL1c3 +-) 16lL1ac5? lLIcb4!.

97

2) Black has no problems either after 15 .i.d3lL1xf416exf4lL1b417 axb4 "xa4:j: Kelei!evic-Przewoznik, Finkenstein 1990. 14••••e81S .i.bS! .i..d716lL\acS Now 16...lLIcb4? is no longer possible: 17 .i..xd7 +-. 16•••l:lcS 17 lLIxd7 .xd7 18 ~bl a619.i..c4 bS 20 .i..xdS exdS 21 .d3 With strong play against the isolated d-pawn, Vaganian-Arfandi, Reggio Emilia 1992/3. Theory is far from settled yet, but 10 liJd2 .i..e7, and either lllLlb3 or 11 .i.e2, gives White freer and more flexible play than in the 10 l:dl .i.e7 lines.

B2} 10.••.i..b4! (D)

A much more direct challenge to White's initiative. 11 cxdS 1) The defensive 11 l:lc 1 is ineffective after 11.. ..i..xc3 12 'ifxc3 'ifxc3 13 l:lxc3 e5 14 .i.g5 (14 .i..g3 d4 15 l:lcl .i.f5 =Gavrikov-Peshina, USSR 1981) 14 ... d4 15 l:lcllL1d7! (taking the initiative; 15 ... lLIg4 16 h3 dxe3 17 fxe3, Benko-T.Berger, Amsterdam 1964,

98

The Queen's Gambit Declined: 5 ~f4!

=

and now 17... h6 ECO; 15 ... dxe3 16 fxe3 l:d8 17 .i.xf6 gxf6 = H.Griinberg-Bonsch, Hanover 1991) 16.i.e2? (16 exd4 exd4 ':F) 16...h617 .i.h4f518 h3 g5 19 .i.g3 f4 20 exf4 exf4 21 .i.h2 l:e8 22 ~dllOc5 23 b4lOa4 24 lObi .i.f5 25 ~d2 1Oc3 0-1 KonikowskiFoeldi, corr. 1978 - not exactly the best advertisement for White's f4bishop. 2) ll.i.e2 d4 120-0 .i.xc3 131Ob3 'ii'a4 14 bxc3 dxe3 15 fxe3 lOa5 00 112_112 H.Griinberg-M.Miiller, Bundesliga 199112. Possibly both sides were unhappy with their position. White has the bishop-pair and the higher FIDE rating, but having all those pawn islands is not much fun. 1l•••exdS 1) 11...e5 12 .i.g3 .i.xc3 13 dxc6 .i.xd2+ 14 'ii'xd2 'ii'xd2+ 15 ~xd2 lOe4+ 16 ~ellOxg3 17 hxg3 bxc6 18 l:h5 l:e8 19 ~d2 as 20 ~c3 leaves several question marks hanging over Black's pawn structure in the endgame, Suba-Cvetkovic, Belgrade 1984. 2) 11...lOe7?! is reasonably promising if White exchanges on e6 (12 dxe6?! .i.xe6 13 l:cl .i.xc3 14 'ii'xc3 'ii'a4 00), but Lputian notes that 12 d6! lbed5 13ll'lxd5ll'lxd5 14 l:d 1 strongly favours White. Black will have to contort himself to regain the pawn. 3) 11 ... lOxdS releases the tension a little too early, giving White chances of play against the isolani. 121Oxd5 exd5 13 .i.d3 (13 l:dl.i.e7 14 .i.d3 h6 15 0-0 .i.g4! 16 1Of3 d4 = O.FoisorPigusov, Sochi 1995) 13 ... h6 14 l:cl .i.e7 (14....i.xd2+ 15 'ii'xd2 ;!; Bareev) 15 0-0 .i.e6 16 lOb3 'ii'b6 17 lOc5 .i.xc5 (17 ... l:ac8?! 18ll'lxe6 fxe6 19 'ii'e2 .i.f6 20 'ii'h5! and White has

successfully transferred his attack to the kingside: 20...'ii'xb2? 21 'ii'g6 followed by a well-timed l:bl wins, while 20 ...lOe7 21 b4 ±, as in Kramnik-Htibner, Bundesliga 1993/4, extends White's bind) 18 _xeS 'ii'xb2 19 l:bl 'ii'd2 20 l:fdl 'ii'a5 21 l:b5 'ii'a4 22 l:al! and White regains the pawn on b7 with a slight edge, BareevDzhandzhgava, Debrecen Echt 1992. 12.i.d3! Another line you won't find in ECO! This has supplanted the older 12 lOb3 .i.xc3+ 13 bxc3 'ii'a4 (13 ....i.f5?? 14 lOxaS .i.xc2 15 lOxb7, StohlFranzen, Tmava 1985) 14 .i.d3 b6 (14 ...l:e8!? 150-0 ll'le5 16ll'ld4 {16 .i.xe5 l:xe5 17 c4 J.e6 =} 16...'ii'xc2 17 .i.xc2 .i.d7 =Agdestein-Gild.Garcia, New York tpd 1994) 15 f3 (15 lOd4!? Agdestein), when Agdestein suggested 15 ....i.a6 16ll'ld4 _xc2 17 J.xc2 lOaS as equalizing. Black has pressure on the queenside, NimzoIndian style, to compensate for the loss of the bishop-pair. In AgdesteinHjartarson, Gjl1Jvik (5) 1985, Black played less convincingly, and after 15 ...l:e8 16ll'ld4 'ii'xc2 17 J.xc21OhS 18 J.a4 J.d7 19 J.xc6 J.xc6 20 J.g5 h6 21 J.h4 J.d7 22 ~d2 White had the better-placed minor pieces. 12...d4 (D) 130-0! This Korchnoi idea put the 10 ll'ld2 variation back on the theoretical map in 1990. 13.••J.xc3 13 ...dxc3?! 14 axb4 'ii'xb4 15 bxc3 leaves White with much the more active pieces, so Black has little alternative but to fall in with White's idea. Korchnoi suggested 13 ....i.e7 in his

Alternatives to the Main Lines for White

Informator notes, but it remains untested, probably with good reason, as 14ltX4! Wd8 ISlDe2 dxe3 16 fxe3!? leaves White with a lot of clout in the centre. 14lbc4 WhS 15 bxc3 Now Black has three choices: B21: IS•••dxe3 99 B22: IS•••dxc3 99 B23: Is...lbclS 100 821) IS...dxe3 16 ~e3 (D)

16....i.d7!? 16....i.e6 and now: 1) Following 17 :abl, 17 ... ltXiS? led to a catastrophe in Korchnoi-Beliavsky, Amsterdam 1990 after 18 :bS

99

:fd8 19 c4! lOd4 20 Wb2lOxbS 21 cxdS .i.xdS 22 j,xbS j,e4 23 j,c7 :dc8 24 :cl, when the two minor pieces easily outweighed the rook. 17 ... WcS!? looks more secure, with counterplay against the a-pawn. 2) In Nogueiras-Gild.Garcia, Matanzas 1994, White kept the a-pawn covered with 17 :fbl :ac8 18 j,e2 (18 :xb7ltXiS! gives Black counterplay; having the rook on the c-file makes a big difference) 18...WcS 19 :xb7 ltXi4 20 cxd4 Wxd4 21 Wb2 Wxf4 22 :xa7 ;1;. White is an isolated passed pawn up, but it will be very difficult to advance it if Black keeps his pieces active. 17 :abl 17 :fdl?! lOe7! 18 c4 :fe8 19 j,e2 WaS 20 Wb2 lOg6 21 j,d6 .tc6 22 cS lOe4 ;: Dreev-Vaganian, Moscow 1991. 17...lOas ;I; Vaganian; GO Dreev. Black hopes to equalize gradually with ...j,c6, ... b6, etc. Not, however, Vaganian's alternative proposal 17 ...lDeS?, which leaves the queen short of squares: 18 j,e2 lOeg4 (18 ... lOfg4 19 j,xe5 lOxe3 20 j,xhS lOxc2 21 :xb7 +- Dreev) 19 lOxg4 .txg4 20 :bS lOdS 21 :xdS WxdS 22 j,xg4 +- Dreev-L.Ravi, Calcutta 1992. 822) IS...dxc3 (D) 16.tg3 16 Wxc3? lOdS ;:. 16 j,d6:d8 17 Wxc3, with the idea 17... bS 18 lOb6, was suggested by Korchnoi, but Black is gaining time after 17 ... j,e6.

100

The Queen's Gambit Declined: 5

16••. bS 17lLld6 b4 18 .i.bS 18 axb4lLlxb4 19 "xc3 lLlxd3 20 'ilxd3 .i.e6 =. 18•.•.i.d7 19 .i.xc6 .i.xc6 20 axb4 :ad8 21 'ii'g4 22 'ii'xc3 l:[xd623 f3 .i.xf3 24 gxf3 'it'g6 Black equalized in Stohl-Korneev, Hamburg 1993 after 2S :xa7?! lLlh5 26 ~g2 :c6, but 2S e4! would have maintained White's initiative.

:as

823)

Is•••lLlds (D)

17...exf2+? 18 :xf2 .i.e6 19:e4 ± Lalie. 18lLlxe3 18 fxe3 is messy, but probably not favourable to White, e.g. 18 ...:ad8 19 lLld6lLleS 20 .i.fS (20 lLlxb7lLlxd3 21 "'xd3 :d7 22 "'bS "g4 GO Lputian) 20 ...:xd6 21 .i.xeS :dd8 22 e4lLle7 23 .i.d4?! (23 :e3 lLlg6 24 .i.d4 b6 GO Lputian) 23 ....i.xfS 24 exfS lLlc6 =F Tukmakov-Lputian, Tilburg 1994. 18...h6 18 ...:ad8!? 19lLlxdS :xdS and now Lalie gives 20 .i.c4 :d2 21 'ilxd2 .i.xc4 =. Instead 20:e4? would be positionally weak due to 20 ....i.fS 21 "'g6, but maybe 20 c4!? 19lLlc4?! Too time-consuming. White still has the initiative after 19 c4!, the weakening of the d4-square being of little concern. 19...:ad8 20 lLld6 20 lLleSlLlce7! 21 .i.e2 "'fS =Lalie. 20•••b6 21 .i.bS lLlde7 = L.HansenB.Lalie, Moscow OL 1994. It would seem that White's prospects in the 10 lLld2line are better than in the 10 :dlline, but not as good as after 10 0-0-0. White has chances for a modest edge in the critical lines.

:h4

C)

Simple piece-play preserves Black's opportunities. Lputian now gives 16 .i.d6? :d8 17 cxd4 bS 18 lLleS :xd6 19lLlxc6 :h6 +. 16 .i.g3 dxe3 17 :ael .i.e6

~f4!

"'as

9 :dl 10 .i.e2 (D) Instead 10 a3 is the Old Main Line. This position was reached via the move-order 8 .i.e2 lLlc6 9 "'c2 (9 0-0I?; 9 cxdS!?) in the game O.Foisor-Morovie, New York 1988. Play continued 1O....i.e7 11 lLld2 eS 12 ~b3 Wb6 13 .i.gS d4 14 .i.xf6 .i.xf6 15 ~dS "d8, and White had indeed saved a tempo by missing out a3, but

Alternatives to the Main Lines for White

since he played 16 .td3 next anyway, this was hardly an improvement. Black can play more vigorously: 10•••~e4!? This is not so convincing if White has already played a3 (Chapter 2.2, Line E), but Black's ability here to place a piece on b4 without coming to harm from the a-pawn has some tactical implications. 11 cxd5 exd5 12 :xd5 ~xc3 13 bxc3 13 "'xc3.tb4 14 :xaS .txaS wins the exchange. 13...~b4 14 cxb4 .txb4+ 15 :d2

101

16...:fcS? 17 .xh7+ eRfS IS 0-01 turns the tables. 17 .xd2 .xd2+ 18 eRxd2 :td8+ IS ... .txe2 19 eRxe2 :acS 20 :dl leaves White perhaps a little better in the endgame, but the weak a-pawn will need attention. 19 .td3 l%ac8!? 19 ...:d7?! 20 :cl :adS 21 :c3 consolidates White's position. The text is complicated, but given that Black's rooks are well placed on open files, while White's minor pieces are far from fully coordinated and his a-pawn is weak, it is likely that Black is holding his own. If for example 20 f3, then 20...:c5!?

.tg416~g5

16 it)d4? :acS 17 "'dl .txd2+ IS "'xd2 :cl+ 19 .tdl :xdl#. 16 e4?! .txf3 17 .txf3 :acS IS "'dl :fdS leaves a killer pin on the as-el diagonal. 16 .tc7! is Burgess's suggestion. After 16......d5 White has the pleasant choice between 17 e4 "'d7 IS 0-0 .txd2 19·~e5 "'e7 20 ~xg4 :acS 21 'ii'xd2 :xc7 22 .td3, followed by ttJe3-d5, or 170-0 .txd2 IS :dl :acS 19 :xd2 and Black cannot exploit the c-file pin. So maybe Black must try 16... b6, and if 17 a3 then 17 ....tc3. 16....txd2+

This is almost totally unexplored, but ultimately unconvincing. White again hopes to save a tempo by not playing a3; Black naturally must respond by making use of the unprotected b4-square. 10....tb4 11 ~b3 .b6 12 .td3 1) 12 a3 .txc3+ 13 .xc3 eS (Filip) leaves Black with a significant lead in development.

102

The Queen's Gambit Declined: 51./4!

2) 12 j,e2?! eS 13 j,gS (not 13 j,g3? d4 14 a3 j,xc3+ IS bxc3 dxe3) 13 ...~ 14 cxdS .txc3+ 15 bxc3 ~g5 16 dxc6 'ifxc617 0-0j,h3 18 f3 j,d7 is slightly better for Black. 12.••e5 13 .tg5 d4 1) 13 ...dxc4 14 j,xc4 j,g4 (Filip), and now IS %ld2 ;to 2) 13 ... j,g4 14 f3 e4 (Filip), and now not IS j,xf6?! exd3 16 'ifxd3 dxc4 (Burgess; 16...gxf6 17 fxg4 /t)eS =F is not so sharp) 17 'ifxc4 (17 .td4 cxd3 18 j,xb6 axb6 19 fxg4 %lxa2) 17 ... j,e6! winning a piece, but rather IS j,e2 exf3 16 gxf3 dxc4 17 lLld2 .te6 18 j,xf6 gxf6 19 'ife4 with attacking chances. 14 j,xf6 gxf6 15 cS?! An unnecessary weakening. 15 a3 is best: 1) IS ...dxc3 16 axb4 cxb2 (White is slightly better after 16... lLlxb4 17 'ifxc3lOa2 18 'ifc2 c!Llb4 19 'ife2) 17

20 %lxd4 'ifeS+, winning a piece after either 21 %le4 j,xc3+ 22 bxc3 'ifh5, or 21 'ife4 j,xc3+ 22 bxc3 %lh8. 17 0-0 dxc3 18 axb4 lLlxb4 19 'ifxc3lLlxd3 =P.Littlewood-Speelman, Hastings 198112.

4.3 White's alternatives on move 9 1 d4 dS 2 c4 e6 3 c!Llc3lLlf6 4lLln j,e7 5 j,r4 0-0 6 e3 c5 7 dxcS j,xcS 8 a3 c!Llc6 (D)

bSlLlb418j,xh7+~g719'ifd2~xh7

20 'ifxb4 ±; if 20....tfS then 21 e4 as (21 ... j,xe4? 22 cS +-) 22 'ifa4 j,e6 23 0-0 (23 cS? j,xb3) 23 ...%lac8 24 cS j,xb3 2S 'ifxb3 %lxcS (if the other rook had moved to c8 to give the king a flight square, 26 'ifxf7+ would now win) 26 'ii'h3+ ~g7 27 'ifg4+ followed by 28 %ld3. 2) IS ... j,xc3+ 16 bxc3 dxe3 17 0-0;;1;. 3) IS ....taS! 16 c!Llxa5 'ifxa5 17 exd4 exd4 18 0-0 dxc3 19 .txh7+ ~g7 20 %ld5 'ifc7 21 'ifxc3 'iff4 22 g3 'ifh6 23 j,e4 j,e6 =F. There are other attacking tries for White, but Black always seems to come out ahead. 15...1i'c7 16 83 ~g7?! 16... j,a5! is again strong. 17 exd4 exd4 18 c!Llxd4 c!Llxd4 19 j,xh7+ q;g7

The main alternative to entering the main lines with 9 'ifc2 is 9 j,e2, although there is a strong argument for suggesting that White does better to play 8 j,e2, and to try to do without a3. This is covered in Chapter 4.4, Line A. 9 b4 and 9 %lcl have also been tried. We discuss: A: 9 adS 102 B: 9 b4 103 C: 9%lel 104 D: 9 j,e2 lOS A)

9 cxd5 has no independent significance, and transposes directly into

Alternatives to the Main Lines for White systems discussed in Chapter S, under 8 cxdS. 9 ...lDxdS 10 lLlxdS exdS transposes into 8 cxdS lDxdS 9 lLlxdS exdS 10 a3, while 9 ...exdS 10 ~e2 transposes into 8 cxdS exdS 9 ~e2lLlc6 10 a3 (instead of the critical continuation 10 0-0). B) 9 b4 (D)

Aiming to show that a3 can be directly aggressive, and not just part of a prophylactic plan. Having said that, it seems a little unusual to use this plan when the Queen's bishop is already outside the pawn-chain; it would be more naturally placed on b2 rather than f4.

9.••~e7 1) 9 ...~d6 and now: la) 10 cxdS exd5 11 ~gS ~e6 12 c!DbS ~e7 13 lLlbd4 c!De4 =Barkhagen-Wedberg, Stockholm Rilton Cup 1992. White's queenside pawn advances actually make it more difficult to play against the isolani. Ib) 10~gSdxc411 ~xc4.!DeSI2 ~e2 lLlxf3+ 13 ~xf3 as 14 bS a4 IS 0-0 hS with equality, Mikhalchishin-Koziak, Lvov 1996.

103

lc) 10 ~xd6 'ilfxd6 11 ~e2! (11 cxdS exdS 12 ~e2 ~g4 = SjOOinBarkhagen, Vaxjo 1992) 11...l:ld8 12 0-0 b6 13 cxdSlLlxdS 14lLlxdS .xdS IS.xdSl:lxdSI6l:lfdl ~d717 l:lxdS exdS 18 l:lcl ~ Mikhalchishin-Ubilava, Volgodonsk 1981. Instructive timing by White: instead of exchanging on dS immediately, and isolating Black's pawn but releasing his bishop, White bided his time and exchanged only when he had completed his own development and had forced Black to commit his bishop. 2) 9...~b6 and then: 2a) 10 l:la2.!De4 11 c!Dxe4 dxe4 12 .xd8 ~xd8 13 lLld2 (13 .!Des i.e7, aiming to hit the queenside pawns with ...as) 13 ...eS 14 ~g3 fS IS bS f4 16 bxc6 fxg3 17 bxg3 bxc6 = StohlPrandstetter, Czech Ch (Tren~ianske Teplice) 1995. 2b) 10 ~e2 lLlh5!? (1O ...dxc4 11 .c2;t; 1O...d4?! 11 exd4lLlxd412 cS! ±) 11 cxdS lLlxf4 12 exf4 (12 dxc6 .f6!) 12...•f6 13 0-0.xc3 14 dxc6 Mikhalchishin-Magomedov, Tallinn rpd 1988. Since this was a quickplay, one should not expect highly refined decisions, and 14...•xc6!? IS l:lci .dS would appear to be more accurate than 14...•f6 IS g3 bxc6 16 "c2 ~b7 17 .!Des ~, as played. lOcxdS 1) 10 "c2 is the main alternative, but does the insertion of b4 really help White's development? Play down the c-file is the natural plan, for example 1O...~d7 (or 10...dxc4 11 ~xc4 i.d7 12 0-0 l:lc8 13 "e2 c!Dhs 14 .i.eS lLlxeS 15 lLlxeS ~f6 16lLlxd7 "xd7 17 l:lacl g6 Mikhalchishin-Renet, Dortmund 1991) 11 ~e2 (11 cxdS can

=

=

104

The Queen's Gambit Declined: 5 ~f4!

be met by 1l...~xd5 12 ~xd5 exd5 :;:

Olafsson; instead 11...exdS 12 b5 ~aS 13 .td3 Ilc8 140-0 lbc4 =was Grynszpan-R.EkstrOm, Lugano 1989) 11...llc8 120-0 dxc4 13 Iladl 'if'e8 14 Ild2 as 15 bS ~b4! 16 axb4 axb4 17 ~e4 b3 :;: Christiansen-H.Olafsson, Reykjavik 1986. 2) 10 1la2 and now 10... aS 11 bS ~b8 12 .te2 b6 13 cxd5 ~xdS 14 ~xdS exd5 Is0-0;t was ArkhipovL.Spassov, Moscow 1985. Olafsson suggested instead 1O...tDe4 00, with 11 ~xe4 dxe4 12 'if'xd8 .txd8 transposing to the Stohl-Prandstetter game given under 9....tb6. 3) 10 h3 dxc4 ll.txc4 'if'xdl+ 12 Ilxdl gains a tempo on the 9 .te2 dxc4 10 .txc4 lines discussed later, but b4 is as much a weakening move as a constructive one. 12 ... a6 130-0 b5 14 .td3 Ild8 15 ~41Dd5 16 .tg3 .tb7 = Granda-Ivanchuk, Amsterdam 1991. 10...~xdS 1O...exdS 11 .te2.te6 121Dd4 Ilc8 13 0-0 ~xb4 14 axb4 Ilxc3 15 Ilxa7 'if'b6 16 ~bS ;t Zsu.Polgar-Geller, Vienna 1993. 11 ~dS exd5 12.td3 .tf613 Ild a6 This is probably a wise precaution. 13 ...lle8 14 0-0 .tg4 15 h3 .txf3 16 'if'xf3 ~eS 17 .txeS IlxeS 18 Ilfdl ;t Mikhalchishin-Heine Nielsen, Copenhagen 1991. 140-0.te6 14... g6 is unnecessarily passive: IS h3 'if'e7 (lS ....tb2? 16 Ilc2 .txa3 17 'if'al.txb418.th6±) 16'if'd2.te617 Ilc2 ;t Heine Nielsen-Barkhagen, Mamaiajr Wch 1991. 14...lle8 IS .tb1 g6 16 .ta2 .te6 17 'if'd3 d4 18 .txe6 Ilxe6 19 e4 'if'e7

20 ~d2 .te5 21 .tg3 .txg3 22 bxg3 ~S 23 11'b3 Ild8 24 IlcS d3 was played successfully in BeliavskyShort, Groningen FIDE Wch 1997. 15 .tbl .e7 16 e4?! 16 'if'd3 00. 16....d7! Black has taken the initiative, DreevLutz, Dortmund 1994.

C) 9 Ild (D)

B

9...dxc4 The simplest, but perhaps not very plausible as a winning attempt. 1) 9...d4!? and now: la) 10 exd4 ~xd4 1b) White can also try 10 ~xd4, e.g. 1O...~xd4 11 b4! ~c6 12 'if'xd8 Ilxd8 13 bxcS ± Knezevic-Szabo, Belgrade 1977. However, Black can improve with 10... eS! 11 ~b3 (11 ~xc6 'it'xd1+ 12 Ilxdl exf4 13 ~d4 fxe3 14 fxe3 lle8 with ample compensation for the pawn - Kramnik) 1l....txa3! 12 bxa3 exf4 13 'if'xd8 Ilxd8 14 exf4 .te6 15 f3 (15 lbcS?! .:r.e8!) and now Kramnik-Beliavsky, Belgrade 1993 continued IS ... b6? 16 ~f21Dd4 17 ~xd4 Ilxd4 18 ~e3 ±.

=.

Alternatives to the Main Lines for White Kramnik notes that Black should not be scared of l'lX:S: lS ...:ac8! 16lDcs lDaS 17 lDxe6 fxe6 =. If anything, Black could be regarded as slightly better. 2) 9 ... a6 and 9 ... j.e7 both transpose to the 'Exchange Variation' of Chapter 4.4, Line B2 after 10 cxdS exdS. White is better. 3) 9 ... j.b6 (since 9 ... d4 now appears to be satisfactory for Black, this is revealed as an unnecessary loss of tempo) 10 j.e2 (10 cxdS!?) 10 ... d4 11 exd4 lDxd4 12 lLlxd4 j.xd4 13 0-0 ;j; Orlowski-Gerigk, Germany 1994. 10 j.xc4 .xdl+ 11 ~xdl 11 :xd1 transposes into the line 9 j.e2 dxc4 10 j.xc4 .xd1+ 11 hd1, discussed in Line D 11 below. The text aims to squeeze a little more from the position. 11•••b6 12 ~e2 j.b713 :hdl :Cd8 14lDb5 lDe8 15 j.d3 as Black aims to stabilize the position by restraining White's b4 idea. Kramnik gives lS ... j.a6 16 lDc7 lDxc7 17 j.xc7 ':xd3 18 :xd3 ':c8 19 j.g3 l:td8 20 :d1 f6 21lLlel eS 22 ~f3 as also equalizing. lS ...h6?! 16 b4 j.e7 17 j.e4 :xd1 18 ~xdl! lDd8 19 j.xb7 lDxb7 20 lDfd4 ±, Kramnik-Vaganian, Lucerne Wcht 1993, is too slow for Black. 16 lDgS lLlf6 17 j.g3 j.a6 18 a4 lDb4 19 j.c4 h6 = Dreev-Vaganian, Tilburg rpd 1993.

D) 9 j.e2 (D) Now: Dl: 9...dxc4 D2: 9•••a6 D3: 9•••

:e8

lOS 110 110

D4: 9 ...lLle4!? DS: 9••• j.e7 D6: 9.••others

105

111 112 112

D1} 9•••dxc4 The most direct of Black's many alternatives. Black aims to take the sting out of the position by exchanging queens. 10 j.xc4 10 'tixd8?! is unhealthily eager, and not worth analysing. 10 .c2!? at least makes sure the queens stay on. Then Black can continue: 1) 1O.••bS! forces White to continue in gambit fashion. 11 0-0 (11 lDxbS?! .as+ 12l'lX:3 lDb4 13 'tiel lDbdS 14 j.xc4lLlxf4 IS exf4 j.xf2+ 16 ~xf2 'ticS+ 17 ~g3 'tixc4 I8iLleS "'b3 =+= Gavrikov-Lutz, Bundesliga 199516) 11...a6 12 :adl (12 lDgS?! j.b7 13 :adl "'e7 14 lDge4lDxe4 IS lDxe4 eS 16 j.gS f6 17 j.h4 112-112, but probably:j:, Adorjan-Helmers, Gjjljvik 1983) 12...Wb6 (12 ......e7 13 j.gS! h6 14 j.h4 gS? IS lLlxgS) 13lDg5 h6 14 lDge4 lLlxe4 15 lLlxe4 e5 16 j.g3 j.e6 17 j.f3 112-112, Adorjan-Karlsson, Gjjljvik 1983, 17....te7!? 00.

106

The Queen' s Gambit Declined: 5

2) 1O...~a5?! is an inferior way to hold on to the pawn: 11 :d 1 "eS 12 ~g5 i.e7 13 i.e5 g614 ~ge4 ~xe4 15 ~xe4 f6 16 i.c3 .b5 17 0-0 and White has a clear advantage, Szymczak-Novak, Poland 19S2. 3) 1O...i.d7 11 :dl i.e7 12 i.xc4 :cS (12 ...•a5 130-0;1; Adorjan-Van der Sterren, Plovdiv Echt 19S3) 130-0 ~h5 14 i.e5 "eS 15 .e2 a6, and now instead of 16 ~e4 ~xe5 17 ~xe5 i.c6 = Gavrikov-Ubilava, Minsk 19S3, maybe 16 :d2!?;I;. 4) 1O...•e7?! 11 i.xc4 :dS 120-0 i.d6 13 i.g5 h6 14 i.h4 ~e5 15 ~4 ~g6 16 i.xf6 gxf6 17 :fd 1 ± MuseBonsch, German Ch 1991. 5) For 10... a6, see 9... a6 10 .c2 dxc4. After 10 i.xc4 play again branches out, according to whether or not Black exchanges queens immediately: 011: 10.....xdl+ 106 012: 10••. ~hS! 107 013: 10•••86 107 014: 10•••b6 109 011) 10••••xdl+ 11 :xdl (D) 11 ~xdl b6 12 ~e2 i.b7 13 :hdl :fdS 14 :xdS+ :xdS 151.dl :xdl 16 ~xdl lh-lh Hertneck-Yusupov, Munich 1992, is probably the most boring game in this book. Black has a slight lag in development he must try to neutralize. White will play for e4-e5. 11•••i.e7 1) 11...a6 12 i.e2!? (12 i.d3 :dS 13 i.g5 h6 14 i.h4 e5 15 i.xf6 gxf6 16 ~h4 i.e6 is unclear, Zsu.PolgarGomez Esteban, Pamplona 1990/1) 12...:dS (12...b5 13:c1 Gurevich) 13

~f4!

:xdS+ ~xdS 14 0-0 i.d7 (14... b5!? Gurevich) 15 :dl i.eS 16 ~d2 ~6 17 ~e4 ~xe4 18 ~xe4 i.e7 19 ~d6 ;I; M.Gurevich-R.Ekstrom, Neuchlitel 1996, demonstrating that Black has to handle his pieces very carefully ifhe is to equalize. 2) After 11 ... b6 12 e4, a top grandmaster even lost a miniature, though admittedly in a quickplay. 12...i.b7?! 13 e5 ~g4 14 ~4 :fdS 15 0-0 ~5 16 ~xc5 ± i.xf3? 17 :xd8+l:xd8 IS lLlxe6 :e8 19lLlc7 :c8 20 e6 ~xc4 21 e7 i.c6 22 :dl :eS 23 ~e8 i.xeS 24 :dS ~f6 25 i.g5 1-0 Kramnik-Yusupov, Moscow PCA rpd 1995. Safer for Black is 12 ...:d8 13 :xdS+ ~xdS 14 e5 ~eS, with Dreev-Vaganian, Tilburg 1993 continuing 15 ~e2 i.b7 16 :dl h6 17 i.e3 i.xe3 IS ~xe3 =, but 15 ~e4 i.e7 16 i.d3 i.b7 17 ~e2 would force Black to work hard to close the .8ap. 12e4~d7!?

12...:dS 13 ~e2 i.d7 14 e5 ~eS 15 :d2 :acS 16 :hdl ~bS 17 i.d3 i.c61S i.e4 :xd2+ 19 :xd2 :dS 20 :xdS i.xdS gave White a slight advantage in mobility in the minor-piece ending in Vaganian-BeJiavsky, Tilburg rpd 1993, although he eventually

Alternatives to the Main Lines for White contrived to lose it by leaving his epawn weak. 13~e2a6

Black did well to delay touching his queenside pawns, as he could thereby wait to see what White did before deciding his own pawn structure. 14 .i.d6 bS IS 1.xe7lDxe7 161.a2 b4 17 axb4lbc6ISl:[d6lDxb4 19 1.c4 l:ta7 20 l:ta1l:[c7 21 ll)d2 lDes 221.b3 l::td7 23 l:[xd7 .i.xd7 112-112 BareevNovikov, Lvov 1987. So Black should equalize, but he must take great care. There is, however, a safer way to exchange queens, as we shall see.

012) 10•••lDhS! (D)

This continuation has the reputation of being a reliable equalizer. Black forces simplification without loss of time. ll'iVxdS 11 1.g5 1.e7 12 1.xe7 (12 .xd8 l:[xd8 transposes into the next note) 12 ...•xe7 13 0-0 l:[d8 14 .e2 1.d7 with equality, P.Nikolic-Ki.Georgiev, Dubai OL 1986. 1l•••l::txdS 12 1.c7

107

12 1.g5 1.e7 13 1.xe7 lDxe7 14 ~e2 (14 0-0 1.d7 15 l:[fdl 1.c6 16 lDe5 lDf6 17 .i.e2 ~f8 = Lechtynsky-

Dobrolowski, Rimavska Sobota 1990) 14 ...1.d7 15 1De5 1.e8 16l:[hdl lDf6 17 f3 (17 e4 ~f8 18 ~e3 ll)c6 19 lDxc6 1.xc6 20 f4 ~e7 = AgzamovI.Zaitsev, Erevan Z 1982) 17...~f8 18 e41Dc6 19lDxc6 .i.xc6 20 .i.b5, and in Liogky-Marciano, St Affrique 1997, Black needlessly landed in an inferior endgame after 20....i.e8?! 21 .i.xe8 lDxe8 22 e5 ~e7 23 g3 f6 24 exf6+ i. Instead, 20 ...1.xb5+ 21 lDxb5 l:[xdl 22l:[xdl ~e7 ought to be a safe draw; if23l:tcl a624l:tc7+??then24...~d8. 12•••l:[d7 13 1.eS lDxeS If 13 ...l:[d8, any advance on 14 1.c7 l:[d7 112-112 Zlochevsky-Kharitonov, Moscow 1996? 13...b614ll)e4 and now 14...lDxe5?! 15 lDxe5 l:[c7 16 1.e2 lDf6 17 lDxc5 bxc5 18 1.f3l:[b8 190-0-0 i was Dautov-Beliavsky, Reggio Emilia 199516. 14... .i.e7, preserving the bishop, is safer, and led to a draw after 15 1.c3 1.b7 16 1.e2 l:[dd8 17 0-0 lDa5 18 lDed2 lDc6 19 lDe4 lDa5 20 lDed2 in Lobron-Lutz, Nussloch 1996. 14lDxeS l:[dSIS 0-0 lDf616l:[fdl 1.d7 17 lDxd7 l:[xd7 IS l:[xd7 lDxd7 19 l:[dl ll)f6 20 ~n 112_112 TopalovGelfand, Vienna 1996. This variation is a very safe choice, provided Black is not careless or ambitious. D13) 10•••a6 (D) Keeping close symmetry, but allowing a bit more tension in the position than after 10...1Dh5. 1l"c2

108

The Queen's Gambit Declined: 5 j:,f4!

A natural square for the queen, and a move that Black cannot reciprocate. 1) 11 0-0 bS 12 .td3 .tb7 13 "'c2 (13 :cl .te7 14 We2 :cS was soon agreed drawn in the game YuferovZhachev, Moscow 1990) 13 ....te7 14 :adl Wb6 (14......a5!?) IS e4lLlh5 16 .te3 "'c7 17 eSlLlxeS ISlLldS! lLlxf3+ 19 gxf3 Wd6 20 lLlb6 Wc6 21.txh7+ ~hS 22 Wxc6 .txc6 23 .te4 .txe4 24 fxe4 ;t Hellsten-San Segundo, Pula Echt 1997. 2) After 11 .te2 White managed to squeeze something out of 11.. .lLlhS 12 .tgS .te7 13 WxdS .txdS in Hauchard-Marciano, French Ch 1996: 14 lLle4 f6 IS lLlfd2 ;t and the knight is awkwardly placed on hS. Instead 11....td7!? should be safe enough. 11•••.te7 The most popular choice, keeping the bishop out of trouble, but not necessarily the most solid. 1) ll...bSI2.td3.tb713lLle4(13 0-0 transposes to HeIlsten-San Segundo) 13 ... .te7 (13 ...WaS+?! is mistimed: 14 ~e2 i.e7 ISlLlxf6+ .txf6 16.txh7+~hSI7 :acl :acSlSWbl b4 19 .te4 ± I.Sokolov-Cifuentes, Oviedo rpd 1992) 14lLlxf6+ .txf6 IS .txh7+ ~hS 16 :dl WaS+ 17 ~e2

:acS IS Wbl b4 with unclear play, Maiwald-Bonsch, Berlin 1994. 2) 11.. ..td7 and now 120-0 lLlhS 13 .tgS .te7 14 :adl h6 IS .txe7 Wxe7 16 .ta2 1/2- 1/2 was Stohl-Ubilava, Trencianske Teplice 19S5. 12 ':dl is much more challenging, and if 12 ...lLlhS?! 13 lLlgS g6 14 lLlge4 lLlxf4?, then IslLlxcS. 3) 11...We7 12 .tgS h6 13 .th4 Wc7!? (challenging White to play 14 .txf6, which, however, leads to no clear results) 140-0 (14 :dl ;t, with ideas of i.a2-bl, is more forceful) 14 ... .td6 IS .te2lLleS 16 .tg3 .td7 17 :ac 1 :acS = Yakovich-Lputian, USSR Ch (Kiev) 19S6. 4) 11...WaS?! is premature. 120-0 .te7 13 b4 WhS (l3 .. :ifb6 14 :fdl Wa7 IS e4 ± Maiwald-Masserey, AItensteig 1994) 14 .te2 WfS IS .td3 WhS 16 lLle2 eS 17lLlg3 Wg41SlLlxeS winning a clear pawn, Skalik-Jaworski, Poland 1996. The queen wasted too much time. 12:d1 1) 12 O-O!? bS 13 :fdl WeS 14 .ta2 .tb7 IS :acl :d8 16 .tbl :xdl+ 17 :xdl WcS 18lLlgS g6 19 lLlce4 ± Lechtynsky-Tozer, CappeUe la Grande 19S9. 2) 12 i.d3 h6 13 0-0 .td7 (or 13 ... lLlhS 14 :fdl ± Adorjan-P.Nikolic, Vdac 1983, reached via 11 .td3 h6 12 "'c2 .te7 13 0-0 lLlhS) 14 h3 :'c8 IS e4 bS 16 eS lLldS 17 lLlxdS exdS 18 1i'd2 ;t Zlochevsky-Arlandi, Fonnia 1995. 12..:.&5 13 0-0 (D) Results favour White, not surprisingly considering his more active minor pieces in an otherwise symmetrical position.

Alternatives to the Main Lines for White

13.•.i.d7 1) 13 ...bS 14 i.d3 i.b7 IS ttle4 h6 16 ttlxf6+ i.xf6 17 i.h7+ Iii>h8 18 .l:td7 eS 19 i.xh6 ttld8 20 i.gS +Stohl-Cvitan, Polanica Zdroj 1985. 2) 13 ...eS 14 i.gS i.g4 IS i.xf6 i.xf3 16 gxf3 i.xf6 17 l:ld7 e4 18 'it'xe4 i.xc3 19 bxc3 l:lac8 20 i.dS 'ifxa3 21 c4 ± Stohl-Dobrolowsky, RimavsIca Sobota 1990, illustrating yet again the potential dangers on the light squares after Black has played ...eS. 3) 13 ...h6 and then: 3a) 14 i.a2 i.d7 (l4 ...ttlhS!?) IS ttld2 ± Miles-Ivkov, Vrbas 1980. 3b) 14 e4 ttlg4 (14 ...eS IS i.d2 'ifc7 16 ttldS ttlxdS 17 exdS ttld4 18 ttlxd4 exd4 19 'ili'd3 'ili'b6 is unclear according to ECO, but White would seem to have a useful lead in development) IS i.a2 i.f6 16 h3 ttlgeS 17 ttlxeS ttlxeS 18 i.e3 bS 19 i.cs liz-liz Stohl-Kotronias, Khalkidhiki 1992. 3c) 14 h3 l:ld8 IS i.a2 i.d7 16 ttld2 i.e8 Stohl-Gavrilaikis, Haifa Echt 1989. White now unnecessarily contented himself with 17 i.bl 'ili'hS 18 ttlce4 t, whereas 17 ttlc4! 'ifhs 18 ttlb6 l:lxdl (18 ...:a7 ±) 19 ttlxdl and then 19 ...l:ld8? 20 i.c7 would have trapped Black's rook in broad daylight.

109

Naturally 19... l:la7 can be tried, hoping for something like 20 ttlc8 l:la8 21 ttlxe7+ ttlxe7 22 'ifc7 ttlfdS! 23 'ili'xb7 i.c6, but after 20 'ifb3 Black's position is miserable. 14 ttlgS 14 e4 i.e8 IS :fel bS 16 i.a2 l:lc8 17 i.bl ttlhs 18 eS g6 19 i.h6 b4 20 i.xfS bxc3 21 i.xe7 ttlxe7 22 i.a2 i.a4 liz-liz Pekarek-Geller, Dortmund 1992. 14•••lUd8 IS ttlds!? A familiar tactical idea from the New Main Line. IS•••exdS 16 l:lxdS .b6 In comparison with the New Main Line, Black has the extra move ... a6, which prevents White from ever playing l:lbS. However, White's king is much safer on gl than on cl, giving White a little more time to mature his attack, without having to wony about counter-tactics. 17 l:lxd7 l:lxd7 18 i.xf7+ 1ii>f8 19 i.a2 g6 20 ttlxh7+ Iii>r:T 21.lbl1ii>xh7 22 .xg6+ Iii>h8 23 .h6+ Iii>g8 24 i.a2+ ttldS 25 .g6+ Iii>h8 26 .hS+ Iii>g8 White has a decisive attack, Efimov-Dutreeuw, Asti 1995. Now 27 e4 'ili'xb2 30 exdS .xa2 32 dxc6 wins, though in the game, White repeated moves before playing this line. D14) 10•••b6 The problem with this move is that the black queen runs dangerously short of squares. 11 .c2 i.b7 12 l:ldl 13 0-0 i.e7 13 ... i.a6? 14 i.xa6 'ifxa6 IS b4! spears a piece.

.cS

110

The Queen's Gambit Declined: 5 i.f4!

14 i.a2 .:r.d8 15 i.b1 i; Y.KozlovShuraev, Tula 1995. The general conclusion concerning 9 ... dxc4 10 i.xc4 is that Black can equalize very easily with 10 ... ~h5, but the position is dull and drawish. Alternatives to 10 ... ~h5 leave White with good prospects of an advantage. If Black wants to keep options of playing for a win, he should vary at move 9. 02) 9 ...a6 (D)

lOadS It is of course possible to transpose

to the previous line, should both sides desire, with 10 Wc2 dxc4 (or, less satisfactorily, 1O... Wa5?! 11 0-0 dxc4 12 i.xc4 Maiwald-Masserey, Altensteig 1994) 11 i.xc4. 10... exdS 1O... ~xd5 11 ~xd5 exd5 is like the Knight Exchange Variation, discussed in Chapter 4.4 Line B 1, except that Black has played ... a6 prematurely, allowing White time to develop his kingside. Goldin gives 12 ~e5 ~xe5 13 i.xe5.:r.e8 14 i.d4 i.xd4?! (14 ... i.d6

:j:) 15 Wxd4 .:r.e4 16 'ii'd2 d4 17 i.f3 l:te8 18 l:tdl ±. 11 i.gS If 11 0-0, simply 11...d4 = GoldinBeliavsky, Yugoslav Cht (Tivat) 1995. 11...d4 12 i.xf6 12 ~e4 i.e7! (12 ... Wa5+?! 13 b4 ~xb4 14 axb4 Wxb4+ 15 ~ed2) 13 ~xf6+ i.xf6 14 i.xf6 'ii'xf6 =. 12..:it'xf6 13 lbe4 We7 14 ~xeS 'it'xeS 15 ~xd4 .:r.d8 16 'it'c2 Wxc2 17 ~xc2 i.fS 18 .:r.c1.:r.ac8 19 g4 i.g6 20 f3 Thus far Psakhis-Sitnik, Portoro~ 1995. Now 20... ~5 21 e4 ~xg4! 22 0-0 .:r.d2 23 ~4 .:r.xcl 24 .:r.xcl h625 .:r.dl, as given in Informator, leads to a drawn endgame. Perhaps 10 Wc2 is to be preferred.

03) 9•••.:r.e8 Black has in mind the ...e51...d4 push. 10 i.gS Mter 100-0: 1) Black wasted a tempo with 10... h6?! in Miles-Lobron, Reggio Emilia 1985, and was soon forced on the defensive: 11 b4 i.f8 12 i.g3 a6 13 Wc2 i.d7 14 .:r.fdl ±. 2) Black can simply play 10... e5! 11 i.g5 d4!, when 12 ~e4 i.e7 13 i.xf6 i.xf6 is assessed as unclear by Sokolov. Ifinstead 12 ~d5?!, Sokolov analyses 12 ... d3! 13 i.xd3 (13 'ii'xd3 e4 14 Wc3 ~xd5! 15 cxd5 Wxd5 16 i.f4 exf3 17 i.xf3 Wf5 18 g4 Wg6 19 Wxc5 i.xg4 +) 13 ... e4 14 i.xe4 ':'xe4 15 ~xf6+ gxf6 16 'ii'xd8+ ~xd8 17 i.xf6 :j:. These lines feature some tactics reminiscent of the Old MainLine.

Alternatives to the Main Lines for White Therefore White contents himself with a quieter move. 10.•. ~e7 11 'ii'c2 h6 12 ~h4 It is not worth isolating Black's dpawn when ... d4 can be achieved so easily. 12•••dxc413 .i.xc4.!Dd5 14.i.xe7 14 .i.xd5 exd5 15 ~xe7 lIxe7 16 0-0-0 ~e6, given as unclear by Sokolov, could be critical. 14...lDcxe7 15 0-0 lDxc3 16 'ii'xc3 'ii'c7 Now 17 lDd4 ~d7 led to equality in I.Sokolov-Lobron, Debrecen Echt 1992. 17 lIac 1 (Sokolov) preserves White's slight edge in piece mobility. D4) 9...lbe4!? (D)

Although this is largely ignored by theory, Black can improve on previous play. This variation might just hold the answer to how Black can play for a win against the ultra-solid 9 .i.e2 variation. 10~xe4!

10 "c2lDxc3 11 "xc3 d4 12 exd4 lDxd413 lIdllDxe2 14 ~xe2"b6 15 b4 ~e7 is given by Dreev, who does not provide an assessment. Unclear

111

perhaps? The position needs further testing. 10...dxe4 l1lDe5 lDbS! Black runs out of tactical tricks after 11.. ...a5+? 12 b4 lDxb4 13 axb4 "xb4+ 14 ~f1 f6 15 lIbl 16

lbd7.

"a3

The text-move is a more sophisticated way of trying to exploit the lack of escape squares for White's minor pieces. Undeveloping to b8 puts the knight on the only square where it does not get in the way; if l1...lbe7?, then 12 "xd8 lIxd8 13 b4 ~b6 14 c5 ~c7 15lDxf7. Only 11..:"f6 12 lDxc6 bxc6 13 "c2 has been previously seen. Then 13 ... a5 14 "xe4! (14 O-O?! e5 15 ~g3 :e8 16 b3 ~f5 17 "b2 "e7; MuseVan der Sterren, Altensteig 1991) 14.....xb2 15 ~e5 f5 16 "f4 "c2 17 ~f3 ~a6 18 0-0 ± Dreev, or 13 ...e5 14 ~g3 ~f5 15 0-0 (15 b4! Dreev) 15 .. :"e7 (l5 ... a5 16 'ifc3 ± Dreev) 16 b4 ~b6 17 c5 .i.c7 18 lIfdl ± DreevKhalifman, Linares 1995. 12b4 Preparing a retreat on c4. 12 'ii'xd8 :xd8 13 b4 ~e7 14 c5 a5!, challenging the queenside pawn majority, is unclear, but probably satisfactory for Black. 12....i.e7 13 c5 f6! 13...a5 14lbc4 axb4 15 axb4 lIxal 16 'ii'xallDa6 17 'ii'bl ±. Other lines are also good for White. 14lDc4 e5 15 ~g3 15 'ifxd8 ~xd8 16 .i.g3 b6 transposes. 15... b6 16 cxb6 axb6 17 'ii'xdS ~xdS Islbd6 ~e6 Leading to unclear play after 19 0-0 ~e7 20 ~c4 ~xc4 21lDxc4 b5. Not,

The Queen's Gambit Declined: 5 j"f4!

112

however, 19 ~xe4?? fS followed by ...f4, trapping White's bishop. 05)

9•• ~e7 (D)

w

11 ~c6 bxc6 12 i.xd6 'iIIxd6 13 0-0 Szymczak-Abramovic, Polanica Zdroj 1983. Black's position looks OK until one asks how he is going to advance any of his central pawns safely, ...cS and ...eS both being answered by cxdS, leaving White with clearly the better pawn structure. And if the pawns remain immobile, White can pressurize pawns on the light squares, and seek outposts, notably cS, on the dark squares. ECO gives it as ± after 13 ...:d8 14

':'cl.

10 lbes 10 0-0 dxc4 11 i.xc4 c!DhS! is just as solid as in the 9 ...dxc4 line, e.g.: 1) 12 'ffxd8 :xd8 13 i.c7 :d7 14 i.eS ~xeS (14 ... gS!? IS h3 c!DxeS = KneZevic-Karpov, Leningrad 1977) IS fOxeS IIc7 16 fObS :cS 17 f4 a618 c!Dd4 c!Df6 Lputian-Aseev, Irkutsk 1986. The position is perhaps slightly deceptive; it looks as though White is building up a significant space advantage, but Black has no real weaknesses to attack, and there are some potential holes in White's position that need watching. 2) 12 'ii'e2 ~xf4 13 exf4 ~d4 14 c!Dxd4 'ii'xd4 IS g3 i.f6 16 lIfdl 'ii'b6 17 :ac1 a6 18 i.a2 :d8 19 ~4 i.e7 20 ~3 i.f6 1/2- 1/2 Lin Ta - SantoRoman, Lucerne Wcht 1985. 10 'ii'c2 transposes to Chapter S.I, Line H. 10•••i.d6 10...~xeS!? lli.xeS dxc412i.xc4 'ii'aS is safer.

=

06) Finally, a couple of 'played-onces' from several years back: D61: 9•••i.d6 112 112 D62: 9•••d4?! 061) 9... i.d6Ied to 'orthodoxy' in Szekely-I.Zaitsev, Moscow 1982: 10 i.gS i.e7 11 0-0 dxc4 12 i.xc4 h6 13 i.h4 a6 14 :cl 1/2- 1h. White could also try 10 i.xd6!? with possibilities of play against the isolated d-pawn. Analogous lines of play are discussed under 9 'ii'c2 i.d6 in Chapter S.I, Line E. 062) 9...d4?! is premature. 10 ~a4 i.d6 11 i.xd6 'ii'xd6 12 ~xd4 ~xd4 13 'ii'xd4 'it'c6 14 ~3 'ii'xg2 IS 0-0-0 eS 16 'ii'xeS 'ifxt2 17 :dfl 'ii'h4 18 :hgl ~h8 19 ~S :e8 20 'ii'c3 ± Szymczak-lasnikowski, Poland 1982.

General conclusion to 9 i.e2 The move is extremely solid, and there are few ways for Black to seek any sort

Alternatives to the Main lines for White of complications, the relatively unexplored 9 ... ~4 being perhaps the major exception. If, however, Black is content with a drawish position, with 9 ...dxc4 10 .i.xc4iOh5, there is little White can do to try for an advantage.

4.4 White's alternatives on move 8 1 d4 dS 2 c4 e6 3 1Dc3 iOf6 4iOf3 .i.e7 S .i.f4 0-0 6 e3 cS 7 dxcS .i.xcS (D)

113

independent possibility, unpromising for White, is 8 .i.d31Oc6 9 0-0 We7 10 cxdS exd5 11 .i.g5 :d8 12 :cl h6 13 .i.h4 a6 14 :el g5 15 .i.g3 lOe4 16 iOd2 .i.b4 :j: Tjiam-Vaganian, Dutch Cht 1997. Developing the bishop to e2 is generally more flexible if the central pawn formation has not been clarified. We note also in passing the game Gabler-Zlatilov, St Ingbert 1990, which ended tamely after 8 .i.g5 .i.e7 9 .i.e2 dxc4 10 .i.xc4 a6 11 0-0 b5 12.i.b3 .i.b7 13 'ife2iObd7 112_112. Our lines are: A: 8.i.e2 113 B: 8 adS 122

A) 8 .i.e2 (D)

We now consider lines where White plays neither 8 'ifc2 nor 8 a3 here. Any line not already considered where White plays 8 Wc2 or 8 a3 will be covered in Chapter 5, under 'Alternatives for Black'. The most important alternative here is 8 cxd5, the 'Exchange Variation'. However, it makes for better continuity in the text to start with 8 .i.e2. We have just looked at 8 a3 iOc6 9 .i.e2, but is a3 really necessary in this line? 8 .i.d3 and 8 :cl could well transpose into the 8 .i.e2 line after 8...dxc4 9 .i.xc4 'ifxdl+ 10 :xdl, although clearly 8 Acl discourages alternatives to the queen swap on move 9. An

This ought in principle to have more bite than the 8 a3 1Oc6 9 .i.e2 system. After all, from the diagram 8...iOc6 is a common enough move, and few would consider 9 a3 as the most dangerous reply. The only drawback to 8 .i.e2 instead of 9 .i.e2 is that Black then has the possibility of developing his queenside more flexibly with ... a6, ... b5, ... .i.b7 and ... iObd7. Even so, White's extra tempo, or

114

The Queen's Gambit Declined: 5 1./41

half-tempo, is arguably more significant. The main lines are: AI: 8...dxc4 114 121 A2: 8...lLlc6 First, one minor alternative for Black may be noted: 8.....e7 9 cxd5 :d8 10 0-0 lLlxd5 I1lLlxd5 :xdS 12 .b3lLlc6 13 :adl :xdl 14 :xdl is a rather time-consuming way for Black to clear the central pawns. After 14...b6. 15 lLlg5!? looks strong (instead of 15 a3. Stahlberg-Chekhover, Moscow 1935).

A1)

8...dxc4 9 .ixc4 Again one must ask whether Black can equalize by exchanging queens. If not. 9 ... a6 or 9 ...lLlc6 comes into consideration. The immediate 9 ... lLlh5?! is not this time an antidote, with the other knight not yet in the game. e.g. 10 .ig5!? .ie7 11 Wxd8 .ixd8 12 .ixd8 lhd8 13 ~e2 and White is much better mobilized for any endgame. We consider: All: 9 ....xd1+ 114 All: 9 ...a6 116 A13: 9"'lLlc6 119

A11) 9....xd1+ 10 :xd1 (D) 10...a6 Again envisaging a development with the knight on d7. 1) 1O... .td7?! 11 lLleS .ic6 relinquishes the bishop-pair much too easily: 12 0-0 a6 13 :d2 b5 14 lLlxc6 lLlxc6 15 .ie2 ±. 2) 1O... lLlc6 11 0-0 is clearly more favourable for White than in the 8 a3 lines.

2a) If then 1l....id7. 12 e4 is a lot more testing than the limp 12 .id6?! .txd6 13 :xd6 :fd8 = of AugustinFaibisovich. Brno 1991. 2b) 11...b6 12 e4 .tb7 13 eSlLlaS! is a more interesting attempt to take advantage of White's omission of a3. After 14 exf6 lLlxc4 15 b3 .ib4 16 lLlbS lLla3 17 fxg7 ~xg7 18 lLlc7 .txf3 19 gxf3 :ad8 White played for control of the d-file with 20 .tg5 in Ibragimov-Lputian, Vienna 1996. but after 20...:xdl 21 :xdl ~g6 22 .th4 lLlc2 23 ~f1 :c8 Black had counterplay. Targeting Black's minor pieces with 20 .ieS+ ~g6 21 .ib2 is more threatening; White should be better.

ulLleS! Given the 'N' symbol in lnformator 56, but actually played by Nimzowitsch in the 1930s. True, Nimzo contrived to lose the game ... Alternatives pose Black few problems: 1) 11 0-0lLlbd712lLleS bS 13 .td3 lLlxeS 14 .ixeS .te7 IS a4 b4 16lLle4 .ib7 17 .id6 .txd6 18lLlxd6 .tc6 19 a5 :fd8 =Zsu.Polgar-A.Mari~. Novi Sad wom OL 1990. 2) 11 .td3 lLlbd7 12 :cl b6 13 lLle4 .ib7 14 lLlxc5 lLlxc5 15 .te2

Alternatives to the Main Lines for White

=

115

IICc8 160-0 ttld5 Donner-Benko, complete mobilization of the black Wijk aan Zee 1972. forces. If enough pressure is exerted 3) 11 a4?! ttlc6 12 ~e5 and now, on the black position, it should be posrather than 12... ~xe5 13 .*oxe5 with sible, if things go smoothly, to force an identifiable weakness without allowequality, M.Markovic-Carlhammar, Stockholm 1987, 12...~h5 would have ing any corresponding increase in activity among the enemy pieces. heen more enterprising. Yet in the game Nimzowitsch4) 11 ~e2 b5 12.*ob3 .*ob7 13 .*od6 ixd6 14 IIxd6 ~6 15 IIhdl :Cd8 = Stablberg, Gothenburg 1934, Nimzowitsch was perfectly happy to exBobotsov-Sobhani, Teheran 1991. change his excellent bishop on eS for a 11••• ~bd7 12 .i.e2! not particularly active knight in order Forestalling ...bS. 12••• ~xe5 13 .*oxe5 bS 14.i.f3 lIa7 to inflict what is in context a trivial (f)) pawn weakness. He was soon overrun by Black's piece-play on the queenside. The game went 15 .*oxf6? gxf6 16 ~4 .i.e7 17 g4.*ob7 18 ~e2 .*od5 19 b3 IIc8 20 IId2 ~f8 21 IIhdl :ac7 22 IId4 a5 23 ~d2 .*oxf3+ 24 ~xf3 IIc2 25 a4 bxa4 26 bxa4 lIa2 27 ~e4 IIcc2 28 IIld2 f5 0-1. In fairness to Nimzowitsch, this game was played in his last tournament, and by then his playing strength had declined. Now let us see the modem treatment with 15 We2. IS•••.i.d7 15 ....i.b7 16 .*oxb7 IIxb7 17 .i.d6 ;!; lS~e2! Those who fondly look back to the Malaniuk. 16 lIel! b417 tDdS! J(lIod old days in chess tend to argue Ihut while our knowledge of openings Leaving Black only one way to premight well have increased immensely serve the bishop-pair. The sequence that follows is forced. liver the last sixty years, our under17•••.*obS+ 18 ~d2 ~7 19.*od4 ~lIl1lding of the inner spirit of chess has nlll; a Capablanca is at least equal to a .*oxd4 20 ~e7+ Wh8 21 exd4 ~b6 22 Knsparov, with corresponding rela~6 IId7 23 b3;t Malaniuk-Arlandi, IllIns applying to lesser players. This Forli 1992. In formal terms, White has even pllsition suggests otherwise. It is an Ilpen position, with White having the slightly the worse pawn structure, but ,,,Ivantage in piece mobility, and the there is little doubt that he has more nnlural way for White to play it would than enough compensatory piece acIll' In aim to preserve and enhance this tivity as a result of the minor tactical ,,,Ivantage by struggling to prevent the skirmishes. The pawn that is most

116

The Queen's Gambit Declined: 5 1./4!

likely to drop for nothing is not the white d-pawn, but rather Black's bpawn. A12) 9 •••a610 0-0 (D)

~xh7 16 'ilfxd7 lIad8 17 'ilfc7 'ilfxc7

18 .i.xc7 lIc8 19 lDxb5 .i.xe3 20 fxe3 axb5 21 b3 f6 22 .i.b6 lIa8 23 lIc7 .i.d5 1/2- 1/2 Lechtynsky-Efimov, Prague 1985. Black's active bishop-pair enabled him to hold the position comfortably despite losing a pawn; White found no alternative but to allow an opposite-coloured bishop ending. 2) 12 lDg5 lDbd7 13 l&e4 (13 lDge4 b4 14 lDxf6+ lDxf6 15 lDa4 .i.e7 =Zsu.Polgar-Schiissler, Vejstrup 1989) 13 ... h6 14 lDxf6+ lDxf6 15.!Do 'ilfb6 16 'ilfe2 and now, in Kacheishvili-Myc, Zaganjr Wch 1997, Black's 16... .i.e4?! 17 .i.xe4 lDxe4 18.!De5 ~ was unnecessary; centralizing the rooks would hold the balance. 3) 12 e4 is probably best answered with 12... lDh5, by analogy with the Smyslov-Kasparov game discussed in the note to move 13. Instead 12... b4 13 lDa4.i.e7 14 'ilfc2 lDbd7 15 .i.c7 'ilfc8 16 lIacl .!De8 17 .i.a5 'ilfb8 18 b3 is messy, possibly favouring White, A.Eriksson-I.Almasi, Budapest 1994.

=

10 'ilfe2 almost invariably transposes; White's standard plan involves 0-0, 'ilfe2, .i.d3, completing his development, and readying himself to initiate play on either side of the board. 10...bS For 1O...l&6, see 9 ...l&6 10 0-0 a6, Line A13 below. ll.i.d3.i.b7 Black can also tease White with l1...lDbd7. After 12lLle4 .i.b7 13 lDxc5 lDxc5 14 .i.e2 .!Dd5 15 .i.g3 lDe4 16 .i.h4 'ilfb6, Danielian-Rustemov, Moscow 1996, the concessions that White has made to preserve his bishop-pair are not worth it.

B

12"e2(D) Alternatives generally involve targeting h7: I) 1211cllDbd713 lDg5!? (13 'ilfe2 reverts to the main line) 13 .....b6!? (13 ... .i.e 7 14 lDce4 lDxe4 15 .i.xe4 .i.xe4 16 lDxe4 ~ Efimov-Ziatdinov, Lenk 1991) 14 lDxh7.!Dxh7 15 .i.xh7+

12...lDbd7

1) 12 ... h6?! should be too slow, if there is any justice in chess, although it does prevent any awkward .i.g5 pins.

Alternatives to the Main Lines for White 13 lIfd 1 'ile7 14 e41Dbd7 IS eS lDdS 16 "'e4 fS 17 exf61D7xf6 co KishnevNegulescu, Cappelle la Grande 1993, hut White has scope for improvement - maybe 13 lIacl ;t. If Black plans to play .. :ile7 anyway, it is best to put the rook on its best square immediately, rather than 'gain' a tempo by attacking the queen. 2) 12 ....te7 13 ':fdl "b6 is potentially transpositional, but there are independent possibilities: 2a) 14 .tg3 lDbd7 (14 ...1Dc6!?) Icads back to the main line. 2b) White could, however, try to tuke advantage of the early withdrawal of the bishop from cS by playing 14 e4!'!, and if 14... lDhS, IS .te3 with a critical gain of tempo when compared with 12 ... lDbd7 13 e4?! lDhS 14.ie3 (sec Smyslov-Kasparov below). If 15 .....c7, 16 eS is extremely dangerous; Black's kingside lacks any realistic defences. 2c) Instead, in the game M.Gurevich-Dutreeuw, Antwerp 1993, White played analogously to the main line with 14 a4 b4 151Dbi (Gurevich notes II possible queen sacrifice after IS as ~c5 l61Da4 "xaS 17 ':dcl ':c8 and then 18 lDcS ':xcS! 19 ':xaS ':xcl+ 20 lDel1Dc6, and Black has the initiative; 18 lIxc8+ .ixc8 191DeS is tempting at first, but there is no obvious hrcakthrough after 19....ib7) IS ...~bd7 16 lDbd2 lDcs 17 as "a7 18 .ic2 li)fd7 with chances for both sides. 3) 12 ... b4 13 lDa4 .ie7 14 lIacl li)dS IS .ig3 and now 16 'ilc2! (lUght to be good for White, the tempo ~uined by threatening h7 giving time for e4 ± next move. Instead, 16 b3? li)d7 17 lIc4 lIfc8 was played in

"as

117

Wedberg-Van der Wiel, Stockholm 1987 and the unnecessary hole on c3 cost White dearly. 4) 12... lDc6 leads to a position normally reached via 8...lDc6 9 0-0 dxc4 10 .ixc4 a6 11 "e2 bS 12 .tb3 .ib7. Then: 4a) Razuvaev-Geller, USSR Ch 1983 was drawn quickly after 13 :adl "b6 (this position actually arose via 12 lIadl 'ilb6 13 .id3 .ib7 in the 8 ...lDc6 move order) 14 a3 lIfd8 15 .igS .ie7 16 h3 liz-liz. 4b) Zsu.Polgar-Sarapu, Wellington 1988 did not last much longer: 13 lIac 1 'ilb6 14 .igS (if White wants to try for more, 14 lDgS! is the way White takes advantage of the fact that after ...1Dc6 the king's knight is no longer supported by its colleague) 14....ie7 IS lIfdl lIad8 16 h3 h6 17 .if4 1Db4 18 .ibl lIxdl+ 19 lIxdl lId8 20 lDes lIxdl+ 211Wxdl "'d822 "xd8+ 11z-112. 1311fdl (D) 1) 13 e4?! lDhS! 14 .id2 (14 .ie3 'ilb6! IS :fel .ixe3 16 "xe3 lIfd8 :;: Kasparov) 14.....c7 was already comfortable for Black in Smyslov-Kasparov, Vilnius Ct (4) 1984. After 15 g3, the main problem for Black is to avoid sacrificing prematurely; Kasparov notes IS ...lDxg3? 16 hxg3 "xg3+ 17 ~hl "h3+ 181Dh21DeS 19 f3 ±. Play continued instead IS ...lIad8 16 .ie3 (16 .igSlDxg3! 17 hxg3 "xg3+ 18 ~hl "h3+ 191Dh2 .id6 20 f3 f6 21 .ie3 lDeS 22 .ic2 lDc4 23 .igl .ieS with a strong attack - Kasparov) 16....ixe3 17 'ilxe3 "cS :;:. Kasparov also gives a couple of move IS alternatives, showing that Black remains untroubled after either IS eS? .ixf3 16

118

The Queen's Gambit Declined: 5 ~f4!

.xf3 g6 + or 15 b4 Ab6 (15 ... Axb4 161Oxb5.aS17iDd6Axd218lDxb7 lOf4! 19 .xd2 .xd2 20 lOxd2 lOxd3 is a complicated liquidation to equality) 16 a41Of4 17 Axf4 .xf4 18 axb5 axb5 19 Axb5 lOe5 with an ample pawn's worth of piece-play. 2) 13 :acl Ae7 (l3 ...•e7 14 Ag5 :fd8 15 Ae4 Axe4 161Oxe4 Ab6 17 :c6 ;t F.Olafsson-H.Meyer, Reykjavik 1984) 14 :tdl .b6 15 a3lDc5 16 Ac2 :fd8 171Oe5 1/2- 1/2 Gulko-Pigusov, Biel IZ 1993.

13.....b6 Black's most popular choice in recent years. Games featuring 13 ...•e7 or 13 ...h6 are generally slightly older. 1) 13 ...•e7 and now: la) 14 Ag5 and then: lal) 14...:ac8 15 a4 b4 16 lObi (161Oe4 Axe4 17 Axe4 h6 = Gurevich) 16... aS 171Obd2 h6 18 Ah4 i.b6 19 Ae4 ;t M.Gurevich-Beliavsky, Moscow 1990. If 19...Axe4, Gurevich gives 20 lOxe4 g5 21 lOfxg5 lOxe4 22 :xd7 .xd7 23 lOxe4 Ad8 24 Af6 ~h7 25 '6'h5 .c6 26 f3 with a dangerous attack. la2) 14... h6 15 Ah4 .e8!? (or 15 ...:fc8 16 Ae4! lOb6 17 i.xf6 gxf6

181Od4 f5 19 i.xb7 .xb7;t Gomez Esteban-A.Hoffman, Salamanca 1991) 16 :acl Ae7 17 i.bl :c8 18 a3lDc5 19 lOe5 /Dd5 20 .c2 g6 21 Axe7 Ih.-Ih. Kaidanov-Van der Sterren, Budapest 1989. The loss of tempo with the queen looks strange at first, but White's bishop has been lured to a less active diagonal. Ib) 141Od2!? is a more recent try for White, which cuts across the ...•e8 defensive plan, and aims to take advantage of the lack of retreat squares for the black bishop. Blees-I.Ivanov, Gausdal1993 continued 14 ...:ac8 15 lOde4 /Dd5 16 lOxd5 exd5 17 lOxc5 lDxc5 18 Af5 iDe6 19 i.e5 f6 20 i.c3 ;to This might well be promising. 2) 13 ... h6 14 Ag3 Ab4 15 iDe5 .e7 16 lOxd7 lOxd7 17 a3 i.xc3 18 i.d6 .h4, and in view of 19 AxfS1! Ae5 20 f4 :xfS 21 fxe51Oxe5, when Black has excellent compensation for the exchange, White played 19 bxc3 and agreed a draw in M.GurevichYusupov, Munich 1992. It is of course open for White to try to improve. For example 14iDe5looks more economical than 14 Ag3. 14i.g3 1) 14 a4!1 b4 15 lObi (15 as .c6! 16 iDa4 e5 17 lOxc5 .xc5 18 i.f5 exf4 19 i.xd7 fxe3 ~ M.Gurevich; one of the points of 14 i.g3 is to prepare a4-a5 without allowing this type of possibility) 15 ...:fe8 (15 ...i.e??! 16 lObd2 :fc8 17 lOc4 ;t E.ShvidlerCiolac, Silvaplana 1993) 16 i.g3 e5 17 Ac4 h618 as .a7 191Obd2 e4 20 iDel AfS (20... iDe5 21 Axe5 :xe5 22 lOb3 i.fS transposes) 211Ob3 iDe5 22 i.xe5 :xe5. An Informator reference tersely gives 23 lDc2 'iVb8 24 :d4

Alternatives to the Main lines for White Cifuentes-Van der Sterren, Dutch Ch (Eindhoven) 1992, while ChekhovGorelov, Moscow 1995 was agreed drawn after 23 l:td4 l:taeS 24 ~2 .i.cS 25 "d2 "bS 26 l:ta4 l:thS 27 g3 .i.g4 2S iClxb4 .i.f3 29 .i.fl "eS 30 .i.g2 .i.xg2 31 Citxg2 "fS 32 h4 "f3+ 33 Citgl l:txh4. So where does the truth lie? Probably Black should meet 23 ~2 with 23 ...l:laeS; 23 ... ~S unnecessarily blocks off the rook. The bpawn will almost inevitably fall, but Black has good chances of achieving a compensatory kingside attack. 2) 14iClgS eS IS .i.g3 b4 16 ~e4 h6 = Ibragimov-Sturua, Peristeri 1993. 3) 14 l:tac1 l:tac8 ISiClgS l:tfd8 16 iClge4 (16 .tbl.te7 17iClge4 b4 =Malaniuk-Heine Nielsen, Katowice 1993; 16 ~e4?! h6; Rustemov) 16....te7 17 .td6 CitfS IS .txe7+ Q;xe7 19 iClxf6iClxf6 20 b4 (20 a3 =; 20 e4?! e5 21 iCld5+ .txd5 22 exd5 l:txc1 23 "xe5+ "e6! ; Rustemov) 20... l:lc7 21 "b2 l:tcd7 22 .te2 l:txd 1+ 23 :xdl "c6 24 .tfl 112_1/2 Kiriakov-Rustemov, Russian Ch 1997. Several minor transpositions of move-order are possible in this line, but the impression given in both '2' and '3'isthatanearly iClg5 is not all that threatening if Black has played ...iClbd7. 14..•.te7 14...:fcS!? - Zsu.Polgar. IS a4 b4 15 ... bxa4 16iClxa4 ;t Zsu.Polgar. 16aS"d8!? 16... 'ii'c6? 17iCla4 l:tfcS IS e4! left Black in a bind in Zsu.Polgar-Geller, Aruba 1992. 16.....c5!? 17 ~bl "h5 IS iClbd2 lCJcS 19.i.c4 l:tfc8 (19 ...lCJfe4 20 iClxe4 iClxe4 21 l:td7iClxg3 22 hxg3 .txf3 23

119

gxf3 .tf6 24 .txa6 ;!; Gurevich) 20 ~d4 "g6 (20 .....xe2 21 .txe2iClfe4 22 ~4;!; Gurevich) 21 f3iClfd7 22 e4

h5! 23 .tf4 h4 aD and Black is well set for counterplay on the dark squares, M.Gurevich-Geller, Helsinki 1992. 17 ~bliClhS 18 iClbd2 18 .txh7+? Citxh7 19 'ii'd3+ Citg8 20 'ifxd7 Wxd7 21 l:txd7 l:tfdS 22 l:lxdS+ l:txdS 23 iClbd2 .i.f6 24 ~4 iClxg3 25 hxg3 .txf3 26 gxf3 l:tcS =+= Tregubov. 18 .te5 g6 19iClbd2iClxe5 20 lClxe5 "c7 21 ~ec4 iClf6 22 iClb6 l:ta7 23 ~c4 l:td8 24:ac1 Wc5! and Black's bishops gradually took control in Blees-Van der Sterren, Brussels 1993. 18... ~xg3 19 bxg3llkS 20 ~ Now Black got into trouble in Tregubov-Van der Sterren, Wijk aan Zee 1995 after20...iCJxd3? 21iClb6! ±.te4 22 lCJel WeS 23 iClxd3 :dS 24 lCJe5 .td6 25lCJed7! l:txd7 26iClxd7 Wxd7 27 Wxa6 l:tdS, when Tregubov gives 28 "b6 .tdS 29 a6 :bS 30 WaS, etc., as winning - a rare case of the knightpair dominating the bishop-pair in an open position. Tregubov suggests that the more patient 20..."eS! is equal. We are still a long way from being able to give a lasting verdict on the 9...a6 line. Maybe White achieves a slight edge, or maybe Black equalizes. What is clear is that both sides must manoeuvre their pieces carefully in positions where direct tactical threats are not far below the surface. A13) 9...~(D)

White could transpose to the 9 a3 line with 10 a3, but it is extremely unlikely that he would want to do so.

120

The Queen's Gambit Declined: 5 j.f4!

w

Emilia 199112. Substituting 0-0 for a3 (comparing this and the 8 a3 lLlc6 9 .te2line) makes it much less likely for Black to secure equality with an early

...lLlhs.

100-0 10 "e2!? and now: I) Black rashly went pawn-snatching in the game Khurtsidze-A.Koglin, Leon 1996: 1O...i.b4 II 0-0 i.xc3 12 bxc3 'iVa5 13 e4 "xc3? 14 :lac 1 "a5 15 e5 tLld5 16 i.d2 "d8 17 :lfdl i.d7 18 i.g5 ±. 2) 1O.....e7 is more cautious: 11 0-0 h6 12 i.g3 i.b4 (the whole idea of ...i.b4 seems a bit of a time-waster; practical experience is lacking, but maybe 12... a6 13 :lfdl 'ire8!? could be tried) 13 l:lfdl e5 1400 lLlxd5 15 i.xd5 i.g4 16 "c4 i.h5. Now 17 a3 .td6 18 i.xc6 :lac8 was M.GurevichShort, Paris rpd 1991, but 17 'ire4!? might be better; indeed it is hard to see how Black holds everything together. 10•••a6 I) 1O.....xdlll l:lfxdl b6?! 12lLlb5 i.b7 13 i.d6 i.xd6 14 lLlxd6lLla5 IS i.bS ± Karlsson-Balashov, Helsinki 1983. Even if Black can improve on move II, the immediate queen exchange is a little too slow. 2) 1O... tLlhS 11 "xd8 l:lxd8 12 i.c7 :ld7 13 i.eSlLlf6 14 i.g3 a6 15 :acl i.e7 16 i.e2 h617 h3 'iPfB 18 a3 :ld8 19lLleS lLlxeS 20 i.xeS i.d7 21 i.f3 ;!; Portisch-Vaganian, Reggio

3) 1O.....e7 11 i.gS (or II "e2, transposing into the Gurevich-Short game above) 11...h6 12 i.h4 :ld8 13 "e2 'iWfB and in H.Meyer-R.Wiemer, Bundesliga 199112 White played quietly and soon drew. Splitting the kingside pawns with 14 i.xf6 gxf6 ISlLle4 i.e7 16 :lac1, as in Parker-McLaren below, is a more attractive option. 4) 10... i.e7 II h3 a6 12 We2 bS 13 :ladl "b6 14 i.d3 i.b7 IS e4 b4 16 tLla4"a5 17 b3 :lfd8 18 "e3lLld7 19 :lcl :lac8 20 eS ;!; Ibragimov-Rustemov, St Petersburg 1996. l1:lct After II "e2 bS, 12 i.d3 i.b7 was discussed under 12...lLlc6 in the 9...a6 line (Line AI2 above), when 13 :lacl "b6 14lLlgS was recommended for White. Instead, 12 l:lfd1 13 i.d3 is less threatening: 13 ...lLlb4 14 lLle4 tLlxd3 IS tLlxf6+ gxf6 16 :lxd3 :ld8 V.Ragozin-Makogonov, USSR Ch (Moscow) 1944. 11••• b5 11...h6 12 tLleS tLlxeS 13 .txeS lLld7 14 i.g3 bS and now IS i.b3 i.b7 16 tLle2 might improve on IS i.d3, Zysk-Hochgrlife, Dortmund 1991. 1l.....e7 12 i.gS h6 13 i.h4 :ld8 14 'ire2 "fB IS i.xf6 gxf6 16lLle4 ;!; Parker-McLaren, British Ch (Eastbourne) 1991. 12 i.d3 lLlb4 13 i.bl 'irxdl 14 :lfxdl i.e7 15 lLle4 lLlbd5 16 i.d6 i.xd6 17 lLlxd6 :ld8 18 lLlxf7 'iPxf7 19 e4;!; Karlsson-Lechtynsky, Copenhagen 1983.

"b6

=

Alternatives to the Main Lines for White A2)

8...lDc6 (D)

With mainly transpositional effects. 9 cxd5 exd5 leads to the Pawn Exchange Variation, Line B3 below. 9 0-0 dxc4leads to variations just examined. 9 'ii'c2 'ii'a5 10 l:tdl is Chapter 4.3, LineC. The main non-transpositional possibility is 9 cxd5 tDxd5 10 tDxd5 exd5. First, we must deal with some odds and ends after 9 0-0. Our sections are thus: A21: 9 0-0 121 A22: 9 cxdS tDxd5 121 A21)

90-086 9...i.e7?! 10 cxd5 tDxd5?! 11 tDxd5 cxd5 12 'ii'b3 is uncomfortable for Black, Kiriakov-O.Danielian, USSRjr Ch (Alma-Ata) 1991. After 12... tDa5 White gets good play with either 13 .c2 or 13 'it'd3 i.e6 14 l:tac 1 (better Ih,lD the game's 14 i.e5 tDc615 i.c3). 10 cxdS tDxdS IO ... exd5 does not fit in well with .. 016. 11 ':cl i.a7 12 'it'b3 (after 12 l,i)c5!'!, 12...tDe7!? 13 'ii'b3 d4?! 14 J:lld 1 i.e6 15 'it'xb7 tDfd5 16 tDxd5

121

i.xd5 17 Wb4 +- was Kaidanov-Geller, New York rpd 1990, but 12...l:te8 13 tDxc6 bxc6 14 tDa4 i.b7 keeps Black alive, Stull-Arencibia, Moscow OL 1994) 12... tDa5 13 Wa3 i.e6 14 l:tfdl :lc8 15 tDd4 tDc4 16 'ii'b3 tDa5 17 Wc2 We7 18 Wbl l:tfd8 19 i.g5 ± Kolev-Polak, Vienna 1990. 11 tDxdS exdS It is not particularly clear how Black has gained by interpolating ...a6 in this line either. 12:lct White gains significant pressure against the isolani after either 12...i.b6 13 tDe5 tDe7 14 tDd3 i.e6 15 Wa4 l:tc8 16 'it'a3 l:tc4 17 i.d6 :le8 18 i.xe7 'ii'xe7 19 Wxe7 :lxe7 20 l:txc4 dxc4 (the pawn is weak even when it is no longer isolated) 21 tDf4 ± LobronRenet, Uzes 1990, or 12...i.a713 tDe5 tDe7 14 W'b3 Wb6 15 Wa3 :le8 16 l:tfdl "f6 17 tDo h6 18 i.e5 W'e6 19 Wb3 ± Kharitonov-Gavrilov, Smolensk 1992. This tends to suggest that 9 ...dxc4 is best. A22) 9cxdS tDxdS For 9...exd5, see Line B3 below. 10 tDxdS exdS (D)

110-0 In comparison with the Knight Exchange Variation (8 cxd5 tDxd5 9 ICJxd5 exd5), White is able to castle without having inserted a3. 11•.•i.b6 1) Abramovic analysed 11...d4?! to a big plus for White, after 12 Wc2 (12 l:tel i.b613 exd4 tDxd414 tDxd4 i.xd415 b3;t D.Gurevich-Frias, New York 1983, is far less challenging)

122

The Queen's Gambit Declined: 5 J.f4! 12•••"'d6 13 l:c1 .ie6 14 "'a4 h6 15 .if4 "'e716lLleslLlxe5 17.ixe5 White has the typical modicum of pressure against the isolated pawn, bu.Polgar-Lobron, Brussels 1987.

12.....e7 13lLlgS g6 (13 ... fS 14 l:ac1 i..b6 IS i..c4+ ~h8 16 exd4lLlxd4 17 "d3 ±) 14lLle4 and then 14... i..fS IS "xcS "xe4 16 i..f3 ±. Geller tried to resurrect Black's play with 14 ... i..b4, but after IS a3 i..fS 16 i..d3 i..xe4 17 i..xe4 i..d6 18 i..xd6 "xd6 19 l:fdl :ac8 20 "a4 l:fd8 21 l:acl the isolani was under mortal pressure in Av.Bykbovsky-Geller, Dortmund 1992. 2) 11...h6!? and then 12 l:cl i..b6 13 lLleS "f6! 14 .xd5 i..e6 15 "bS lLld4 16 exd4 "xf4 = was ConquestMotwani, British Ch (Swansea) 1987, while 12 .c2 i..b6 13 l:adl i..e6 14 lLle5 "f6 leads to a position assessed by Abramovic as equal, but maybe 12 "b3, angling for a transposition into Kaidanov-Abramovic, could be tried. 12i..g5 12 "c2 h6 = Abramovic; see above. 12 "b3 h6 (12 ...d4 13 exd4lLlxd4 14lLlxd4 i..xd4 15 :adl "f6 16 "g3 CID Abramovic) 13 lLle5 i..e6 14 l:adl d41S"a3 (IS"a4 "f6=Abramovic) Kaidanov-Abramovic, Vienna 1989, and now Abramovic suggests 15.....f6! 16 lLlxc6 bxc6 17 i..d6 l:fe8 threatening ... i..xa2. The text aims to keep the game quieter.

In conclusion, 8 i..e2 is a genuine try for an edge, even if only a modest one. White aims for less, and risks less, than in the New Main Line, but play is more lively than in the 8 a3 lLlc69 i..e2line. From Black's point of view, the most promising plan is 8 ...dxc49 .ixc4 a6 followed by ... b5 and a queenside fianchetto, with the knight going to d7. 8...lLlc6 gives White good prospects after 9 0-0. B)

8 cxd5 (D)

For convenience we call this the Exchange Variation, and hope that no one confuses it with the 'other' exchange variation (1 d4 d5 2 c4 e6 3 1Oc3lLlf6 4 cxd5 exd5, etc.). Black is not absolutely obliged to complete the exchange. since 8......a5. although tried only once in grandmaster chess, has not been refuted. I.Farago-l.Zaitsev, Sochi 1981 continued 9

Alternatives to the Main Lines for White .1d3lLlxdS 10 0-0 lLlxc3 II bxc3 j.e7 12 "c2 h613 :abl, when 13....tf6!?, instead of 13 ... lLld7 14 :bS 1id8 IS 1 i, is worthy of consideration. Farago in his notes suggests 9 dxe6 .1xe6 10 .td3 :d8 11 "e2 lLldS 12 0-0 ±, but it is not so clear why White should be so much better if Black continues 12 ...lLlxc3 13 bxc3 "xc3 14 :fdllLld7 - something for the adventurous to explore perhaps. If Black is to recapture, the choice is with the knight or the pawn. The knight capture leads to lively play, with White facing the problem that if he wants to castle kingside after 8...lLlxdS 9lLlxdS exdS he will have to spend a tempo on 10 a3. The Knight Exchange Variation has been popular in recent years, but as with most other white options in the .tf4 system, it has been seen a bit less often since the successes of h4 in the New Main Line. Black generally feels happier in the Knight Exchange Variation, where he has free development, lively pieceplay, and the opportunity to liquidate his isolani with ...d4, than in the Pawn Exchange Variation, 8 ...exdS, which leads to stodgy positions where Black has to stick his pieces behind the isolated pawn, and hope that White can't build up any initiative. Many of the games in the Pawn Exchange Variation were played with the move-order 7... lLlc6 8 cxdS exdS 9 j.e2 j.xcS, instead of 7 ....txcS 8 cxdS exdS 9 .te2 lLlc6. For some reason I've forgotten, and fail to fathom now, 7 ... lLlc6 was once regarded as slightly more accurate than 7 ....txcS; few now believe this.

:d

123

We divide our coverage as follows: Bl: 8.••lLlxdS 123 B2: 8•••exdS 13S B1)

8..•lLlxdS (D)

9lLlxdS Mter 9 .td3, 9 ...lLlxf4 10 exf4lLld7 II 0-0 lLlf6 12lLleS .id4 13 :c I .ixc3 14 :xc3 b6 00 Dementiev-Tavadian, Erevan 1983 is possible, but 9...lLlxc3!?, splitting White's queenside pawns, is a simpler route to equality. 9 •••exdS (D) As well as the Knight Exchange Variation, there is also a Queen Exchange Variation, with 9 .....xdS 10 "xdS exdS, agreed drawn immediately in Nogueiras-Damljanovic, Saint John 1988, and after 11 a3 lLlc6 12 :cl.tb613j.bSj.d7140-0:ac8IS :fdl j.e6 16 lLleS lLlxeS 17 .txeS :xc1 18 :xcl :c8 19 :xc8+ j.xc8 20 ~f1 in Ruban-Smirin, Novosibirsk 1995. Supposing that White is not content with a quick draw, then what? II O-O-O?! lLlc6! 12 :xdS .1b6 wins a pawn, but at the cost of development. Black's rooks and bishops will soon

124

The Queen's Gambit Declined: 5

swann around the white king after ....i.e6, ...:ac8, etc. For example 13 a3 .i.e6 14 :d2lDa5, or 13liJe5 .i.e6 14 lDxc6 bxc6. Maybe the most promising shot is 11 a3 lDc6 12 .i.b5!, waiting a little before deciding the best squares for the rooks. The attempt by Black to reach an opposite-coloured bishops position a pawn down is unconvincing: 12... d4 13 .i.xc6 bxc6 14 ~xd4 .i.xd4 15 exd4, and White's general plan will be to exchange a pair of rooks on the efile, and to target Black's weak queenside pawns.

w

Now White must decide whether to allow the bishop check with 10 .i.d3 .i.b4+, or whether to spend a tempo preventing it with 10 a3. Our sections: B11: 10.i.d3 124 B12: 10 a3 126 B11) 10 .i.d3 .i.b4+ An opportunity too good to miss. 1O.....e7?! 11 0-0 ~c6 12 :cl .i.b6, Ek-Nordstrom, Swedish Ch 1974, and now 13 .i.bl!? is slightly better for White. 11~e2

~f4!

1) II liJd2 takes the knight out of position: la) 11 .....f6120-01Dc6?! (it makes more sense to snatch the pawn) 13 a3 .i.xd2 14 Wxd2 .i.f5 15 .i.bS d4 16 .i.xc6 dxe3 17 Wd6 ;t; Granda-J .Armas, Palma Soriano 1986. Ib) 11...lDc6 120-0 .i.e7 13lDb3 (13 a3 .i.f6 14 "c2 h6 15 ~b3 :e8 16 :labl "e7 17 "cS liJes BlagojevicAbramovic, Nik~ic 1997) 13 ....i.f6 14 "d2 .i.e6 IS :ac1 :c8 16 ~cS We7 17 :c2 ~S Lputian-Timoshchenko, Irkutsk 1986. 2) 11 ~f1 has a venerable pedigree, O.Bernstein-Rubinstein, Ostend 1906 continuing 11.. ..i.e7 12 h4lDd7 (this was questioned by Bernstein, but 12... ~c6Ieaves the kingside bare; in his notes, Bernstein mistakenly remembered Janowsky-Em.Lasker, London 1899 as continuing 12...1Dc6 here, but in fact this was in a different position via a very inaccurate move-order, with Lasker not yet castled) 13 :cl (after 13 g4, 13...:e8?? walks into 14 .i.xh7+ ~xh7 IS Wc2+ ~g8 16 .i.c7 trapping the queen, Bernstein, but 13 ...lDf6 and a speedy ... ~e4 makes White's kingside activity seem premature) 13 ...lDf6 14 liJd4 "b6 IS "b3 ;t;; White has good play against the isolani. Donaldson and Minev, in Akiba Rubinstein: Uncrowned King, suggest that 11.. ..i.e7 is better than the 'modern' 11.. ..i.d6, but I find it hard to agree; Black wants to lessen the pressure against his kingside. Vaiser-Speelman, Sochi 1982 continued 11.. ..i.d6 12 .i.g3 ~c6 13 .i.c2 .i.e7 (now that White's bishop no longer covers gS) 14 h4 .i.f6 IS "d2 d4 16 :dl .i.g4 17 hS :e8 Ill_Ill.

=

=

Alternatives to the Main Lines for White White generally prefers 11 ~e2, even though it leaves his king a little more exposed, as it allows the rooks to be connected without wasting time. 1l•••lZk6 1) This time 11....td6 is not so impressive: 12 .txd6 'ifxd6 13 'ifa4! (13 'ifc2 g6 14 l:laci .tg4 15 'ifc5 'ife5 16 'iFc3 'iFh5 17 'iff6 ~d7 18 'iff4 l:lac8 "" A.Petrosian-Beliavsky, Lvov 1981) 13 ... ~c6 14 l:.hdl ~e5 15 ~xe5 'ifxe5 16 'ifh4 h6 17 l:ld2 d4 18 e4 l:le8 19 ~fl ;t Smyslov-Beliavsky, Moscow 1981. 2) Il...l:.e8?! 12 'ifc2 h6 13 l:lhdl ~c6 14 c,ftfl ;t Bischoff-Kuligowski, Reykjavik 1982 allows White the smooth development that he wants. 3) Similarly, 11....te7?! 12 'iVc2 h6 13 l:lhdl ~c6 14 a3 .te6 15 l:ld2 l:lc8 16 'ifa4 .tf6 17 ltadl 'iVb6 18 'ifb5;t D.Gurevich-Gudmundsson, Reykjavik 1982. It is essential for Black to get his queenside pieces out quickly, before White can consolidate in the centre. 12'iVc2

1) 12 h3 and now, instead of 12... .td6 13 "a4 ~b4 14 .txd6 'ifxd6 15 'ifa3 'ifb6 16 l:lhd 1 .td7 17 ~d4 ;t de Firmian-Handoko, Dubai OL 1986, 12 ...l:le8 is more harmonious; Black aims for ... d4 before White can sort out his king's position. 2) 12 "bl!? h6 13 l:.dl l:.e8 14 c,ftfl .tg4 15 .te2 'ifd7 16 a3.tfS 17 b4 a6 18 1:Ia2 ltad8 and now 19 b5!? might well improve on 19 l:lad2 'ife6 "" Seirawan-Campora, Buenos Aires 1993. An approach worth considering, even though White lost the game in question. 12.••h6

125

This compromises the kingside the least, but 12 ... g6 has also been played. Then: 1) 13 ~5 ~xe5 14 .txe5 1Ig5 15 .tg3 .tg4+ 16 ~fl l:lac8 = DjuricAbramovic, Yugoslav Ch 1995. 2) 13 a3 .te7 14 h3 .tf6 15 l:lac1 a5 16 :lhdl .te6 17 ~f1 a4 18 .tb5 'ifb6 19 .txc6 bxc6 20 .te5 with a slight advantage for White, I.Soko10v-Van der Sterren, Rotterdam 1997. 3) 13 l:lhdl .tg4 and then: 3a) 14 ~fl .txf3 15 gxf3 1If6 16 ~g2 (16 a3 .ta5 17 ~g2 .tb6 =F A.Petrosian-Lputian, Erevan 1983) 16...~d4 17 'ifa4.tc5 18 :lac 1 ~6 O.Foisor-Oll, Tbilisi 1983. 3b) 14 'ifa4!? 'ife7 15 .tb5 :lfd8 16 .txc6 bxc6 17 h3 .txf3+ 18 gxf3 .td6 19 .txd6 'ifxd6 20 l:ld2 l:ld7 21 l:lcl c5 22 1Ib5 d4 23 b4 ± D.Gurevich-Rechlis, Beersheba 1987. The weakness of Black's hanging pawns outweighed White's pawn weaknesses on the kingside. 13l:lhdl 1) After 13 h3?! White had an embarrassing accident in LechtynskyLputian, Berlin 1982: 13 ...l:le8 14 a3 .ta5 15 l:lhdl 'iff6 =F 16 ~f1? g5 17 .tg3 .txh3 18 .te2 .tf5 19 .td3 .tg4 -+. Black played less vigorously in Lematschko-Gaprindashvili, Bad Kissingen 1982, 13 ....td6 14 .txd6 'ifxd6 15 :lhdl .te6 16 l:lacl l:lac8 17 'ifc5 ;t, and in Kndevic-P.Meyer, Wuppertal 1986, 13 ....te6 14 l:lhdl l:tc815 'iVa4 'iff616l:labl d417 l:lbcl 112_ 1h. 2) 13 a3 .td6 14 .i.xd6 'iVxd6 is a bit too simplistic to promise White much, and indeed Gavrikov-Geller, Erevan 1982 was agreed drawn here.

=

126

The Queen's Gambit Declined: 5.t./4!

IS :hd 1 .tg4 should give Black no problems. 13••••f6(D) 1) 13 ....td6?? 14 .th7+ ~h8 IS :xdS ~b4 16 :xd6 +- Seirawan-Li Zunian. Biel IZ 1985. is a massive miscalculation rather than a real trap. 2) ECO quotes Zaltsman-Benko. USA 1983 as leading to a slight edge for White after 13....tg414.th7+~h8 IS .tfS j,xf3+ 16 gxf3 ~e7 17 .tc7. even though Black won a miniature in the game in question. After 17...•e8. 18 a3! improves on the game's 18 j,h3?! .taS 19 j,d6 :d8 20.ta3 (20 'iVcS? b6 21 'iVa3 :xd6 22 'iVxd6 'it'bS+) 20 ....tb4 21 j,xb4 Wbs+ 22 :d3 'ilxb4 23 :b3 'it'h4 24 :xb7 d4 2S exd4 ~6 26 'ilxc6 0-1 (26 ...:fe8+ would follow). 3) 13 ... j,e614 :acl 'it'f6 and now. rather than IS ~f1 :fd8 16 .tbS j,d6 17 j,xd6 :xd6 18 'ilc3 d4 19 exd4 :ad8 20 j,xc6 1/2- 1/2 MikhalchishinBalashov. USSR Ch (Riga) 1985. maybe IS a3 could be considered. e.g. IS ....taS 16 b4 .tb6 17 bS. while if IS ....te7. the flashy 16 "xc6?! bxc6 17 .teS leads nowhere after 17 ....tg4. but 16 .b3 looks good.

This critical position remains vastly under-explored. certainly when compared with positions in the more popular. but not necessarily better. 10 a3. A combination of ECO. Informator, and computer database gives only the two examples below. 14a3 14 ~f1 .taS IS .te2 .tb6 16 .d2 ':d8 17 ~4 ~xd4 18 exd4 = Quinteros-Najdorf. Mar del Plata 1982. 14•••.te7 15 j,e6 Is .. .'ii'xc3 16 bxc3;1;.

.c3

16~d4

Maybe this is the critical place to look for an improvement. 16 ':acl?! leads to sudden tactical problems after 16.....xc3 17 ':xc3 .tf6 18 ':c2 d4!. but Black still has some work to do to equalize after 16 "xf6!? j,xf6 17 ':abl. 16...lDxd4+ 17 .xd4 .xd4 18 exd4 Ill_lIz Tukmakov-Geller. Tbilisi 1978. 812) 10a3(D)

B

w

White eliminates the possibility of the bishop check at the cost of a tempo. This line was a popular try for

Alternatives to the Main Lines for White White in the early 1990s; Black has to play very actively to stay in the game, and sometimes has to sacrifice the h7-pawn with check. 10...lDc6 The natural move, preparing to neutralize the weakness of the d-pawn with an early ...d4. 10....i.f5 has also been tried, but aftcr 11 .i.d3 .i.xd3 12 "xd3 "a5+ 13 ~c2 "b6 14 b4 .i.d6 15 .i.g3 (15 .i.xd6!? "xd6 16 l:thdl ;t) 15...lDa6 16 :ac 1 ~7 17 lLld4 Speelman-Chandler, Hastings 198112, White has the sort of slight but stable positional advantage that the Queen's Gambit player should feel very comfortable with. ll.i.d3 (D) Threatening .i.xh7+. I) 11 :cl has only limited point, as the bishop is going to leave c5 anyway: la) 11.. ..i.b6 12 .i.d3 (12 .i.b5?! can be met by 12...d4 13 .i.xc6 bxc6 14lLlxd4 c5 15lLlb3 "xdl+ 16 ~xd1 :d8+ 17 ~el 112-112 Browne-Christiansen, SurakartalDenpasar 1982, or 12 ....i.xe3! 13 .i.xe3 "a5+ 14 b4 'irxb5 15 l:tc5 "a6 16 b5 "a5+! 17 'ird2 +) 12.....e7 13 lLlg5?! (13 0-0 .ig4 14 h3 .ixf3 15 "xf3 d4 00 Gaprindashvili, Ubilava) 13 ...h6 14 .i.h7+ ~h8 15 .i.bl d4! 16lLle4 (16 'it'd3 f5 17 lLlf3 dxe3 18 fxe3 :f6 +Gaprindashvili, Ubilava) 16....i.f5! 17 .i.d6 "c6 18 .i.xfS .i.xe4 19 .i.c5 dxe3 20 .ixe4 "xe4 21 "f3 exf2+ 22 ~fl 'irg6 ; Vaiser-Ubilava, Kislovodsk 1982. White has cause to regret not having castled earlier. I b) 11.. ..i.e7 is also possible, although the example cited by ECO as

127

good for Black is unconvincing. Tolonen-Kimelfeld, USSR 198213 continued 12 .i.d3 .ig4 13 "c2 l:tc8?! (13 ...h6!?; 13 ... g6!?) 14 .i.xh7+ ~h8 15 .i.f5 "a5+, and now instead of 16 lLld2?! lLlb417 "xc8lLld3+! 18.i.xd3 .ixc8 ;, White should have tried 16 ~dl, when Arkhangelsky and Kimelfeld's suggestion of 16....i.xf5 17 "xf5 "b5 is well met by 18 .ie5! consolidating the extra pawn. 2) 11 .i.e2!?, taking the sting out of ....i.g4, is under-explored: 2a) 11.. ..i.b6 12 b4 a6 130-0 l:te8 and now Meduna-Abramovic, Prague 1983 was soon drawn after 14 "c2?! "f6 15 :adl d4 16 .ig5 'ii'g6, but 14 "d3!? is worth trying, and if 14...d4, 15 lLlg5 g6 16 e4 ;t, the white queen having a good square in reserve on g3. 2b) Black played more actively in D.Cramling-Schiissler, Swedish Ch 1983, but White still stood better after 11.. ...f6 12 "c2 .ib6 13 .ig5!? .i.a5+ 14 ~fl "d6 15 l:tdl .i.e6 16 .i.f4 "e7 17lLlg5 g618lDxe6 fxe6 19 h4.

11••..i.b6 This strategic retreat, keeping an eye on d4, is Black's most popular

128

The Queen's Gambit Declined: 5 ~f4!

here, although there are several alternatives. Not, however, 11 ....*.e6??, the problem being not 12 ~5 ;t, as given in a hasty note in an old Informator (I'm not saying which!), but rather 12 .*.xh7+ winning a pawn. 1) 11 ...h6? is a horrible little move, wasting time and weakening the kingside. White's most direct plan is to pile up on the bl-h7 diagonal, for example 12.*.bl (12 .*.c2!? Farago) 12.. J%e8? 13 "c2 "as+ 14 b4 .*.xb4+ 15 axb4 "xal 16 'ifh7+ 1-0 Ehlvest-Martinovsky, Linares 1994. 120-0, though less ambitious, is also good for an edge, for example 12 ... .*.d6 13 "a4 .*.xf4 14 "xf4 I.Farago-Griinfeld, Malta OL 1980, and White is effectively a tempo ahead of the 11 ....*.d6line; or 12....*.e6 13 .*.bl ~ 14 "d3 g6 15 .*.xh6 l:[fd8 16 b4 ± A.Nikitin-Svidler, Lugansk 1989. 2) 11 ....*.d6 aims for gradual equality. 12 "a4 (12 .*.xd6 "xd6 13 0-0 and now rather than 13 ... h6?! 14 "a4 .*.d7 15 lIadl lLle5 16 "f4 ;t Skembris-D.Jano§evic, Belgrade 1988, Black should prefer 13 .....f6! 14 b4 .*.g4 15 .*.e2 lIad8 = Gavrikov) 12....*.xf4 (12 ... ~7 13 .*.xd6 'ilfxd6 140-0 'ilff6 15 lIfdl g6 16 lId2 lLlc6 17 lIadl ;t M.Letelier-L.Sanchez, Moscow 1956) 13 'ilfxf4 d4!?, when Gavrikov has failed to prove an advantage for White in two attempts: 2a) 14lLlxd4 'ilfaS+ 15 'ite2lLlxd4+ 16 "xd4 lId8 17 'ilfb4 'ilfg5! 1811hdl .*.g4+ (18 ... 'ilfxg2? 19 .*.xh7+ <;Pxh7 20 "h4+ 'itg6 21 lIgl .*.g4+ 22 "xg4+ ± Gavrikov) 19 f3 .tf5 20 .*.xf5 Wxg2+ 21 <;Pel "gl+ with a perpetual, Gavrikov-Beliavsky, USSR Ch 1986.

2b) 140-0-0 .*.e6 and now 15lLlg5 g6 16lLlxe6 fxe6 17 ..g3 'ilff6 18 lIhfl was Gavrikov-Timoshchenko, Irkutsk 1986. Instead, IS .*.c2 is tempting, but Black should have enough, in terms of the white king's exposure, to compensate for the pawn after 15 ...lIc8, e.g. 16lLlxd4lLlxd4 17 'ilfxd4 'ilfaS 18 'itbl (18 "e4?? loses to 18...lIxc2+ or 18....*.f5) 18...lIc419'ilfd3 g620.*.b3!? (20 'ital lIfc8 21 .*.b3? lIc3) 20....*.f5 21.*.xc4 .*.xd3+ 22 .*.xd3. Still, 13 ...d4 does not feel entirely convincing, and White can think of ways to sacrifice a pawn or two for the initiative, for example: 2c) 14 O-O!? dxe3 15 lIadl!? ext2+ 16 l:txt2 or even ... 2d) 14 lIdl!? 'ilfa5+ 15 b4 'ilfxa3 160-0. 3) 11 ... .*.e7 (perhaps slightly too passive) 120-0 (12 h3?! .*.f6 13 "c2 'ilfa5+ 14 'ite2 Alexandria-A.Maric, Biel 1991) 12....*.f6 and now: 3a) 13 'ilfc2 h6 14 lIadl, Ih-lh Yuferov-Doroshkevich, Krasnodar 1991, is uninformative. 3b) 13 h3?! is slow. Black can improve over 13 ... h6 14 IIbl .*.e6 IS .*.c2 ;t Titov-M.Iosif, Sumperk 1990, with 13 ... .*.xb2 14 .*.xh7+ 'itxh7 15 'ilfc2+ .tf5 16 'ilfxb2 (16 'ilfxf5+?! g6 and now 17 'ilfg4 is bad due to 17 ... .txal 18 lLlg5+ 'itg7!, but even after the better 17 "g5! W'xg5 18 lLlxg5+ <;Pg8 Black's queenside pawn majority might start to become threatening) 16... 'ilfb6 =. 3c) 13 'ilfb3 g6 (13 .....b614 'ii'xb6 axb6 IS lIabl .te6 16 .*.c7 .td8 17 .*.d6 .*.e7 18 .*.xe7 lLlxe7 19 lLlg5 ;t Lputian) 14 l%adl lLla5 15 'ilfa4 .te6 16 .th6 l%e8 17 .tb5 gave White a

=

Alternatives to the Main Lines for White slight advantage in the game Van der Sterren-Bonsch, Munich 1991. 4) 11...g6 12 h3 (12 O-O!?) 12 ...d4 13 e4 ~d6 14 ~h6 :e8 IS 0-0 a6 16 :cl ~f817.d2~Panno-Amura,Ar­ gcntine Ch 1992. Black is weak on too many dark squares. None of these alternatives is likely to supplant 11...~b6 as the main line. 120-0 1) 12 b4!? is a move-order tease, with White generally castling next move and transposing into the main lines, for example 12...~g4 130-0, as in the Gofshtein-Kharitonov game discussed later. 2) 12 h3 is unenterprising; ...~g4 is not fearsome enough to justify the loss of a tempo in its prevention. 12 ... d4 13 e4 h6 14 0-0 ~e6 IS:tel :'c8 16 .e2 1/2- 1/2 Keene-O.Rodri/luez, SurakartalDenpasar 1982. 3) 12 ~gS is ferocious, but is soon neutralized: 12... h6 13 ~h7+ ~h8 14 J.c2 d4! IS .hS dxe3 (lS ...d3!? 16 n·()·o ~fS 17 ~xf7+ l:txf7 18 .xf7 Wc8 19 ~xh6 gxh6 20.f6+, Anikaev, is " perpetual) 16 fxe3 1i'e8 17 ~ fS I H "xe8 1/2- 1/2 Anikaev-Geller, Mosrow 1982. 4) 12 .c2 h6 13 0-0 .f6 14 b4 i.g4 = Skembris-Abramovic, VrnI,,~ka Banja 1989. The queen is misplaced on c2. After White castles, Black must demle how to complete his development. 1.! ... ~g4 is the only line given by ,·ro, and is still critical, but 12...•f6 IIIIlI 12 ... d4 have been played with suc, I·SS:

11121: 12••."f6 129 11122: 12•••d4 131 11123: 12••• ~g4 131

129

8121) 12.....f6 (D)

13b4 13 .c2 h6 transposes to Skembris-Abramovic above, and so is not particularly promising for White. 13...~f5 1) 13 ... a6?! 14 ~gS .d6 IS l:ta2 leaves White ahead of similar lines where Black has played 12... ~g4. Arlandi-Van Dongen, Cannes 1993 continued IS ...h6 16 ~h4 g5 17 ~g3 'iVf6 18 l:td2 l:td8 19 ~bl d4 20 ~xd4 ~xd4 21 exd4 :xd4 22 :xd4 ~xd4 23 'iVd3 ~g4, and now 24 would have been strong, meeting 24 ... h5 with 25 h3 and 'iVh7+. Instead White played 24 "h7+?! ~f8 2S 1i'e4 hS with an unclear position. 2) 13 ... ~g4 14 h3 ~fS has been tried a couple of times, although one would imagine that the extra move h3 slightly improves White's position. When playing through the variations that follow, bear in mind that Black can reach them with h3 inserted. 14 b5 Lputian suggests that 14 ~xfS "xfS 15 .bl also maintains a slight edge for White, for example IS .....e6

.e4

130

The Queen's Gambit Declined: 5 i.f4!

16 a4 as 17 bxaSlbxaS 18 :dl :fd8 19 i.eS;t. 14••• lbd4! (D)

A spectacular idea of Beliavsky's, and the tactical justification behind 12...•f6. However, it seems that with accurate play White can still keep a slight edge. 15 i.xfS 1) Islbxd4i.xd416:tcl (l6exd4 i.xd3 =) 16... i.b6 17 i.xfS "xfS 18 i.c7 .e4 is on the whole drawish, although Seirawan suggested that 19 might keep a slight edge for White. Instead, Seirawan-Beliavsky, Belgrade 1991 continued 19 .c2 "xc2 20 :xc2 d4 21 exd4 :ac8 22 :fcl i.xd4 23 'ii?fl f6 =. Thnik-Sulipa, Minsk 1993, saw the same position, but with the white pawn on h3 rather than h2 (see note to Black's 13th), and this slight difference in pawn structure eliminated any back-rank wornes White might have had. Play continued 20 .c2 h6 21 'ii'xe4 dxe4 22 :fdl :fc8 23 i.xb6 :xc I 24 :xc I axb6 2S :C7 ±; the extra tempo proved useful. 2) IS lbeS i.xd3 16 .xd3 :fd8! 17 a4 (l7lbg4 Wg6 18 Wxg6lbe2+; Lputian) 17 ...lbe6 18 as i.c7 19lbg4

"f3

should not be too dangerous. Black can equalize either with 19...•e7 20 i.xc7 "xc7!? (20... lbxc7 21 f3 = Lputian) 21 :fcl "d6 22 .fS d4 23 lbes dxe3! (Lputian; 23 ...We7 24 a6;t Lputian-Beliavsky, USSR Cht (Azov) 1991) or, more simply, with 19...Wg6 20 "xg6 hxg6 21 i.xc7 lbxc7 22 :fbl (22 :abl f6 =Lputian) 22 ... d4 23 exd4:d5 24 b6 axb6 2S axb6 :xal 26 lbal lbe6 =Brenninkmeijer-Van der Sterren, Dutch Ch 1992. Is..•lbxf3+ 16.xf3 'ibfS 1784 After 17 :fdl, 17...:fd8 18 a4 transposes, while M.Gurevich in InJormator 53 gives 17... d4 18 exd4 "xbS 19 Wxb7 :ad8 as equal, but Goldin in InJormator 56 claims a slight edge for White after 20 a4, e.g. 20... WfS 21 i.e3 :dS (21...:d7 22 "f3 ;t) 22 :ac 1 :fd8 23 h3 h6 24 :c4. One can remain sceptical as to whether White is achieving anything in another of the lines that Goldin gives; after 20...Wb2 21 i.e3 i.xd4 22 'iVxb2 i.xb2 23 :abl a6 24 :xd8 :xd8 2S h4 (;t Goldin) Black looks safe enough on 2S ... i.c3. With the white pawn on h3 (via the move-order 13 ... i.g4 14 h3 i.fS instead of 13 ...i.fS), Yusupov-Beliavsky, Linares 1990 continued 18 :fdl :ad8 19 :ac1 d4 20 i.c7 .xf3 21 gxf3 ;to 17...:fd8 17...:ac8 18 :fdl :fd8 19 as i.cs 20 :acl ;t Gurevich, but again Goldin disputes this assessment, giving 20...b6 21 axb6 axb6 22 h3 (if 22 .g3, then 22 ...d4 =) 22.....d7 =. 18 as 18 :fdl transposes to YusupovGoldin, Tilburg 1992 (reached via 17

Alternatives to the Main Lines for White Afdl) and White achieved nothing afIcr 18 ... Aac8 19 h3 (19 a5 transposes 10 the previous note) 19... h6 20 a5 J.c7 =. 18••..i.e5 19 .i.e7 .xf3 20 gxf3 Ad7 21 b6! Black has a difficult endgame to defend, M.Gurevich-Beliavsky, Belgrade 1991. If now 21...i.d6?, 22 a6!. It should be noted though that Gurevich gave 17... Aac8 as a more accurute defence. The general impression is that 12 ..."f6 gives Black a couple of moves of tactical fun, but then forces him to play carefully to hold in a very Icchnical sort of position. A reasonahly solid drawing line, perhaps, rather Ihan a winning attempt.

8122) 12... d4 (D)

When even 13 "c2 remains untried, it is difficult to come to firm conclusions on this line. 13 ... h6 14 exd4 (14 Aadl .i.g4!) 14... lLlxd4 ISlLlxd4 tlt'xd4 16 .i.e3 "f6 17 i.xb6 "xb6 18 Aac 1 i.e6 is nothing, though. 13•••i.g4 14 h3 i.h5 After 14.....f6!?, 15 i.h2 .i.xf3 16 "xf3 17 gxf3lLla5 was SalovBeliavsky, Linares 1992, but Beliavsky notes 15 hxg4 "xf4 16 gS as a possible improvement, continuing 16... f6 17 g3 'it'g4 18 gxf6l1xf6 19lLlh2 'iVh3 20 "g4 "xg4 21lLlxg4 Ag6 22 .i.c4+ ;1;. However, 16.. .f6?! is a strange move, handing White a passed e-pawn, and removing his headaches over the g5pawn. 16 ...:ae8!, preparing ... lLle5, looks good for Black. 15 g4 .i.g6 16 Ad Ae8 17 Ael ~h8 18 ~g2 f6 19lLlh4 .i.e7 Black is comfortable, Karpov-Beliavsky, Tilburg 1993. 12...d4 would appear to be a sound enough way for Black to avoid the masses of theory after 12... .i.g4, to which we now turn.

"ill

8123) 12•••i.g4 (D)

To knock out the pawn protection the bishop on f4, so that .....f6 will l"IImc with gain of tempo. There is not much experience of this, but Beliavsky has twice guided the position h, equality against world-class opponcnts. l3e4 III'

131

=

132

The Queen's Gambit Declined: 5 i.,f4!

13b3 13 b4 :e8 (or 13 ... a6) 14 h3 transposes; White will have to prod the bishop sooner or later. 13•••.t.bS 14 b4 14 .t.e2 is uninspiring. 14...:e8 15 b4, as in Spassov-Georgadze, Line B1231 below, leads to a minor white deviation from the main line, while 14....t.xf3 15 .t.xf3 d4 16 exd4 .t.xd4 17 "d2 Wb6, Hiibner-Karpov, Tilburg 1983, is extremely drawish, notwithstanding White's bishop-pair. 14•••:e8 (D) Reckoning that b5 is no threat, and that bringing another piece into play is useful. 1) Even so, 14... a6 has been tried a few times: la) If then 15 d4 16 e4 (or 16 g4 .t.g6 17 .t.xg6 fxg6! Georgiev) 16....t.c7 17 .t.xc7 Wxc7 18 g4?! (18 :c2!? :ad8!? Georgiev), 18 ....t.g6 19 :e2 :fe8 ; Seirawan-Ki.Georgiev, Dubai OL 1986; White must watch out for the f4-square. Ib) White's main choice is 15:Cl d4 16 g4 .t.g6 17 e4: Ibl) After 17 ...:c8 18 :el, Epishin-Van der Sterren, Ter Apel 1992 continued 18 ...:e8 19 Wg2 .t.c7 20 Wd2 .t.xf4 21 Wxf4;t, but by analogy with the Karpov-Beliavsky game in section B122 (with b4 and ... a6 omitted), 18 ...~h8 followed by ...f6 might be a more reliable choice for Black. Ib2) Beliavsky-Ki.Georgiev, Biel 1992 saw instead 17...:e8, but again we may doubt whether this is the best place for the rook. Play continued 18 ltld2 (18 :el!? Beliavsky) 18 ...:c8 (18 ...f6!? Beliavsky) 19 .t.g3 (19 Wb3! ;t Beliavsky) 19... f6 20 b5 (20 f4!

:a2

Georgiev) 20 ... axb5 21 Wb3+ Wh8 22 "xb5 .t.c7 =. Evidently Beliavsky as White learnt a lot from this game of use for his encounter as Black with Karpov the following year. Piling pieces on the e-file is not important in this line; it is better to restrain the pawn roller with ...f6, and get the bishop back into play without loss of time. 2) 14.....f6?! 15 g4 .t.g6 16 .t.xg6 hxg6 17 Wxd5 grabbed a pawn for insufficient compensation in GavrikovVaganian, Tallinn 1988, but this was only a quickplay encounter.

w

This position has been reached several times in recent years, with White trying a variety of moves, but Black, with ...d4 and ...ltle5 both in the air, has enough play to remain comfortable: B1231: 15.t.e2?! 132 133 B1232: 15 B1233: 15 g4 134 B1234: 15 :&2 134

:ct

81231) 15 .t.e2?! d4! Georgadze's suggestion. Alternatives:

Alternatives to the Main Lines for White I) IS ... a6 16 l:a2 Wd7 17 lDes ~)xeS IS i.xhS o!ik4 L.Spassov-

=

T.Georgadze. Bulgaria 19S1. 2) IS ... i.g6 16 l:a2 d4 17 exd4 li)xd41SlOxd4 Wxd419 Wxd4 i.xd4 ::: Gofshtein-Kharitonov, Aktiubinsk 1985. 16exd4 16 lOxd4 i.xd4 17 i.xhS i.xal =+= Gcorgadze. 16•••l:xe2 17 .xe2lLlxd4 18 'iVe4 i.g6! 19 We3 1Oc2 20 'iVb3 lOxal :j: Georgadze. 81232)

ISl:c1 (D)

IS•••d4 This thematic thrust leads to a sharp tactical sequence which liquidates to what should be a drawn endgame. Others: 1) 15 ... a6!? is simple and reliable. 16 g4 i.g6 17 i.xg6 hxg6 IS Wd3 d4 19 e4 '12-'12 Speelman-Geller, London 19S2. 16 i.e2!? i.g6 =. 2) 15 ... aS (with the point 16 b5 lOeS 17 i.xeS i.xf3) 16 g4 i.g6 17 i.xg6 hxg6 IS bSlDe7!? (1S ... lLla7 19 Wd3 ~ Lputian-Geller, New York 1990) 19 i.eS?! (exchanging bishops

133

leaves holes in the white position; 19 Wd2 ~) 19 ... Wd7 20 i.d4 i.xd4 21 lOxd41Ocs 22 Wc2 l:e7 23 WcS WdS 24 l:fd 1 lOb6 2S 1Of3 l:eS 26 Wc7 l:e7, Tunik-Kharitonov, Russian Ch (Elista) 1995, and now repetition with 27 WcS would have been advisable. Instead White exchanged queens, and his far-advanced queenside pawns proved weaIc:er than Black's isolated d-pawn in the endgame. 3) IS ...lOe5, Lputian, is an untested idea. 16 g4 i.g617 i.xg6 bxg618 bS Now Dreev-Geller, Helsinki 1992 headed for a quick but bright draw after IS ... lOe7!? 191Oxd4 (19 e4 might improve) 19 ...lLlds 20 lDe2 (20 i.g3!? ao) 20 ... gS 21 i.g3 lOxe3! 22 Wxd8 :axdS 23 l:fellOdS 24 ~f1 :d7 2S l:edl l:edS 26 o!ik3 f6 27 l:d3 '12-'12. Dreev showed a simpler, tactical path to safety, and a Brazilian GM followed the road a year later... 18•••dxe3 19 bxc6 e2 20 WxdS exnW+ 21 ~ l:axdS 22 cxb7 f6?? A move too far down the track. Dreev gives this as equal, but it loses. 22 ... a6, Lima, prepares the defence ... i.a7-bS, and forces White to win back the exchange while he still has the chance. Fta<:nik has suggested 23 lOeS!? l:bS 24 lOc4 l:xb7 25 lOd6, and if 2S ...l:bbS, 26 lOxeS l:xeS 27 l:c6 with endgame pressure. Black is fine though after 25 ...i.c7! 26 l:xc7 :xc7 27 lOxeS l:c3, and if 2S i.d6 f6 White's minor pieces are in a tangle. Instead... 23 l:bl! gS 24 i.g3 l:d7 24 ...l:bS 2S i.xbS l:xbS 26 a4 ~f7 (26 ...l:xb7 27 as +-) 27 l:bS +-; 24 ... l:d3 2S ~g2 l:xa3 26 l:el! l:f8

1.\4

The Queen's Gambit Declined: 5 J.J4!

(26...l:ld8 27 l:lel +-) 27 -*.d6 l:lb3 28 -*.xf8 -*.c7 (28 ...-*.a5 29 -*.b4! +-) 29 l:le7 -*.d6 30 l:le8 +- is analysis by Lima. 25 bS'it' l:lxbS 26 -*.xbS And the extra piece won the endgame in Lima-Milos, Sao Paulo Z 1993. A game with a moral: when following published analysis, keep checking the position on the board, and don't just rely on the analyst, particularly if the move in question is a quiet one at the end of a long forcing sequence.

81234) 15 l:la2 (D)

81233) 15 g4 (D)

The most popular choice. White is prepared to double rooks on any of the central files if required. 15•••d4 The most enterprising reply, but it would seem that Black may also kill the tension with 15 ... lbe5!? 16 -*.xe5 -*.xf3 17 "xf3 l:lxe5 18 l:ldl "g5 19 l:lc2 l:lae8 =Ehlvest-Kotronias, Reykjavik 1994. 16b5 16 g4 -*.g6 17 -*.xg6 hxg6 18 :ld2 (after 18 b5, 18 ...lba5, as in the Seirawan-Speelman game given below, would transpose to the main line, whileI8 .....d519bxc6 ..xa220cxb7 is unclear) 18 .....d5 19 exd4 l:le4, Kharitonov, and Black's central blockade nullifies White's extra pawn. Kharitonov also gives 16 l:ld2 "f6 17 g4 -*.g6 18 -*.xg6 hxg6 19 -*.g5 "e6 20 exd4 "d5!, and again White's extra pawn means very little. 16•••~ 1) 16... lbe5 17 -*.xe5 -*.xf3 18 "xf3 l:lxe5 and now White stands better in the centre after 19 e4, followed by "e2 and f4 - M.Gurevich. White

B

Possibly releasing the tension a little too early. 15•••-*.g616-*.xg6hxg617 b5 ~ IS 'it'd3 ~4 19 'it'b3 'it'r6 20 l:ladl l:ladS 21 ~g2 'it'e7 22 l:ld3 'it'e4 23 ~g3

In the game Chiburdanidze-Geller, Aruba 1992, Black headed for the draw with 23 ...lbb2 24 l:ld2lbc4, etc. Black may well be able to try for more, for example 23 ... f6 24 h4 l:le7 followed by ... l:led7 and White's kingside is over-extended.

Alternatives to the Main Lines for White could try picking up a pawn with 19 exd4 1i'xd4 20 1i'xb7 instead, but after 20 .. J:ld8 21 :d2 1i'c3 22 :e2 g6 Tunik-Komeev, Smolensk 1991, his winning chances are slight, in view of the opposite-coloured bishops and Black's active pieces. 2) 16 ... i.xf3?! 17 "xf3 dxe3? (l7 ... 11k5 ;t) is unsound, but highly entertaining, with promotion, underpromotion and a king wandering coming up. Gavrikov-Kharitonov, USSR Ch(Moscow) 1988 continued 18bxc6 'iVxd3 19 cxb7 exf2+ 20 Wh2!! (20 :fxf2 'iWxf3 21 gxf3 .!:tab8 = Kharitonov) 20 .....xfl (20... 'iWxf3 21 gxf3 :ab8 22 i.xb8 :xb8 23 :b2! +Kharitonov) 21 bxaS....gl+ 22 Wg3 fl tD+ 23 Wh4 i.dS+ 24 'iWxdS .!:txd8 25 .!:tf2 f6 26 :xfl g5+ 27 Wh5 "c5 28 i.e3? (a move-order slip; Kharitonov notes that 28 "b3+! Wh8 29 i.e3 'iWe5 30 'iWf7 would have won easily) 28 .....c4! (now the white king is in danger) 29 Wh6:e8 30 i.d2 "f7 31 "d3 :e2! (threatening ...:xd2) 32 'iWdS+ .!:teS 33 'iWd3 with a draw. 17 g4 i.g6 18 i.xg6 18 exd4 'iWdS 19 l:e2 i.xd3 20 "xd3 :xe2 21 'iWxe2 h6 22 :dl tDc4 23 a4 :d8 is another line where Black successfully blockades the light squares after having sacrificed the isolani to give White an isolani of his own. Neither side can do anything constructive if the opponent holds his ground. Gavrikov-M.Gurevich, USSR Ch play-off Vilnius (4) 1985 continued 24 Wg2 WfS 2S "c2 WgS 26 .!:td3 i.aS! 27 1i'e2WfS 28 Wh2 :e8 29 "c2 wgS and the deadlock remained unresolved. 18••• hxg6 19 tDxd4 1i'dS 20

:c2

135

Here Seirawan-Speelman, London 1984 continued 20... tDc4 21 a4 :acS 22 tDf3 "xdl 23 :xdl tDeS 24 tDel tDf3+ 2S Wfl tDxel 26 :xc8 :xcS 27 .!:txel :c4 28 :al and Black was ultimately unable to hold the endgame. The simple 20... i.xd4, again aiming to fix a white pawn on d4, seems more satisfactory, when after 21 'iWxd4 "xbS 22 :cS 'iWa6!, Black is comfortable, or 21 exd4 :e4 22 i.e3 b6 (to prevent :cS) and Black can think about attacking with ...fS-f4. Various attempts have been made to bolster the Knight Exchange System in recent years. but Black seems OK if he plays actively. If there is any scope for improvement for White. one intuitively feels that it is in the less well explored 10 i.d3 system, rather than 10 a3, after which the plan of development with ...i.b6, ..."f6 and ....tg4. often in association with ...d4. is difficult for White to attack.

82) 8 ...exdS (D)

This is the Pawn Exchange Variation. Play normally continues with 9

136

The Queen's Gambit Declined: 5 ~f4!

.i.e2 ~c6 here, leading to a position that can be arrived at by several different move-orders after 1 d4 dS 2 c4 e6 3 ~c3 ~f6 4 ~f3 .i.e7 5 .i.f4. For example: 5 ...0-0 6 e3 c5 7 dxc5 .i.xc5 8 cxdS exdS 9 .i.e2 ~6 (the move-order being considered here); 5 ... 0-06 e3 c5 7 dxc5 .i.xc5 8.i.e2 ~c6 9 cxdS exd5; 5 ... 0-06 e3 c5 7 dxc5 .i.xc5 8l:lcl ~6 (8 ...dxc4!?) 9 cxd5 exd5; 5 ... 0-0 6 e3 c5 7 dxc5 ~c6 8 cxd5 exd5 9 .i.e2 .i.xc5; 5 ... c5 6 dxc5 ~c6 7 e3 J.xc5 8 cxdS exd5 9 J.e2 0-0. If one is looking for something against the Tarrasch, there is also the further possibility 1 d4 d5 2 c4 e6 3 ~3 c5 4 cxdS exd5 5 ~f3 ~6 6 .i.f4 ~f6 7 e3, and if 7 ....i.e7, 8 dxc5 .i.xc5 9 .i.e2 0-0. The .i.f4 system against the Tarrasch is well out of fashion, with no recent InJormator references. ECO (Kasparov) gives 7 ...cxd4 8 ~xd4.i.b4 as equal, but the material is sparse, and there must surely be scope for exploration and improvement. I leave this to the reader. The diagram position is generally a pleasant one for White to play; he has chances of manoeuvring against the isolated queen's pawn structure, while Black will find it difficult to establish counterplay. Black certainly does not have the fluidity of piece movement associated with the Knight Exchange Variation. White's main. choices are 9 .i.e2 and 9l:lc 1. With the c3- and f6- knights still on the board, it makes sense to keep the isolated pawn under restraint, rather than, as in the 9 cxd5 ~xdS 10

~xd5 exd5 line, to play for pressure along the bl-h7 diagonal. We consider two minor lines first: 821: 9 J.d3 136 822: 9 J.eS 136 823: 9l:lcl 136 824: 9 .i.e2 137

821) 9J.d3lllc6I00-0andnow 1O...J.g4 11 l:lcl d4! 12 ~b5 .i.b6 13 ~bxd4 ~xd4 14 exd4 Wd5! = TeichmannJanowsky, Hastings 1895, was, as we noted in the Introduction, a remarkably avant-garde idea for its time. Instead 1O....i.e6 lll:lci lIc8 12 .i.bl ~ as in the game Steinitz-Burn, Hastings 1895, leaves White's initiative unchallenged. 822) 9 .i.eS .i.e6 10 .i.d4 shouldn't be too dangerous: 1O....i.d6 (1O....i.e7!?) 11 .i.e2 ~c6 12 0-0, Juarez-Campitelli, Villa Ballester 1996, and now 12...~ seems simpler than 12...~d4 13 ~xd4. 823) 9l:lcl~(D)

w

Alternatives to the Main Lines for White This will usually transpose, after 10 i.e2, back into the main line, B24, though 1O... d4! should equalize (for example 11 exd4li)xd4 12ll)xd4 'ilxd4 13 'ilxd4 .i.xd4 14 li)b5 .i.b6 15 0-0 i.e616 a3li)dS =Ye Rongguang-Van der Sterren, Antwerp 1997) suggesting perhaps that White's rook move is best delayed. There are also a few minor offshoots.

10a3 1) 10 li)xdS?! (this tactic rarely works, as we shall see in other settings as well) 1O... li)xdS 11 ':xcS li)xf4 12 'if'xd8 (12 exf4?? 'ile7+) 12...li)xg2+ 13 .i.xg2 ':xd8 = Reinfeld. If anything, Black may be a little better. 2) 10 .i.d3?! d4 11 exd4li)xd4 12 0-0 .i.g4 ~ Steinitz-Em.Lasker, St Petersburg 189516. 10...d4! If you can play ... d4, do so! 1) 10....i.g4 11 .i.e2 l:te8 12 0-0 .i.e7 13 'ii'b3li)aS 14 'ila2li)c6 is also reasonably solid, Larsen-B.lvanovic, Bugojno 1982. 2) 10... a6 was reached by transposition (8 a3 li)c6 9 ':cl a6 10 cxdS exdS) in Dreev-Short, Linares 1995. 11 .i.e2 would lead towards standard positions. Instead White tried for more, and achieved much less, with 11 .i.gS?! d4 (11.. ..i.e7?! 12 .i.e2 .i.e6 13 0-0 ± Hamann-Filip, Lugano OL 1968) 12li)b5? (12 exd4) 12... dxe3 13 'ilxd8 exf2+ 14 ~e2 ':xd8-+. 11 exd4 White probably does best to settle for dull eqUality. 1) llli)a4?! 'if'aS+ 12 b4li)xb4 13 li)d2 (captures on cS are met by ... ltXl3+; 13 axb4 .i.xb4+ 14 'it'e2 d3+ 15 'it'xd3 .b5+ -+) 13 ...dxe3 14 fxe3

137

(14 li)xc5!? exd2+ ~) 14...li)a6 15 .i.xa6 .txa3! 16 ':al 'i'xa6 17 ':xa3 bS + Slutskin-Kibalnichenko, Russia 1992. 2) 11 li)e2? dxe3 12 'i'xd8 exf2+ and Black wins. 3) 11 li)bS 'ilaS+ 12 b4 (12 'ild2 'i'xd2+ 13li)xd2 dxe3 14 .:xcS exd2+ 15 .i.xd2li)e4 ~) 12... li)xb4 13 axb4 .i.xb4+ 14 li)d2li)e4 15 exd4 ':e8 16 li)c7 li)xd2+ 17li)xe8 li)b3+ 18 'it'e2 li)xcl+ 19 'if'xcl .i.g4+ +. In this analysis by Kibalnichenko and Zinov, the play is fast-moving, but White's lack of development always tells. 1l...li)xd4 12li)xd4 'ii'xd4 13 'ii'xd4 .i.xd4 Equal, according to Kibalnichenko and Zinov. If anything, Black is to be preferred, as White is behind in development and exposed to checks on the e-file. The f2-square could also cause problems. An editorial note in In/ormator 57 gives, after the suggested 14 li)b5.i.xb2 IS ':c2, 15 ....:e8+ as good for Black. If then 16 .te2 li)dS 17 l:txb2 li)xf4 18 li)c7, Black wins a pawn by 18... li)d3+! 19 'it'd2li)xb2. 824) 9 .i.e2 (D) 9•••li)c610 0-0 After lO':c 1, 1O....i.e6 11 0-0 leads back to the main line, while 10...d4! 11 exd4 li)xd4 12 li)xd4 .xd4 13 .xd4 .i.xd4 14li)b5 .tb6 15 0-0 (IS li)c7? .i.xc7 16 .i.xc7 ':e8 +) IS ....i.e6 16 a3 ':fd8 17 li)d6 l:ld7 18 .i.bS ':e7 19 ':fel Ill-If'], was ~~~~Karpov, Seville Wch (22) 1987 (where the move-order was 8 ':c 1 li)c6 9 cxdS exdS 10 .i.e2).

138

The Queen's Gambit Declined: 5 1./41

lO....te6 1) 10...h6 was not tested very hard in Bagirov-Guseinov, Baku 1993: 11 ~e5 .td6 12 ~xc6 bxc6 13 .txd6 'iVxd6 14 ':cl .te6 15 Wc2 :lab8 16 b3 ':fc8 17 ':fdl 112-112. It makes more sense to force the black bishop to commit itself before playing ~5. 11 ':cl!? .tb6 12 ~5 ~e7 13 ~a4 ;t Nikitin-Shur, USSR Cht (Azov) 1991. 2) 10... a6 11 ~5!? (11 ':cl .te6 leads back to the main text) 11...~7 12 ~d3 .ta7 13 .te5 ;t D.JohansenFederau, Berlin 1984. U':c1 The natural developing move, and the most popular choice, but there are several reasonable alternatives, most of which seem to keep a plus for White. 1) 11 ~b5!? a6 (11...':c8 12 ':cl transposes to the main line) 12 ~bd4 'iVb613 ~b3 (13 ~xc6 bxc614 Wc2 .te7 15 ~g5 .td7 16 ':acl also gives White a slight advantage, E.MedunaAbramovi~, Moscow 1982) 13 ... .te7 14 ~fd4 ~ 15 .td3 ~xd4 16 ~xd4 .tf6 17 ~xe6 Wxe6 18 We2 ~c5 19 .tc2 g6 20 ':adl d4 21 b4 iDd7 22 .tb3 "e7 23 "d2 dxe3 24 'ifxe3 ;t Portisch-T.Bauer, Hungary 1993.

2) 11 ~a4 .te7 12 ~5 ~e4 13 ':cl ~xe5 14 .txe5:C8 15 :Xc8 "xc8 16 ~c3 "c5 17 .td3 ~xc3 18 .txc3;t Zsu.Polgar-Franco, Salamanca 1989. 3) I1lL1e5 .td6 (11...~7 12.tO ;t Mikhalchishin; 11.. .':c8 12 ~xc6 ':xc6 {12...bxc6? 13 .ta6 traps the rook} 13 .te5 ~d7 14 .td4 ;t Gavrikov-Magomedov, Tallinn rpd 1988) 12 ~xc6 bxc6 13 "a4 (13 .txd6!? 'ii'xd6 14 "d3 c5 15 ':fdl Mikhalchishin) 13 ....txf414 Wxf4c5 (14 ...Wb8 15 'ii'xb8 ':axb8 16 b3 c5 17 :edl ':fc8 18 ':acl g6 {18...':b4!? Cebalo} 190 ;t Cebalo-Rukavina, Yugoslavia 1982) 15 b3 WaS 16 ':acl :lac8 17 'ii'a4 Wxa4?! (17 ...Wb6! 18 ':fdl ':fd8 GO Mikhalchishin) 18 ~xa4 ;t Mikhalchishin-Beliavsky, Moscow tt 1981. 4) 11 a3 a6 (1l...':c8 12 ':cl transposes to the main line) 12 ':cl.te7 13 ~5 ':c8 14 ~xc6 ':xc6 15 .te5 ~4 16 .td4lL1xc3 17 ':xc3 ':xc3 18.txc3 .tf6 19 'ifd2 ;t Golod-Suetin, Alushta 1993. 5) 11 J..g5 seems slightly unthematic: 11.. ..te7 12iDd4 h6 13 .th4 ':c8 14 ':cl WaS 15 ~xe6 fxe6 16 "b3 Wb4 = Shamkovich-Yudovichjr, USSR 1964. These lines exhibit a wide range of typical isolated and hanging pawn p0sitions. The general impression would be that White plays natural moves and keeps a slight edge. This is a pattern we shall witness several times in the second half of this book, as we examine ways in which Black avoids the double-edged positions of the main lines. Black avoids being quickly overrun, but makes it easier for White to achieve a modest positional plus.

Alternatives to the Main Lines for White After 11 lIcl, Black has several choices, with 11.. .:c8 and 11.. ..tb6 being the most popular. 11..."e7 may be disregarded, as it loses a clear tempo after 12 tLla4 .td6 13 .txd6 "xd6 14 tLlc5 b6 15 tLld3 it L.Popov-Marjanovie, Albena 1977. We discuss in detail: B241: 11 ....te7 139 B242: 11•.•a6 139 B243: 11.••.tb6 140 B244: 11••• 140

:ca

8241) 11.•..te7 (D)

139

2) 12 "a4 "b6 13 'iWbS "xbS 14 tLlxb5 tLle4 IS :fdl .tf6 =OrnsteinSchussler, Eksjo 1982. Here White simplified a little too early and a little too directly, allowing Black to position his minor pieces actively. 12••• 13 tLlxc6 13.tn'iWd7 14 h3 tLlxd4 IS "xd4 (IS exd4!? it) IS ...bS 16 :fdl :c4 17 "e5 :fc8 18 :bl :4cS 19 e4 ± Tartakower-Klein, Paris 1935, although one can certainly question the accuracy of Black's play. Piling everything on the c-file leads to neglect of the d- and efiles. 13•••:Xc6 13 ... bxc6?? 14 .ta6 wins the exchange, a recurring theme in this type of position where White has a bishop onf4. 14 .to 15 :d8 16.b5 .tb417 .xaS .txaS 18 tLlb5 .tb619 tLld4 ± Browne-Cifuentes, Santiago 1981. These isolated queen's pawn positions can be difficult to judge. When is White better? When is Black equal? The general answer is that when, as here, White establishes the more active piece positioning in the centre he is clearly better. When Black, as in the Ornstein-Schussler game, is able to establish the more active cluster of pieces in the centre, he is comfortable at least.

:ca

.as .d3

This should be regarded with some suspicion, as Black takes a guard off the blockading square in front of the isolated pawn. 12tLld4 The natural reply. Others allow Black good prospects of equalizing. 1) 12 .tg5 :c8 13 a3 (13 tLld4 tLlxd4 14 "xd4 "a5 15 a3 :c6 16 e4 dxe4 17 tLlxe4 J:d8 18 "e3 112-112 Reshevsky-R.Byrne, USA Ch 1981) 13 .....a5 14 tLla4 and now, rather than 14 ... tLle4 15 b4 "d8 16 .txe7 tLlxe7 17 tLlcI4 it Petrosian-Filip, Buenos Aires 1964, Black should try 14...b5!? IS tLlcs b4 IIC.

8242) 11... a6 12 tLla4 12 tLleS!? is a more promising try for White. 12.••.ta7 12....te7 13 tLlcs .txcS 14 :xcS it was played in Uhlmann-StAhlberg,

140

The Queen's Gambit Declined: 51./4!

Marianske Lazne 1965, but it makes questionable sense for Black to vacate a7 for the bishop, then play it to the inferior square e7. 13a3 13 J.g5 d4!? 13...:c8 14 ~cS liz_liz NotkinHorin, Moscow 1991, possibly in view of 14 .....e7. 8243) 1l...J.b6 (D)

Yugoslav Ch (Borovo) 1981) 13 'ifa4 (13 .i.g5 d4 14 exd4 'ifxd4 15 .txf6 gxf6 oo ECO) 13 ...d414exd4(14:fdl ~! Bukie) 14.....xd415~f3 .xa4 16 ~xa4 .i.xa2!? (16...~d5 =is simpler) 17 .i.d6 :fe8 18 :al ~5 19 J.b5 J.c4 20 .i.xe8 :xe8 21 :fdl .i.b3 22 ~xb6 .i.xdl 1/Z-1f2 VaganianBeliavsky, Baden-Baden 1980. A flurry of tactics leads to a peaceful conclusion. 12....i.d7 12.. :ii'e7 13 :fdl :fd8 14 .tg5 h6 15 .th4 :ac8 16 ~b5 'ifb4 17 'ifxb4 ~xb4 18 .i.xf6 gxf6 19 a3 ~c6 20 ~fd4 ~xd4 21 ~xd4 .txd4 22 exd4 :xc 1 23 :xc1 :c8 24 :xc8+ .txc8 and White should almost certainly win the bishop ending, Seirawan-R.Byme, USACh 1981. as 13 :fdl ~4 14 14...~xe2+ 15 ~xe2 .i.g4 16 ~d4 ~ 17 h3; Karpov-BeJiavsky, Moscow 1981. 15.a3 ~xe2+ 16 ~e2 .tg417

.b4

Taking the bishop out of range of the rook, in order to be able to play ... d4 if given the chance. 12 'i'a4! Deterring ... d4, because of the possible pin on the d-file. In other lines, ... d4 equalizes, for example: 1) 12 ~a4 d4 (another possibility is 12...~e4 13 ~5 ~xe5 14 J.xe5 "g5, as in Schiissler-S.Heim, Reykjavik 1981) 13 exd4 ~xd4 14 ~xd4 'ifxd4 15 'ifxd4 J.xd4 TaimanovAn.Bykhovsky, USSR Ch (Tallinn) 1965. 2) 12 ~e5 ~7 (or, more directly, 12 ...d4 13 ~xc6 bxc6 14 exd4 .xd4 15 'ifxd4 .txd4 =Kne!evie-Rukavina,

=

lOed4

17 .d6?! :a6! 18 'ifxd8 .txd8 = Vaganian-Beliavsky, Leningrad 1977. 17...:c818 b3; Gheorghiu-Soylu, Athens 1981.

8244) 1l...:c8 (D) The main line. White now generally chooses 12 ~b5 or the perhaps most accurate 12 a3, while there are a few others too: B2441: 12 ~dS? 141 B2442: 12 J.g5 141 B2443: 12.a4 141 B2444: 12 ~ 141 B2445: 12 ~b5 141 B2446: 12 a3 142

Alternatives to the Main Lines for White

141

82444) 12lDeS(D)

82441) 12lbxdS? This fails tactically. The same trap is also sprung after 12 a3 h6 13 lbxdS? 12.....xdS 131fxdS lbxdS 14 :XCS lL\xf41S exf4lbd4! 16 :eS lbxe2+ 17 llxe21.c4 Black wins the exchange. 82442) 12 1.gS 1.e7 This inter-transposes with 11 .i.gS .i.e7, or with 11 :cl .i.e7 12 .i.gS (12 lbd4!?); see B24 and B241 respectively. The second move with the bishop is not very convincing. 82443) 12"a4""6 As Black has no need to fear'ifbS, 12 ... a6 13 %lfdl 'ii'b6 is too cautious: 14 "c2 1.e7 IslOes :fdS 16 .i.f3 is slightly better for White, Ribli-Unzicker, Baden-Baden 19S1. 13'irbSd4! 13 .....xbS ;t Chekhov. 14 lba4 'ifxbS 15 .i.xbS .i.e7 16 exd41.xa2 Unclear, according to Chekhov. Black should be OK.

Il....i.d6 12... lbe7 13 "a4 ;t, although ECO is mistaken in citing this as Quinteros-Bolbochan, Buenos Aires 1976, since a3 and ...h6 had been inserted. 13 lbxc6 bxc6 14 .a4 1.xf4 15 .xf4 cs 16 :Cdl :b8! This seems satisfactory for Black, despite the loss of tempo when compared with lllOes lines (B24, note to move 11). 17 b3 :b4 18 .eS :e8 = J~:­ Keres, USSR Ch (Tbilisi) 19S9. -lJlacK's- pieces are so active that White can launch no really effective attack on the hanging pawns. 82445) 12lbbS (D) The direct approach. l l... lbe4 Black in return aims for immediate piece activity. 1) 12....e7?! runs into an awkward pin. 13 1.gS .i.b6 14 "d3 .i.g4 IS h3 .i.hS (lS ....txf3 16.txf3lOeS 17 "dl ± Chekhov) 16 ~3 .tg6 17 "bS .te4 ISlL\d2 :cdS 19l:lcdl :reS

142

The Queen's Gambit Declined: 5 i.f4!

B

fS 17 i.c4 ;t, but IS .....d2 is a con· vincing reply. 14.xdlJ.b4 14.....b6!? IS :xcS .xcS 16 J.d6 "b6 17 J.xf8 ~xf8 18 :cI a6 19 lOd4 lOxd4 20 :xc8+ olxc8 21 exd4 a5 Kasparov. IS.dl IS "d3!? W'f6 16 a3 (16lDd6 i.xd6 17 J.xd6 J.fS! "") 16... J.e7 17 b4 as 181Oc7 i.fS 19 W'b3 W'g6!, Kasparov, provides adequate counterplay. 15••••b6 16 a3 ole7 17 b4 as 18 lOd6 :cd8 19 lDxb7! .xb7 20 bS

=

20 lDdxe4 dxe4 21lDds ± ChekhovWilhelmi, Hamburg 1996. 2) 12...ole713 tDes (13lilld4 'ifb6 14lDxe6 fxe6 IslDd4lDxd4 16 "xd4 olcs = Tischbierek-Lechtynsky, Halle 1981) 13 .....a5 (why not 13 .....b6immediately?) 14 a3 'ifb6 IS b4 tDe4 16 lDxc6 bxc6 17 lDd4 old7 =P.MeyerH.Eng, Bundesliga 199112, although one feels the play on both sides can be tightened up. 13lDd2 13lDfd4?! .b6 14lDb3 ole7 IS f3 lDf616.d2 a6 17lDsd4 i.b4 18 "dl lDh5 =+= A.Petrosian-Magomedov, Tallinn rpd 1988, merely shows that strong players can make substandard strategic decisions in a quickplay. 13 lDgS lOxgS 14 :xcS (14 i.xgS =) 14... lDe4 IS :cl .b6 is given by Kasparov as unclear. Black appears comfortable. 13...lDxd2 After 13 ...ole7!? 14lDxe4 dxe4 IS lDd6 i.xd6 16 olxd6 :e8, given as unclear by Kasparov, the possession of the bishop-pair does not promise much for White, as Black's pieces are centralized and active, and his pawns straightened out. IS "a4 (Kasparov) at first looks good, e.g. IS ... a6?! 161Oc3

:c8! If the knight moves, 21 :c7. After the text, a draw was soon agreed in Korchnoi-Kasparov, Brussels 1986: 21 bxc6 :xc6 22 :bl :b6 23 :xb6 'i'xb6 24 i.eS 112-112. 12 lObS is an interesting try, but its defect is that it abandons the e4-square too quickly.

82446) 12a3(D)

The most subtle of White's tries, and perhaps the best chance of carrying through an advantage. It is essentially a waiting move; White makes a

Alternatives to the Main Lines for White little pawn move which will probably be useful later on (in preparing b4, or taking b4 away from a black piece), and waits for Black to make a piece move before deciding his own piece placement. 12••• h6 Black too waits. This cuts out .i.gS, and leaves open the potential threat of ...lbhs. Others: I) 12 ... .i.d6 13 .i.xd6 Wxd6 14 'ifa4 a6 IS ':fdl ':fd8 16 h3 t. White, who has the classic slight edge against the isolani, made ground using the standard plan with 16...h6 17 l:ld2 "fIc7 18li)d4li)xd4 19 Wxd4li)e4 20 lldc2li)xc3 21 ':xc3 Wd7 22 .i.f3, etc., in Korchnoi-Seirawan, Las Palmas 1981. 2) 12... .i.b6 and now 13 'iWa4 "e7 14 .i.gS h6 IS .i.h4 gS 16 .i.g3 ':fd8 17 ':fdl li)hS = Quinteros-Spassky, Amsterdam 1973 is about the first chance of dynamic play for Black that we have seen in this system. White can try instead 13li)bS!? li)e4 14li)bd4;1; (Minev) or 13 li)a4!? 3) 12....i.e7 13ll)eS (13 .i.gS transposes to Petrosian-Filip, section B241; 13 h3!?) and now 13 .....aS 14 "d3 ':fd8 ISli)xc6 ':xc6 16 b4 Wxa3 17 li)xdSWxd318li)xe7+~fSI9li)xc6!

"ii'xe2 20 li)xd8 ± Suba-Resende, Hastings 199011, unexpectedly shows a theme we saw a lot in the Old and New Main Lines - the queen getting into trouble on as. 13 ... .i.d6!? might be worth trying, despite the apparent loss of tempo, the point being that in any hanging pawn position (after 14 li)xc6 bxc6, etc.), it is more difficult for White to restrain ...c4, and to withstand pressure along the b-file, if the

143

white pawn is on a3 rather than on a2, as protection is removed from b3. 4) 12... a6 13li)bS "b6 (13 ...li)e4 14 li)bd4 ;1;) 14 ':xcS "xcS IS .i.d6 "b6 16 .i.xfS ~xfS 17li)bd4 "xb2, and now, rather than 18 a4?!, PortischIoseliani, Monaco 1994, 18 li)xe6+! fxe6 19 'ifd3 gives various attacking chances, though one would not blame the player of the white pieces for distrusting this plan. 13 'iWd3!? is a more thematic method of development. 13.i.g3 Another quiet move which, it should not be forgotten, now threatens li)xdS. For 13li)xd5? see B2441. 1) 13ll)eSll)e7 (13 ....i.d6 ECO) 14 'ifa4 a6 IS ':fdl ;I; Quinteros-Bolboehan, Buenos Aires 1976. 2) 13 li)bS and now 13 ....i.b6? 14 li)fd4? li)xd4 ISli)xd4 ':xc116 "xcI .i.xd4 17 exd4 "b6 112-112 O'KellyTimman, London 1973 was very odd, as 14lOci6 wins material. Black should play 13 ... .i.e7!? or 13 ...li)e4!? 3) 13 "a4 a6 14 li)eS li)xeS IS .i.xeSli)d7 16 .i.d4 =ECO, the point presumably being 16...li)b6 17 'ifdl .i.xd4 18 "xd41Dc4. 13....i.b6 14li)e5 14li)a4li)e4. 14...li)e7 15 li)a4 li)e4 16 :Xe8 .i.xc8 17 li)f3 White keeps a slight advantage, Fischer-Spassky, Reykjavik Wch (14) 1972. This is the sort of clear-cut position with well-defined strategic objectives that Fischer normally played superbly in his peak years. It is all the more surprising then that his position quickly fell apart after 17 ....i.d7 18 .i.eS? (allowing Black to exchange his bad bishop, rather than his good

=

144

The Queen's Gambit Declined: 5 J.f4!

bishop, for the knight; 18lOxb6 .xb6 19 .te5 ~ Gligoric) 18....txa4 19 .xa4 1Oc6! 20 .tf4 .f6! and Black had a serious initiative, although after further mistakes the game was drawn. The heavy positions with an isolated d-pawn are not to everyone's taste, demanding from both players good manoeuvring skills and clear positional judgement. Nevertheless, White keeps a slight pull, and can use this line as a legitimate winning try. The player who seeks the main lines, with .c2, a3, 0-0-0, etc., will encounter this line only if Black plays 7 ...lOc6 instead of7 ....txc5, though even here, White could, if he so chooses, offer the main line with 8 .c2!? lOb4 9 .bl.txcS 10 a31Oc6 11 .c2!?, as in Ribli-Gligoric, BledlPortoroz 1979.

moves a piece out to a reasonably good-looking square, and ignores the pawn structure. Black's main priority is to find a favourable way to fix the central pawn structure; this is usually easily enough done. We consider the following moves:

A: 7.c2 B: 7:et

C: 7h3 D: 7.td3 E: 7.te2!? F: 7 adS

144 144 144 14S 146 147

A) 7 .c21Oc6 8 0-0-0 WaS 9 a3 cxd4 10 exd4 .td7 11 lOe5 :fd8 ao Sanchez-0stergaard, Copenhagen 1996. This is certainly no improvement on the New Main Line.

B)

4.5 White's alternatives on move 7 1 d4 dS 2 c4 e6 3lOc3lOf6 4lOn .te7 5 .tf4 0-0 6 e3 c5 (D)

7:let.aS 7 ... cxd4 8 exd4 b6 9 .te2 lOc6 10 0-0 dxc4 11 .txc4 .tb7 = KapstanD.Allan, Canada 1996. 8 .d2 :d8 9 cxdS cxd4 10 exd4 ll'lxdS lllOxdS .xdS 12 .tc4 'if'e4+ 13 .te3 .td7 14 0-0 .tc6 SavitskyRauzer, USSR Ch (Leningrad) 1934.

=

w

C)

These mostly come under the category 'naIve developing moves'; White

7 h3lOc6 7 ... cxd4 8 exd4 a6 and rather than 9 cxdslOxdS 10 lOxdS .xdS ; TorbinTokmachev, Kazan 1997, White should try 9 cS!? 883 8 cxd5 lOxdS 9 lOxdS .xdS 10 dxcS .xcS 11 .td3 :d8 12 Wbl :xd3 13 .xd3lOb4 is slightly better for Black, Schwierskott-G.Schmidt, Bundesliga 1996f7.

Alternatives to the Main Lines for White 8•••a6 ~H[c1 cxd4 10 exd4 dxe4 11 i.xc4 b512 i.al i.b7 = Anikaev-Listengarten, USSR 1972. This line is too quiet.

D) 7 i.d3 (D)

B

Played a lot in the early days, but still a 'naIve developing move'. 7•••lLle6 1) 7 ...dxc4 8 i.xc4 and now: la) S... a6 9 dxc5 will transpose into the 7 dxc5 i.xc5 8 i.e2 dxc41ine. Less to the point is 9 O-O?! b5 10 i.b3 i.b7 11 We2 Wb6 P.lohner-Teichmann, Karlovy Vary 1907. 1b) Similarly, White does best to meet 8 ... lLlc6 with 9 dxc5; instead, 9 lLle5?! cxd4 10 lLlxc6 bxc6 11 exd4 Wb6 12 Wd2 :d8 =was S.Booth-Sandager, Las Vegas 1994. 2) 7 ... a6 8 cxd5 (8 0-0 lDc6?! {8 ... dxc4 9 i.xc4 transposes into 1ohner-Teichmann} 9 h3?! cxd4 10 exd4 dxc4 11 i.xc4 b5 = TarraschSchlechter, Ostend 1905) 8...lDxd5 9 lDxd5 W'xd5 10 0-0 cxd4 11 exd4.id7 12 :el i.c6 13 :e5 Wd8 14 :e3 ;I; Schlechter-Tartakower, Karlovy Vary 1907.

=

145

80-0 1) 8 cxd5 exd5 9 dxc5 i.xc5 10 0-0 leads to a Pawn Exchange Variation (7 dxc5 i.xc5 8 cxd5 exd5, etc.) where White has taken the unusual decision to develop the bishop to d3 rather than e2. As noted in the Introduction, Black equalizes with 1O...i.g4! 11 :cl d4!, Teichmann-lanowsky, Hastings 1895. 2) 8 :cl (or 7 :Cl lLlc6 8 i.d3) 8...cxd4 9 exd4 dxc4 (9 ... b6 10 cxd5 lLlxd5 11lLlxd5 Wxd5 120-0 i.b7 13 :el :ac8 14 i.e4 Wd7 15 d5 exd5 16 Wd3 "g4 17 i.xh7+
The Queen's Gambit Declined: 5 1./41

146

3) 8 ... b6 9 l:lcl .i.b7 10 cxdS exdS leads to positions discussed under ... b6 systems. White should be better. 9 exd4lDb4 10.i.e2 dxc4 11 .i.xc4 lDbdS Averbakh-Castillo, Thessaloniki 1988. 11...b6!? could also be tried.

=

E) 7 .i.e2!? (D)

This is the one move in this section that might have been under-estimated. Black does not appear to have any quick and certain route to equality. The modest bishop move allows White to keep some influence on the d-file. 7•••cxd4 After 7 ... lDc6 8 0-0, 8 ... dxc4 9 dxcS .i.xcs 10 .i.xc4, V.Ragozin-Makogonov, USSR Ch (Moscow) 1944, is yet another way to transpose to the 6 ... cS 7 dxcS .i.xcs 8 .i.e2 variation, while 8 ...cxd4 9 exd4 transposes to the next note. 8lDxd4 This poses the sternest test. 8 exd4 leads to an attacking IQP position for White, but in such positions the bishop is misplaced on f4, its ideal square being gS. There might follow:

1) 8 ... dxc4 9 .i.xc4 lDbd7 10 0-0 lDb6 11 .i.d3lDbdS 12 .i.eS ;t Skvortsov-Bitman, Moscow 1968. 2) 8 ...lDc6 9 0-0 dxc4 (9 ... b6 10 a3 .i.b7 11 l:lcl l:lc8 12 cxdS lDxdS 13 .i.g3lDxc3 14 l:lxc3 .i.f6 IS l:ld3 .i.a6 16 l:ld2 .i.xe2 17 ti'xe2lDe7 Bradbury-T.Sl/.lrensen, Copenhagen 1997) 10 .i.xc4 b6 (1O ... a6 11 .i.b3?! {II a3!?} 11...lDa512.i.c2bSI3lDeS.i.b7 14 a3 l:lc8 IS ti'd3 g6 16 l:lfel l:le8 17 l:ladllDds ~ P.Costa-C.Santos, Portuguese Ch 1997) 11 a3 .i.b7 12 ti'd3 l:c8 =Gloria-Campora, Bern 1993. 8•••lDbd7?! Missing a tactical trick. 8 ... dxc4 9 .i.xc4lDbd7 is a more accurate moveorder, aiming for the same equalizing position. 8 ...lDc6 9 cxdS exdS 10 O-o;t ti'b6? lllDdbS .i.e6?? 12 .i.c7 1-0 GirettiPerssinotto, Ceriano Laghetto 1997. Possibly a premature resignation, but Black is at the very least losing a lot of dignity after 12...ti'cs 13 a3, threatening to trap the queen with lDa4. However, one of the main databases suggests however that the game actually finished 12 a3 l:lac8? 13 .i.c7 1-0, as if 13 ... l:lxc7, 14 lDa4 ti'aS IS b4 wins material. 9.i.g3?! 9 cxdS! is strong, since 9 ...eS?! is met by 10 lDfS! exf4 (1O... .i.b4 11 .i.gS! lDb6 12 .i.xf6 ti'xf6 13 lDg3 leaves Black wondering how to regain the pawn) 11 d6.i.xd6 12 ti'xd6 fxe3 13 lDxe3, when White's pieces are much better mobilized in an open position. 9 •••dxc4 10 .i.xc4 lDb6 11 .i.d3 lDbdS Brestian-Murugan, Moscow OL 1994.

=

=

Alternatives to the Main Lines for White

147

=

F)

7 adS (D)

There are obvious comparisons to be made with the Exchange Variation (7 dxcS .i.xcs 8 cxdS), but this particular move-order is less accurate in that Black has more flexibility in how to recapture. 7.../oxdS 7...exdS?! gives White the opportunity, should he so desire, of entering the Pawn Exchange Variation with 8 dxcS. White could also consider 8 i.e2, for example 8...~c6 9 0-0 c4 10 ~S i.e6 11 ~xc6 bxc6 12 b3 WaS 13 ~a4 ltac8 14 bxc4 dxc4 IS ~b2 11'b4 16 Wc2;t J.Gonzales-H.Leyva, Cienfuegos Capablanca mem 1991, or 8...cxd49~xd4~4?! lOWb3~xc3 11 bxc3 WaS 120-0 l:td8 13 l:tfdl ± Chachalev-Rabara, Bratislava 1992. SlOxdS

8 i.g3 cxd4 9 exd4 WaS lO "b3 =B.Sadiku-Wiegner, Berlin 1996. S.....xd5 After 8... exdS: 1) 9 dxcS transposes to the Knight Exchange Variation. 2) 9 i.e2 1I'aS+ 10 Wd2 "xd2+ 11
i.b4+ (13 ... d4!?) 14
10 i.d3 lO a31Dc6 (10...ltd8 11 Wcl as 12 i.e2 b6 13 o-o;t Wislez-G.WinkIer, Brussels 1993) 11 l:tel WaS+ 12 Wd2 Wxd2+ 13 ~xd2 eS 14 i.g3 was assessed by Minev in Informator 33 as unclear. Korotkova-Strygina, Kaluga 1996 soon headed for the draw after 14...ltd8 IS i.c4
148

The Queen's Gambit Declined: 5 J.f4!

IIl£ld2?? :dS with a killing pinMinev.

1l•••l£lc6 12 a3 .tf6!

After 12...eS? 13 b4 We7 14.tg3 a6 IS We2 fS 16l£lxeS White wins a pawn, G.Goldberg-Chekhover, Kiev 1940. 13 b4 WhS 14 l:lel eS 15 .tg3 After the incorrect IS ...:dS? White won a brilliancy in Ribli-Chandler, Surakarta/Denpasar 1982: 16 Wc2 .te6 17 .te4 ± l:lac8 18 Wbl g6 19 l:[hdl l:lxdl 20 ~xdl .tg7 21 ~el fS 22 .txe6 bxe6 23 l£lxeS .tdS 24 Wd3 f4 2S l:lxc6! l:ld8 26 l:le8! 1-0. But roles would have been reversed after the even prettier IS ...l:le8! 16 Wc2 e4 17 .i.xe4 .i.f5! 18 .i.xfS l£ld4+ - Minev.

5 Black Avoids the Main Line: ... c5 Systems The rest of the book examines ways in which Black avoids the main line. In this chapter we consider lines where Black plays an early ...c5, but deviates later.

5.1 Black's alternatives on move 9

"c2

1 d4 dS 2 c4 e6 3 lDc3 lDf6 4lDf3 J..e7 5 J..f4 0-0 6 e3 cS 7 dxcS J..xcS 8 lDc69 a3 (D)

We are looking for alternatives to 9 ... 'Wa5, of which the most important are 9 ...J..d7 and 9 ...J..e7. First we dispose of the unquestionably weak lines. Our sections are: A: 9•••e5? 149 B: 9...a5? 149 C: 9...d4?! 150 D: 9...dxc4?! 150 E: 9...J..d6 151 F: 9.....e7 152 G:9••• 153 H: 9 ...J..e7 154 I: 9••• J..d7 155

:e8

A) B

In so many of the main lines, Black plays something like ... J..e7xc5-e7 and ...Wd8-a5-d8. If this gives the impression of being an inelegant waste of time, we may sympathize to some extent, although Black is of course trying to provoke weaknesses in the white position.

9•••e5? Is this pure bluff? Or even a misprint for 9 ... d4 10 0-0-0 e5? I can see no reason why White shouldn't just swipe the pawn with 10 lDxe5, e.g. 1O... lDxe5 11 J..xe5 lDg4 12 J..f4 d4 13 O-O-O! (Burgess). White played more respectfully in Sobolev-Purgin, Yaroslavl 1995, with 10 J..g5 d4 11 0-0-0 J..e6 12 exd4 lDxd4 13 lDxd4 exd4°o. B)

9...aS? This just eats a tempo. 10 :dl J..e7 11 J..e2 a4 12 e4 13 cxd5 exd5 14 exd5 +- MilesFranco del Valle, Seville 1993.

"as

150

The Queen's Gambit Declined: 5 J.f4!

C) 9 •••d4?! (D)

This is premature. The tactics favourWhite. 100-0-0 eS 1) 1O... dxc3 11 .:txd8 cxb2+ 12 ~xb2 .:txd8 13 .td3 .td7 14 g4 leaves Black with insufficient compensation for the queen, Gheorghiu-Voilescu, Romanian Ch 1977. 2) 1O.....e7 11 exd4 ~d4 12 ~d4 e5 13 tOf5! (13 .tg5 .txd4 14 /Od5 "d6 15 .txf6 gxf6 16 .td3 ~g7 17 o!tle3 b5! and Black is developing dangerous counterplay, Filip-Guimard, Buenos Aires 1964) 13....txf5 14 "u5 exf4 15 .td3 g6 16 "xf4 ± ECO. The text-move is a recent attempt to enliven Black's play, and is tricky, though one would suspect not wholly reliable. 11 tOxeS 11 exd4 exd4 "". II •••'ite8 12 o!tla4 tOxeS 13 tOxcS dxe3 Black is aiming for the sort of position where, although a pawn down, he has enough counterplay on the h7-bl diagonal and along the c-file to make it difficult for White to consolidate.

14.txe3 14 tOe4 is met not by 14... exf2? 15 tOxf6+ gxf6 16 .th6 ±, but 14.....e7! is unclear. White could try 14 fxe3!? 14....c6 IS h3?! This seems to be one preparatory move too many. It is the queenside rather than the kingside that White needs to be consolidating. 15 .te2 makes it more difficult for Black to justify the pawn sacrifice. White can then think in terms of .td4 and, if the knightmoves,.tf3. IS••..:te816 tOb3 .td7 17 iOd4 Wc7 18 ~bl .:tacK 19 .:tel Wb8 It is not so easy for White to consolidate. Mohandessi-Dutreeuw, Belgian Ch 1996 continued 20 .te2 /Od5 "", while 20 1Ifb3!? iOe4 is not clear either.

D) 9•••dxc4?! 10.txc4 (D)

B

Exchanging on c4 before White has played .te2 is a pointless loss of tempo. Examples: 1) 10... a6 11 0-0 tOIt5 (11. ...td7 12 .:tad 1 "e8 13 o!tlg5?! {13 .ta2!?;t} 13 ....te7 14 "d3?! .:td8 15 .tc7 .:tc8

Black Avoids the Main Line: ... c5 Systems

151

16.id6.ixd617.xd6.e718~ge4

'/2-'12 Nickoloff-I.Findlay, Toronto 1997) 12 :adl .e7 13lbe4 ~xf4 14 exf4.ia7 IS :fel :d8 16lOegS g6 17 .c3 bS 18 .idS .ib7 19 .ixc6 :ac8 20 :xd8+ .xd8 21 lDxf7 ~xf7 22 lDgS+ ~g8 23 Wh3 We7 24 .txb7 :c2 2S :dl :xf2 1-0 (26 :d8+ wins) Nogueiras-J.Diaz, Santa Clara 1991. White was too far ahead in development throughout. 2) 1O... .id7 11 0-0 lDhS and now 12lDgS?! g6 13lDge4lDxf4 14lDxcs .gS IS .idS lDxg2 16 .txg2 .xcS can't be bad for Black, even though he later lost in Kallai-Eberlin, Bundesliga 1990/1. 12 :adl is better, as in the Nogueiras game - the centre needs to be watched. 3) 10....td7 11 0-0 :c8 12 :adl .te7 13 e4 WaS 14 eS lDhS IS .tcl :fd8 (1S ... lDxe5? 16lDxe5 .xeS 17 :xd7 :xc4 18 :xe7 +-) 16.e2! g6 17 b4 ± Vera-Borges, Cuban Ch (Matanzas) 1997. 4) 1O...•e7 11 .ig5 h6 12 .th4 a6 13 .ta2 :d8 14 lDe4 g5 and now 15 lDxc5 gxh416 :cl b617lDa4 .ib718 ~xb6 h3! 19 gxh3lDb4 20 axb4 .txn led to a short but sharp draw in SubaS.Mirlcovic, Belgrade 1984. White must surely have improvements, starting perhaps with the simple 15 .ixg5. E)

9•••.td6(D) This move has a slightly oldfashioned feel to it, and the average age of the games below confinns the impression. Black is not too concerned with minor pawn weaknesses; he just wants to blunt the white initiative with exchanges.

lO.txd6 The obvious choice, but the unexplored 10 .ig5!? might be more testing. lO••.•xd6 11 :dl If immediately II cxdS exd5, the black bishop can get to g4. 11••.•CS! A new improvement in an old line. 1l ...:d8 12 .ie2.e7 13 cxd5 exdS 14 0-0 .te6 15 .a4 :ac8 16 ~4 lDxd4 (16... a6 17lDxc6 :Xc618 :d4 'fIc7 19 :fdl :cd6 20 h3 h6 21 .tn ~ Gipslis-Naglis, Moscow 1970) 17 :xd4 a6 18 :fdl :d7 19.tn .c5 20 h3 h6 21 .b3 :cd8 22lDa4 'fIc7 23 'fIb6 .xb6 24lDxb6 :d6 25 ~a4 :6d7, and in Capablanca-Yates, Hastings 1929, the players settled for a draw after 26 ~b6 :d6. Capa was evidently in an easy-going mood that day. He could certainly have squeezed a lot harder with, for example, 26 b4, with standard play against the isolani.

12b4 12 cxd5 lDxd5 13 :c 1 lDxc3 14 .xc3 .e7 leaves White little to work with. 12.•••d6 Black has sacrificed two tempi, and in return has succeeded in luring

152

The Queen's Gambit Declined: 5 j"f4!

White's queenside pawns forward before they are fully backed up by pieces. The immediate plan is to play for ...as. Korotylev-Duchov, Moscow 1996 was agreed drawn after 13 cxdS exdS 14 tl)bS "e7 IS "cS. F)

9.•:fle7 (D)

w

A straightforward plan of development, which avoids a lot of the to-ing and fro-ing of the .....as lines. The drawback is that a pin with ~gS could be awkward.

lO:dl 1) 10 ~gS is perfectly playable. After 1O...:d8, White can choose between 11 :dl and 11 0-0-0, transposing to lines considered elsewhere in this section. 2) 10 cxdS exdS gives an Exchange Variation, where Black's extra ... 'ille7 is probably more useful than White's a3. White kept a slight edge in SubaLobron, Dortmund 1983 after 11 ~e2 ~e6 12 0-0 l:.ac8 13 :tfdl l:tfd8?! (13 ... a6! Suba) 14 tl)a4 i.d6 IS ~xd6 'ifxd6 16 tl)c3. 3) 10 O-O-O!? might well tum out to be the most promising line against

9 .....e7, as well as against 9 .....aS. If the d-file is cleared suddenly, White's rooks are cOMected. After 1O...l:td8 11 ~gS d4 12 tl)e4 (the simpler 12 exd4 tl)xd4 13 tl)xd4 ~xd4 14 ~d3 gives a comfortable lead in development, but the text too is worth trying) 12...eS 13 ~d3 ~g4 14 h3 ~h5, Hemdl-Martinovsky, Chicago 1994 continued IS tl)xf6+?! gxf616~xh7+~g717tl)b4 ~xdl 18 l:txdl "d7 19 ~fS "c7 and White had over-extended himself - one should beware the pawn snatch. IS ~xf6! gxf6 16 tl)g3 ~g6 (16 ... ~xf3 17 gxf3 dxe3 18 tl)fS ±) 17 e4 followed by h4 may already be positionally lost for Black. 4) If 10 b4: 4a) Black can try the piece sacrifice 1O...tl)xb4 11 axb4 ~xb4 12 tl)d2 ~d7 13 ~e2 :fc8 14 0-0 bS, and it would be a brave annotator who described it as anything other than unclear. In the game Van Wely-Ziatdinov, Amsterdam 1994, White was starting to gain control after IS ~eS dxc4 16 ~f3 tl)dS 17 fudS (17 ~xdS?! exdS I8fudS"xeSI9fub4aS!) 17...exdS 18 ~d4. Black's queenside pawns are not as mobile as they might look. 4b) Black does not have to sacrifice, however. 10... ~d6!? should be perfectly playable, and if 11 ~xd6?! 'ii'xd6 Black is a tempo ahead of the Korotylev-Duchov game, Line E above. Or 11 ~gS dxc4 12 ~xc4 as 13 bS tl)eS.

1O•••l:td8 1O...d4? 11 tl)xd4! (11 exd4 tl)xd4 12tl)xd4eS) 11...tl)xd412exd4~xd4 13 l:txd4 eS 14 lle4 gives White a clear advantage - Ziatdinov. 11~g5

Black Avoids the

Mai~

1) After II b4, Ziatdinov tried the piece sacrifice again, but with the rooks in more favourable positions for him; his rook is centralized on dS, while on dl White's rook no longer covers the queenside. 11...~b4!? (11.. ..i.d6 12 .i.g5 dxc4 13 i.xc4 as 14 b5 1&5 15 ~xe5 J.xe5 16lhdS+ "xdS 17 ~e4 ± Miles-Kogan, USA 19S0; this time the inserted rook moves favour White) 12 axb4 J.xb4 13 ~2 J.d7 14.i.e2 dxc4 (14 ...l:lac8!? Miles) 15 .i.0 e5 16 J.g5 J.f5 17 "xf5? (17 e4! 00 Ziatdinov) 17...J.xc31S"c2~ 1ge4 b5 20 0-0 "b2 + Miles-Ziatdinov, Montpellier 1994. 2) 11 J.e2 and now: 2a) 11...h6 12 cxd5 exd5 13 0-0 J.e6 14 ~a4 J.d6 15 J.xd6 l:lxd6 16 li'lc5, Ribli-Gligoric, BledIPortoroz 1979, and now Gligoric assesses 16....:tcS as equal, though White might be able to try for a little more with 17

Line: ... c5 Systems

153

12l:lxdS+ 12 i.xc4 l:lxdl+ 13 "xdl h6 14 .i.h4 a6 15 i.d3 "dS 16 0-0 .i.e7 Petrosian-M.Yudovich, Moscow 1967. 12.....xdS 13 i.xc4 i.e7 14 0-0 i.d7 15l:ldl with an equal position, Doroshkevich-Liberzon, USSR Ch (Tbilisi) 1967. But 10 0-0-0 is critical.

=

"cS

G)

.

9•••l:leS (D)

~xe6.

2b) Black can of course avoid the isolani with 11...dxc4, but Black's pieces are a little awkwardly placed when compared with the 9 i.e2 lines (Chapter 4. 1, Line D). After 12l:lxdS+ "xdS (12 ... li'lxdS!?) 13 0-0 i.d7 14 i.xc4 a6 15 l:ldl i.e7 16 h3 l:lcS 17 "e2 "as IS e4, White held the initiative in Stean-B.Toth, Biel19Sl. 11•••dx:c4 Safe, ifa little dull. l1...d4?! favours White after 12 ~e4 dxe3 (12 ...i.b6?! 13 c5 i.aS+ 14 b4 ~xb4 15 axb4 i.xb4+ 16 ~fd2, Chemin-Plachetka, Prague 19S9, is an implausible sacrifice) 13 ':xdS+ ~xd8 14 ~xc5 "xc5 15 i.xe3 with the bishop-pair in an open position, Lucas-Hochgriife, Bargteheide 19S9.

This could lead to positions rather like the main line, except that Black has omitted ... "dS-a5-dS and White has omitted ~f3-d2-b3. The difference probably favours White.

100-0·0 1) 10 .i.e2 transposes after 1O... e5 11 i.g5 d4 (11...e4 12 ~2 d4 13 ~xe4 i.f5 14 i.d3 ±) 120-0-0; it is unlikely that Black will want to try 1O...dxc4. 2) White missed an important tactical trick in the game Pinter-Hurme, Helsinki 19S3: 10 :dl e5 11 i.g5 d4 12 ~5 i.e7 13 ~xe7+?! (13 i.xf6 =) 13 .....xe7 14 i.e2?? d3! 15 i.xd3 e4 winning a piece. So the knight does not always stand we)] on 0.

154

The Queen's Gambit Declined: 5 i..f4!

3) 10 cxdS exdS 11 .i.e2 is enough for a slight edge, as there does not seem to be any way for Black to take advantage of the e-file, e.g. 11.. ..i.g4 120-0, or 11...d4?! 120-0-0. SzaboIvkov, Hungary-Yugoslavia 1966 continued 11.. ..i.e6 12 0-0 h6 13 l:r.fd1 "e7 14111a4;1;. 10...eS 11 .i.gS Not 11 cxdS? exf4 12 dxc6 "c7! Forintos. 11•••d4 12 .i.e2 Calmly completing his development, and waiting for Black to show his hand. The more turbulent 12111dS .i.e6 (12 ... .i.e7! - Forintos; then perhaps 13 .i.xf6 .i.xf6 14 .i.d3 ;1;) 13 .i.d3 .i.xdS 14 cxdS "xdS IS .i.xf6 gxf6 16 .i.xh7+ c,tJf8 17 .i.e4 is also good for White, Forintos-Vaiser, Tallinn 1986. 12.....&5 13 exd4111xd4 14111xd4 .i.xd4 IS lllbS ± Langeweg-Van der Wiel. Dutch Ch 1983.

H) 9....i.e7 (D)

that on the current state of theory. Black has no need to fear 10 l:r.dl transposing into the Old Main Line. while if White angles for the New Main Line with 100-0-0. Black could consider moves other than 1O......aS. Black can. for example. try 1O....i.d7!? and if 11 cxdS?! lllxdS 12111xdS exdS 13 l:r.xdS?! l:r.c8 then the dark side of queenside castling clearly reveals itself. Naturally White does not have to be so compliant. but if. for instance. 11 c,tJbl. then again Black need not transpose to the Main Line with 11.. ...aS. but could try for more immediate pressure along the c-file with 11....l:lc8. After l1111gS, 11...h6?! 12 h4 is dangerous. but 11...l:r.c8 12 c,tJb1 g6 is unclear. All this remains. surprisingly, untested. at least according to the database. 10.i.e2 10 .i.d3 dxc4 11 .i.xc4 transposes. White avoids seeking the main lines. but reaches one of the lines of the 9 .i.e2 variation (Chapter 4.3, Line DS). which on the whole is uninspiring. 10...dxc4 11 .i.xc4.i.d7 The most reliable. Black came horribly unstuck trying to fianchetto in I.Farago-I.Almasi. Hungarian Cht 199213: 1l...a6 12 .i.a2! (12 l:r.d1.i.d7 130-0 "c8 14 .i.d3lllaS ISllleS i.bS = Mikhalchishin-Bonsch. Budapest 1989) 12 ... bS 13 0-0 i.b7 14 l:r.adl IS i.b1 ;!; l:r.ac8?? 16 l:r.d7 i.a8 17 .i.gS g6 18 l:r.xe7111xe7 19 .i.xf6 +-. Elementary tactical themes in the i.f4 Queen's Gambit... 120-0 12 l:r.dl transposes to 8 a3 lllc6 9 .i.e2 dxc4 10 "'c2 .i.d7 11 l:r.dl .i.e7

"as.

"as This unpretentious move might in future be destined to become the main line. replacing 9......aS. The point is

Black Avoids the Main Line: ... c5 Systems 12 -*.xc4 (Chapter 4.3, Line 01). White might perhaps be able to claim a small advantage, since if 12... ltlh5, 13 -*.e5. 12••• 12 ... m!? 13 :rdl "'6 14 -*.d3 14 -*.a2!? 14... h6 15 b4 .!DdS 16 ltlxdS exdS 17 'Ii'd2 -*.g4 18 -*.e2 :fd8 with equality, Mirkovic-Cvetkovic, Yugoslavia 1985.

:c8

155

12 Wxc5ltlxf4 13 exf4 :e8+, Forintos, leaves Black with too much play for the pawn, and 11 -*.e2 d4, Bangiev, is satisfactory for Black. 10... White can now choose between: 11: 11 -*.gS 155 12: 11 -*.e2 156 13: 11 o:dS 157

:c8

11) l1-*.gS (D)

I) 9...-*.d7 (D) B

w

Finally we come to the most popular of Black's 9th move alternatives to 9 ... 'it'a5. With the rook coming to c8 in minimum time, White would be foolhardy rather than brave in castling queenside, and this in itself must count as a minor positional victory for Black. On the negative side, the bishop is not necessarily well placed on d7, particularly if the position is opened up. White generally exchanges on d5 at some stage. IO:dl After 10 cxd5 exd5, 11 :c 1 lIc8 transposes. Instead 11 .!Dxd5?! ltlxd5

11...-*.e7! Black's simplest is to let the pawn go. Forintos's suggestion of 11 ...dxc4 could be very uncomfortable to play after 12 -*.xf6 gxf6 13 -*.xc4. 12 -*.xf6 12 cxd5ltlxd5 =. 12...-*.xf6 13 cxdS exdS 14 :xdS 'We7! Black sought immediate compensation for the pawn in terms of creating direct threats in Forintos-Averkin, Hungary 1969: 14...ltle7?! 15 :d2 -*.xc3 16 bxc3 Wc7 17 -*.d3 g6 180-0 Wxc3 19 -*.xg6! 'fIg7 20 Wbl ;1;. However, Black's main assets are active piece-play, backed up by the bishoppair and a lead in development, and it

156

The Queen's Gambit Declined: 5

makes sense to build on this by activating the remaining pieces. rather than by trying to cash in immediately. 15 .i.d3?! g616 0-0 .i.e617 l:Lb5 Not the ideal square for the rook. 17•..l:Lrd8 18 .i.e4 a6 19 l:Lb6 ti)a5 20 "a4 "cS 21 l:[b4 "c7 22 ti)d4 .i.c4 +Browne-Christiansen. USA Ch 1990.

12) ll.i.e2 (D)

B

1l... dxc4 After 11.. ..i.e7. 120-0 dxc4 transposes. as in Ruban-Geller below. while 12 cxdS ti)xdS 13 ti)xd5 exd5 14 Wb3 1i'a5+ is considered unclear by Ruban. 120-0 A refinement in move-order to discourage Black from early activity. If now. for example. 12 .....a5 13 ti)e4 tDxe4 14 "xe4 (14 l:Lxd7?! tDf6 15 l:Lxb7 tDd8!). Black's kingside is exposed. 12 .i.xc4 has also been played: I) 12.....a5 13 0-0 l:Lfd8 14 e4 (14 tDe4.i.e7 15 ti)d6.i.xd6 16 .i.xd6 b5 17 .i.a2 tDd4 18 Wd2 tDxf3+ 19 gxf3 Wxd2 20 l:Lxd2 .i.c6 =) 14...1Oh5 15 .i.cl! (15 .i.g5 .i.e7 =) 15 .....c7?!

~f4!

(15 ....i.e7 16 "e2;t) 16 tDbS "bS 17 .txe6.txf2+ IS"xt2.i.xe619l:LxdS+ ti)xdS (19 ...l:Lxd8 20"h4 ±) 20 tDxa7 ± Gelfand-Yusupov. Linares 1991 (notes based on those by Gelfand and Kapengut). 2) On 12.....e7. 13 lOgS ;t is probably best. Instead. in Kohlweyer-Konings. Netherlands 1995. Black took the initiative after 13 .tg5 .tb6 140-0 (141Oe4 .ta5+ 15 ~e21ObS!) 14... h6 IS .i.xf6 gxf6 16 tDe410bS 17 "e2 f5. 12....te7 12... a6 13 .i.xc4 lOh5 (13 ...b5 14 .td3.i.e7 ISlDeS!;t EGO) 14 lOgS g6 151Oge4.i.e7 16 .i.d6 .i.xd6 17 l:Lxd6 tDbS IS .te2 lOg7 19 l:Ld3 "e7 20 "d2 .i.c6 21 l:Ldl ± Kramnik-Yusupov. Dortmund 1997. 13 .i.xc4 Wb6 13 .....a5?! 14 "b3 leaves the bpawn weak. Portisch-Westerinen. Nice OL 1974 continued 14...l:LfdS 15 "xb7 e5 16 .i.g3 l:[b8 17 "a6 "xa6 IS .i.xa6 e4 19 tDe5 (19 .i.xbS? tDxb8) 19... l:lxb2 20 l:Lxd7 :xd7 21 tDxc6 ±. 13 ... lOh5?! 14 .td3 g6 15 .i.h6 ± Ruban. 14tDgS _ 1) 14 .i.a2!? l:LfdS IS tDd2 ;t. Ruban. with the knight possibly coming to c4. may be more accurate. 2) 14 .td3!? threatening .i.xh7+. is a try. but after 14...l:Lfd8 IS tDgS h6 White must be careful not to get carried away with the attack: 2a) 16.i.h7+?!~f8171Oxf7~xf7 18 "g6+ 'it?f8 19 .i.xh6 wins attractively for White if Black takes the second piece. for example. 19... gxh6? 20 "xh6+ ~f7 21 .i.g6+ ~g8 221Oe4! lOeS 23 lOgS .i.f8 (23 ....i.e8 24.i.h7+ ~h8 2S .tfS+ ~gS 26 .i.xe6+) 24

Black Avoids the Main Line: ... c5 Systems .i.f7+ t:j)xf7 25 'ifg6+ J..g7 26 'ifxf7+ ~h8 27 ':d4 'ifxd4 28 exd4 +-, but

19... t:j)e8! is a good and surprising defence, however, as 20 ':xd7 ':xd7 21 lIhe6 t:j)f6! covers everything. 2b) 16 t:j)ge4 might keep chances for a slight edge. 14..•h6 14 ...Itfd8? ISlike4 ± Ruban. IS t:j)ge4 Not 15 ':xd7?? hxg5 -+ Ruban. IS••• t:j)aS (D) Ruban gives some detailed analysis of IS ... t:j)xe4 16 Wxe4, the basic verdict being that White is slightly better after, for example, 16...lDaS 17 ':xd7 ':xc4 18 Wf3 J..f6 19 J..d6 ':d8 20 ':xd8+ Wxd8 21 J..b4. There is no reason to quote the analysis in detail here, as the text is simpler and better. Why hring the white queen into play?

157

19 J..xh6? J..xfl 20 J..xfB J..c4! wins the exchange - Ruban. 19•••WcS 20 l:.xe6!? Stirring up the mud. Ruban gives 20 lDxb5 Wxb5 21 Wdl Wxb2 (2l...J..g7 22 b4!) 22 .lxh6 as slightly advantageous for White, but this is highly questionable. After 22 ... ':fd8 White's bishop-pair does not accomplish much, while his queenside is on the point of collapse, with Black's a- and b-pawns ready to fill the vacuum. 20••• Wc4! 21lDxbS Wxc2 22 hel fxe6 23 .txg6 q;g7 lh-lh RubanGeller, Sochi 1989. Black was generous in agreeing the draw. With the c- and d-files fully open, the exchange counts for more than White's bishop-pair and extra kingside pawns. In addition, White's queenside pawns remain weak. 13) 11 adS exdS (D)

16lDxf6+ J..xf6 17 .lal .lbS 17 ....lxc3!? 18 J..bl fS 19 bxc3 .lbS, Ruban, is also to be considered. Black's centre and kingside look fragile at first, but the picture will soon change if Black can play ...lik4 and take control of the long light-squared diagonal. 18 .lbl g6 19 ':d6

By a process of elimination, this pawn exchange is White's one realistic try for an advantage. As already noted, White can exchange a move earlier (10 cxdS exdS 11 ':dl ':c8). 12J..e2

158

The Queen's Gambit Declined: 5 l.f41

White aspires to the long-term positional pressure that the Pawn Exchange Variation offers. The 1i'c2, l:[dl development is less economical than %lcl and keeping the queen at home, especially since c2 is not a particularly safe square for the queen, but on the other hand Black will soon have to play ... ~e6, losing a tempo. 12...~e6 12...1i'e7 13 0..0 ~e6 transposes. 12...i.e7 13 0-0.b6 14 1i'bl ~e6 induces White to playa queen move before Black has to play ...~e6. YrjOlii-Dutreeuw, Debrecen Echt 1992 should have been uncomfortable for Black: IS ~d3 tOaS 16 ~eS g6?! (16 ... h6 ~) 17 ~d4 1i'd8 and now White went for the sacrificial option with 18 tOgS1Dc6 19 ~xg6 hxg6 20 tOxe6 fxe6 21 1i'xg6+ ~h8, following up with 22 e4 tOxd4 23 lIxd4 (23 lId3? tOh7 24 l:[h3 ~h4) 23 ...•e8! with an unclear position. At first I assumed that 22 lId3 improved, but 22 ...1i'e8 231i'h6+ ~g8 24 e4 .hS! (Burgess) defends. However, Black's whole structure seems so unconvincing that sacrificial measures should not be necessary. I do not see anything immediately crushing, but 18 e4!? (instead of 18 tOgS) opens up the position in White's favour, for example 18...dxe4 (l8 ...tOe6 19 ~xf6 ~xf6 20 cxdS) 19 tOxe4 tOdS (19 ... tOc6 20 ~xf6 ~xf6 21 i.a6!) 20 b4 tOc6 21 ~b2 ±. 130-0 'fie7 13 ...i.b6?! 14 1i'a4 reaches a position that can occur through the Pawn Exchange Variation, except that White now has the rook on dl rather than on c 1. The difference favours White, as Black has no time to put the queen

under pressure with 14...i.d7 in view of IS tOxdS. In Sturua-B.Toth, Biel 1996, Black was very passive after 14...1i'e7 15 tOes lIfd8 16 tOxc6 lIxc6 17 i.eS tOd7 18 i.g3 tOes 19.b4 ±. 14 tOeS Angling for the endgame. The more complex 14 ~gS gave Black sufficient play to compensate for his pawn weaknesses after 14 ... lIfd8 IS1i'a4 (1S lId2 h6 16 ~xf6 1i'xf6 17 lIfdl tOe7 ao Bangiev) IS ...h6 16 i.h4 a6 (16... gS!? Bangiev) 17 lId2 gS 18 ~g3 ~fS 19 :tfdl ~e4 in Bangiev-Yusupov, Bundesliga 1992/3. 14 lId2 :fd8 IS lIfdl h6 16 tOeS ~d6 17 tOxc6 bxc6 18 ~xd6 lIxd6 ao A.Martin-Dutreeuw, London 1988, is a slightly less flexible mode of development for White. 14•••~d6 15 tOxe6 bxe6 16 ~xd6 'fixd6 17 lId2 'fieS 18 'fia4 lIe7 19 lid lIb8 20 ~f3 hS 21 g3 ~g4 22 ~g2 ;t Van der Sterren-M.Martens, Dutch Ch 1995. White can make Black suffer for a very long time in this type of position. In conclusion, we have noted no fewer than nine alternatives to 9...'fiaS. Most of these are clearly inferior, but 9 ...•e7, 9 ...~d6, 9...~e7 and 9... ~d7 all seem viable, although not necessarily sufficient for speedy equality. 9 ...~d7 has been popular, but the exchange of pawns on dS leaves White with a nagging positional advantage. If 9...~e7 is really as effective as implied in the notes, then this is one of the best pieces of news for Black in the book, although Black still has to be prepared to enter the Old Main Line after 10 lIdl1i'aS. The most important

Black Avoids the Main Line: ... c5 Systems position for further investigation in Ihis section is the one after 9 ... j.e7 10 O-D-O j.d7!?

A6: 8 •••a6 A7: 8 ••• j.e7

159

162 162

A1)

5.2 Black's alternatives on move 8 I d4 dS 2 c4 e6 3 ~c3lOt'6 4lOt'3 j.e7 5 .tf4 0-0 6 e3 c5 7 dxc5 j.xcS 8 'ifc2 ur8a3 Mostly these alternatives involve a tempo-wasting ...dxc4, and only a desire for completeness keeps these sub-optimal variations in the book. There are, however, a few lines of independent significance to be considered. Our coverage divides as follows: A: Alternatives after 8 a3 159 B: Alternatives after 8 'ii'c2 163 A)

After 8 a3 (D):

8...aS?! This misses the point. White plays a3 not to prepare b4, but rather to stop Black using b4 for a piece. White should simply develop, with, for example, 9 'iVc2 ~6 10 :d1, etc. R.Letelier-l.Clement, Marostica 1990 was agreed drawn after 9 cxd5 exd5 10 .te2 ~6 11 0-0 h6 12 :cl, although even this seems over-friendly on White's part. A2)

8...'''aS?? Possibly played very early in the morning. Though b4 is not top of White's agenda, ifit wins a piece, then play it. 9 b4 j.xb4 10 axb4 'iVxb4 11 1t'd2 ~6 12 cxd5 exd5 13 .te2 and White wins, Chabanon-C.Bouton, Clichy 1992. A3)

8... b6 (D)

First two very minor alternatives, then the marginally playable: AI: 8 .••aS?! 159 A2: 8 ••:"aS?? 159 159 A3: 8 ...b6 A4:8...~ 160 AS: 8 ...dxc4 160

The natural reaction is now to exchange pawns and force Black to

160

The Queen's Gambit Declined: 5 J.f4!

block d5 before the bishop takes control of the long diagonal. 9cxdS 9 Wc2 .i.b7 10 lIdllbbd7 11 cxd5 exd5 12 .i.e2 a6 = Forintos-Stefansson, Budapest 19S9. Black's pieces tum out to be very flexibly placed. White did not help his cause by playing 13 b4? (13 0-0), and after 13 ... .i.e7 14lbd411c8 15 Wb3 b5 16lbf5 lieS 17 lbxe7+ Wxe7 IS 0-0 lbb6 Black was already starting to take advantage of White's weakened queenside. 9•••exdS 9 ...lbxd5?! 10 lbxd5 Wxd5 11 Wxd5 exd5 12 lIcl a5 13 .i.b5 .i.a614 .i.xa6 lbxa6 15 ~e2leaves Black with seriously weakened pawns for the endgame, Miles - Altan-Och, Moscow OL 1994. 10 .i.d3 .i.b7 11 0-0 h6 12 b4 12 ':cl!? 12....i.d6 13 .i.xd6 "xd6 14 lbbS "d8 15 :lel lbc6 16 lbbd4 "d6 17 lbfS "d8 18 "84 ± Van WelyStefansson, New York 1994. Compared with the Pawn Exchange Variation, having the bishop on b7 rather than on e6 is no advantage for Black. A4)

8•••lbe4 Here there is just a single reference in ECO to go on. Adorjan-Sibarevic, Banja Luka 1983 continued 9 'ilfc2 'ilfa5 10 cxd5 lbxc3 II bxc3 exd5 12 .i.d3 h6 130-0 lbc6 14 lIabl WdS 15 l:fdl 'ilfe7 =. White might want to consider instead 9 lbxe4!? dxe4 10 WxdS lIxd8 11 ibd2 f5 12 .i.e5 (or even 12 g4), when White's queenside pawn majority will be more effective

than Black's kingside majority, constrained by the doubled e-pawns. A5)

8...dxc4 Leaving the choice of 9 .i.xc4, when White is a tempo ahead of the 8 .i.e2 line, or the more spirited 9 Wc2: ASl: 9.i.xc4 160 AS2: 9 'it'c2 161 A51) 9 .i.xc4 (D)

White's extra tempo is only a2-a3 though, and Black can attempt to prove that this little pawn-push is not such a great barrier to his attempts to consolidate. 9.....xdl+ For 9 ... a6 10 Wc2, see 9 Wc2 a6, section A52. 9 ...We7100-01ldSll Wc2lbc612 lIadl b6 13 lIxdS+ lbxdS 14 lIdl a5 15 .i.g5 .i.b7 16 .i.a2 allowed White to build up the standard kingside attack in Roiz-Bodnar, Sao Lorenco jr 1995. 10 lIxdl 86 11 .i.d3 Taking the sting out of the advance ... b5.

Black Avoids the Main line: ... c5 Systems 1) 11 0-0 and now 11...b5 12 .i.e2 .i.b7 13 .i.d6 i.xd6 14 :xd6 lOc6 liz-liz, Portisch-St8hlberg, Erevan 1965, is hopelessly dull, but satisfactory for Black. Alternatively, 11 ...lObd7 12 lOgS .i.e7 13 lOge4 lOxe4 14 lOxe4lOb6 15 i.b3 .i.d7 16 .i.d6;!; LechtynskyPrandstetter, Czechoslovak Ch 1986. 2) lllOes, the recommended move in the 8 .i.e2 line (thus with White's pawn still on a2, Chapter 4.4, Line A 11) is less appropriate here. After II...lObd7 12 .i.e2lOxes 13 i.xeS b5 14 i.f3 l:a7, Black would be ready to force a favourable pawn exchange with ... b4, and if 15 b4 .i.e7, Black can aim to consolidate with ...:d7. 3) 11 b4 .i.e7 12 i.d6 .i.xd6 13 :xd6 bS 14 .i.b3 i.b7 15 lOe5, and Black eagerly gobbled the poisoned pawn in Knezevic-Halldorsson, Reykjavik 1984: 15 ... .i.xg2? 16 :gl .i.h3 17 e4 a5 18 lId3 axb4 19 axb4 lIal+ 20 i.dl +-. Simply 15 ... lIc8 keeps an acceptable position. H ...lObd7 l1...b5 12lOe4 ± Miles. 120-0 After 12.!Od2 b5 13lOde4 i.b7 14 f3 .!Ods IS lOxdS i.xdS, PetrosianA.Oliviera, Buenos Aires 1964, one suspects that White won because he was World Champion, rather than as the result of any gains made from the opening. 12••• b5 l2 ... b6 l3lOe4 i.b7 14lOxc5.!OxcS 15 .i.e2 .i.d5 16lOe5 b5 (anyway!) 17 lOd3 lOxd3 18 i.xd3 ;!; PekarekPrandstetter, Czechoslovak Ch 1986. 13 lOe4 .i.b7 14 lOxc5 lOxc5 15 .te2 ;!; Miles-Li Zunian, Lucerne Wcht 1985.

161

White's bishop-pair ensured him a slight but enduring pull in the endgame. The rooks soon came off, and Black proved unable to hold the position . A52) 9 'ilc2 (D)

A sharper approach, attempting to refute Black's play.

9.••a6 1) 9 ... .!OdS?! 10 O-O-O! "as 11 lOxd5 exdS 12 lOgS g6 13 lIxd5! .i.xe3+ 14 fxe3 "xdS IS .i.xc4 gave White a winning attack in AdorjanR.Behling, Lugano 1983. The game finished 15 .....f5 16 "c3lOc6 17 g4 "a5 18 b4 "d8 19lOe4 lIe8 20 lOf6+ f8 21 lIdll-O. 2) 9 ... bS?! 10 lOxb5 "a5+ l11Oc3 lOdS 12 .i.xc4lOxf4 13 exf4.i.b7 14 lOgS g6 IS 0-0 leaves Black with very questionable compensation for the pawn, Muse-Lechtynsky, Budapest 1986. 10i.xc4 10 lid llObd7 11 .i.xc4 transposes. lO... .!tJbd7 1O... b5 11 .i.d3lObd7 (11...i.b7 12 i.g5! Suba) 12 lIdl ;!; leads back to

162

The Queen's Gambit Declined: 5 i.,f4!

the main line. In I.Farago-W.Schuett, Velden 1995, White gained nothing from the liquidations after 12 liJe4 .i.b7 13 :dllilxe4 14 .i.xe4 .i.xe4 IS "xe4 lilf6 16 lIxd8 lilxe4 17 l::txa8 ':xa8=.

nlldl 1) 11 e4lilg4 (ll...lilhS!?) 120-0 "f6 13 .i.g3 bS 14 .i.a2 .i.b6 IS eS "g6 (ls ...lilgxeS?? 16lilxeSlilxeS 17 "e4 +-) 16 "xg6 hxg6 17 lIfdl.i.c7 18 lIel .i.b7 19 lIadililb6 20 lilgS;!; Stohl-Bus, Hradec Kralove 1988. 2) 11 0-0 bS 12 .i.d3 "b6 13 ':'fdl .i.b7!? 14 .i.xh7+ lilxh7 15 ':'xd7 :ac8 16 ':'adllilf6 17lilg5!? "c6 18 ':'xb7 "xb7 19 .i.eS ':'fe8 20 .i.xf6 gxf6 21 "h7+ ~f8 22 "h6+ ~e7 23 lilge4 ':'g8 24 "xf6+ ~e8 2S g3 .i.e7 is a complicated sequence leading to an unclear position, Maiwald - SchmidtSchiiffer, German Ch 1993. n •••bS 12.i.d3 Wb6 13 lDes After 13 b4 .i.e7 14 e4: 1) Spassky notes that Black can't develop the bishop due to 14....i.b7? IS eSlildS 16 fudS .i.xdS 17 .i.xh7+ ~h8 18 ':'xdS exdS 19 "fS with a winning attack. 2) Suba gives some analysis leading to an unclear position after 14...a5 15 eS lilhS 16 .i.e3 "b8 17.i.xh7+ ~h8 18 .i.e4 axb4, and now 19 axb4 :a3 20 "b2 ':'a6, etc. However, White can just play 19 .i.xa8 in all this! 3) Black's correct plan would seem to be 14...eS! ISlilxeSlilxeS 16.i.xe5 .i.b7 followed by ...:ac8, when White will find it very difficult to disentangle, despite his extra pawn. Once again, White must be very careful about playing an early b4, in view of the weakening of the c-file.

13•••h6 14lDe4! Black is under great pressure. Less effective is 14 0-0 lilxeS IS .i.xeS lilg4 (IS ....i.b7 16 .i.xf6 gxf6 17 .i.e4 ;!; Suba-Spassky, Thessaloniki OL 1984) 16 liJe4 lilxeS 17 "xeS "xeS 18 lilxcS :a7 19 b4 lild7 20 lilxd7 1/2- 1/2 Piket-Beliavsky, Groningen 1992. Suba recommends 14 "e2, in view of 14....i.b7 IS lilxd7lilxd7 16 "hS lilf6 17 "h4, but 14 ....i.e7 keeps a solid position. Also, on another Suba suggestion, 14 lilxd7, Black can improve over 14 ....i.xd7 IS .i.eS .i.c6 16 .i.xf6 gxf6 17 "e2 ;!; with 14... lilxd7. 14•••.i.b7 Mistakenly given as equal by Suba. It is difficult to see reasonable alternatives, as if 14....i.e7, IS lilc6, or 14... lilxeS IS .i.xeS lilxe4 16 .i.xe4 ':'a7 17 b4! .i.e7 18 .i.d4 winning the exchange. 15 lilxd7 lilxd7 16lilxc5 lilxc5 16.....xcS?? 17 .i.h7+ ~h8 18 ':'xd7+-. 17 .i.h7+ ~h818 ':'d6 Wa719 b4 with a clear advantage for White. A6) 8•••a6 9 Wc2 dxc4 transposes to the line we have just been looking at, which White can feel happy about. A7) 8•.•.i.e7 This is the one fully satisfactory alternative to 8...lilc6, but its effects are almost entirely transpositional. 9 "c2 lilc6 leads to Chapter S.I, Line H above (9 ....i.e7). and this is the most natural play.

9.te2lilc6

Black Avoids the Main Line: ... c5 Systems The simplest. 9 ...dxc4 immediately is also possible, as after 10 "'c2 tOdS, 11 O-O-O"'aS 12 ~xdS exdS 13 e4 c3! recoils badly on White, while 11 .txc4 ~xf4 12 exf4 "fIc7 13 .i.d3 h6 =, De Roode-Hrsec, AmsterdamlAmhem 1983, is unthreatening. 10 0-0 dxc4 11 .i.xc4 ~h5 = Kne!evic-Karpov, Leningrad 1977. Black has used a simple equalizing plan from Chapter 4.3, Line D.

163

White stands well. See B2, note '2' to Black's 9th, for the continuation.

B2) 8•..dxc4?! 9 .i.xc4 (D)

B

B) After 8 "fIc2 (D):

The normal move is 8 ... ~6, but there are several alternatives: 81: 8 .••~bd7 163 82: 8••.dxc4?! 163 83: 8......85 164 B1)

8...lbbd7 Unthematic; the knight generally wants to come to c6. 9:dl 9 cxdS must also be good for a secure edge; White will not have to worry about lines where Black forces ...d4. 9..•dxc410 .i.xc4 Mancebo-Cabrara. Alicante 1977.

This is even less appealing for Black than after 8 a3, as White's extra tempo is a developing move rather than a prophylactic pawn-push. 9•.•86 9 ...~bd7 and now: 1) After 100-0?! lQb511:tad1 (11 .i.gS!? ECO) ll...llhl412exf4 "fIc713 ~gS ~f6 14 ~e4 ~xe4 IS 1Wxe4 g6 16 fS?! (too adventurous) 16....i.e7! 17 ~xe6 fxe6 18 .i.xe6+ Wg7 Black had slig1ltJy the better of the draw in Larsen-tiric, Beverwijk 1964. 2) The best answer is probably 10 :dl, discouraging ...~hS. From this position, Mancebo-Cabrara, Alicante 1997 continued 10... a6 11 0-0 bS 12 .i.d3.tb7 13lbeS ± h6 14 ~ "'b6 IS ~xd7 ~xd7 16 ~xcS ~xc5 17 .i.h7+ ~h8 18 :d6, leaving White effectively a tempo ahead of the already promising line given under A52 above. 10.td3 To meet 1O... b5 with 11 ~ ±. 10 0-0 b5 11 .i.d3 .tb7 is less well timed.

164

The Queen's Gambit Declined: 5 j,{4!

10•••h6 10... iLlbd7 110-0 b5 12lDe4 Wb6 13 :acl ~e7 14 ~c7 Wa7 15 ~d6 ± Petrosian-R.Byme, Moscow 1975, is horrible. 11 O-O?! Again, fractionally early. 11 :dl!? iLlbd7 120-0 ±. 11•••iLlc6 Ensuring that White has to spend a tempo on a3. 12 a3 iLlh5 13 b4 iLlxf4 14 ~h7+ 'it'h8 15 bxc5 iLlxg2! 16 'it'xg2 g6 17 .ixg6 fxg6 ao Tabatadze-Michenka, Ostrava 1992. Entertaining, but White could have kept better control.

83) 8.....aS (D)

This is a respectable try. The main line is similar to S... iLlc6 9 a3 Wa5 10 iLld2, but without White having to spend a move on a3. 9lLld2 1) 9 a3 may obviously be met by 9 ... lLlc6 10 0-0-0, while 9 ....ie7 and 9 ... dxc4 10 .ixe4 a6 are fully worth considering. 9 ... a6 10 iLld2 .ie7 11 iLlb3 Wb6 120-0-0 Wa7 13 e4 dxc4 14

.ixc4iLlbd7 15 e5 WbS 16.ig3iLlg4 17 :hel g61SiLld4 Wc719.ia2~hS is too elaborate to be convincing for Black. 20 h3 iLlh6 21 h4 might be an improvement on 20 We2iLlh6 21 .if4 iLlgS ao Pelletier-Dubiel, Altensteig 1994, when White's queenside is a little loose. 2) 9 .ie2 and now: 2a) The simplest reply is 9 ... iLle4, and if 10 iLld2?!, then 1O... iLlxd2 (1O...iLlxe3?! lliLlb3) 11 Wxd2 dxc4 12 .ixc4 =. 2b) Another try is 9 ... iLlc6 10 0-0 dxc4 11 .ixc4 .ie7 and now 12 :fdl?! .id7 13 a3 :fcS! 14 b4 WdS was fine for Black in Greenfeld-Kholmov, Pardubice 1995. However, Kholmov admits that 12 a3! .id7 13 b4 Wh5 14 .ie2 would have been good for White. 9•••.ib4 9 ....ie7!? seems more secure, although after 10 .ie2, 10...lLlc6 11 0-0 e5 Ih-Ih Bellon-Martin Gonzalez, Alicante 19S9, is not a rigorous test. Indeed, in the final position White may be regarded as ahead of the Old Main Line, in that he has substituted .ie2 and 0-0 for a3 and :dl, thus sorting out the problem of his kingside development before the centre opens up. 10....id7 is more flexible for Black, 11 cxd5?! iLlxd5 12 iLlxd5 Wxd5 13 .if3 Wb5 giving White nothing. 10.id3 10 .ie2?! .ixe3 11 bxe3 iLlc6 12 0-0 e5 13 .ig5lLle4! =F O.Foisor-Ubilava, Tbilisi 1986. 10•••lLlc6 11 0-0 d4 1) 1l ...e5 12 .ig3 d4, and now 13 a3?! .ixc3 14iLlb3 Wa415 bxc3 dxe3 =F G.Sachs-Emeste, Biel 1991, was

Black Avoids the Main Line: ... c5 Systems over-elaborate. Simply 13 ~b3 "d8 14 exd4 is strong. 2) 11.. ..i.xc3 12 bxc3 e5 13 .i.g3 .te6 14 .!:tabl ;!; Sanna-Hawes, Novi Sad OL 1990. 12 ~b3 .d8 13 exd4 ~xd4 14 ~xd4 .xd4 15 lLle2 "d8 16 .!:tadl 17 a3 .i.e718lDc3 White has an imposing lead in devclopment, Shneider-Kakhiani, BudaJlcst 1990, though, as noted, Black could have improved earlier.

"'as

5.3 Black's alternatives on move 7 1 d4 d5 2 c4 e6 3 M ~f6 4 ~f3 .i.e7 5 .tf4 0-0 6 e3 c5 7 dxc5 (D)

B

165

~xd5 exd5 10 .i.e5 ~xc5 11 .te2 .i.f6 12 .txf6 Wxf613 Wd4 Wxd4 14 fud4 ;!; Gavrikov-Begovac, Swiss Cht 1993. This is certainly not a winning attempt for Black. 7...dxc4?! is dubious, if for no other reason than that it leaves White a tempo ahead of 7 ....i.xc5 8 .te2 dxc4 9 .txc4. For example, 8 .txc4 .i.xc5 9 'ifc2 a6 10 ':dl (1O ... ~bd7 11 .i.d3 b5 12 ~e5 Wb6 13 .i.xh7+ ~xh7 14 ':xd7 ± Miles-Barua, British Ch (Edinburgh) 1985) 11 a3.te7 12 0-0 b5 13 .ta2 .i.b7 14.i.bl ':c8 15 ~g5 g6 16 Wb3 'ifb617 e4 ~bd7 18 e5 ~d5 19 ~xd5 .i.xd5 20 .g3 ~f8 21 h4 ±V.Bukaljr-Jukic, Zagreb 1995. 7....aS!? This is a more interesting try, still relatively under-explored. White has a wide choice: A: 8.i.d6?! 165 B: 8 ':c1!? 166 C: 8.c2 166 D:8~ 167

.a5

A) 8 .i.d6?! (D)

Back in the early 1970s, 7 ... ~6 was regarded as the main line, Black not being frightened of the isolated pawn, or at least willing to accept the slight disadvantage it entailed, after 8 cxd5 exd5 9 .te2 .i.xc5. We have already considered this variation under the move-order 7....i.xc5 8 cxd5 exd5 9 .i.e2 ~6 (Chapter 4.4, Line B2). Another way of accepting the isolated pawn is 7 ... ~6 8 cxd5 ~xd5 9

This is premature. White will not be able to keep the extra pawn.

166

The Queen's Gambit Declined: 5 JJ..f4!

S••..i.xd6 Heading to what western players often derogatorily refer to as the 'Russian draw' . If Black is more ambitious. he may try 8...1Dc6 9 ~d4 .i.xd6 10 cxd6 Wb4 11 ~b5 dxc4 12 'ii'a4 a6 13 Wxb4 ~xb4 141Dc7 l:lbS 150-0-0b516a3 ~c617 .i.e2l:lb618 f4 l:ldS 19 e4 e5 with sharp and unclear play, Csaba-Kapu, Hungary 1993. 9 cxd6l:ld8 10 Wd2 dxc4 11 .i.xc4 l/2-1fz Mikhalchishin-Azmaiparashvili, Dortmund 1992. 11. ..Wc5 =. B)

Sl:lcl!? This looks impressive from the two examples given in ECO: 8...dxc4 9 .i.xe4 Wxc5 10 We2 ~c6 11 0-0 :dS 12 e4 ± Govbinder-M.]ohanessen, corr. 1972; or S....i.xc5 9 ~2 .i.e7 10 .i.e2 ~c6 11 0-0 .d8 12 cxd5 ~xd5 13 ~xd5 exdS 14 ~b3 ± Larsen-Lombardy, Monte Carlo 1967. However, these are old examples, and one should always be a little suspicious of lines which have not been tested recently. S•..ll)c6!? 9 a3lOe4 10 Wa4 Black has no reason to be scared of 10 cxd5 exdS 11 b4 Wxa3 12 ~xd5 .i.e6. 10•••Wxa4! 11 ~xa4.i.f6 Black has excellent compensation for the pawn; the knight is very uncomfortably placed on a4. If, for example, 12 ~d2, then 12... ~xd2 13 'i>xd2 ~a5! with ....i.d7 to follow after White has averted the knight fork on b3.

C) SWc2(D)

8

Sticking closely to the themes of the main line. S...~ 1) S...dxc4 9 .i.xc4 Wxc5 and now 10 .i.b3 ~c6 is ebl.i.eS 13 ~2 .i.f8?! Black feels obliged to provide a retreat for the queen, and loses the thread of the game. There is a promising exchange sacrifice with 13 ...dxc4

Black Avoids the Main Line: ... c5 Systems 14 lbxc4 bS! 15 o!M6 (15 b4lbxb4) IS ... b4 16lbxc8 l:txc8 and Black's atlack is a few tempi ahead of White's, always useful when the kings are caslIed on opposite flanks. 14 gS 'DhS 15 adS lbxf4 16 exf4 lbd417 jM3lbfS 18 .i.h3 ± GoldinGeller, Sochi 1989.

167

always needs watching. The prophylactic 10 .tg3! ;I; is best. 10•••eS 11 b4 "d6 12 .tg3 l:td8 13 lbo d4 14 exd4 lbxd4 15 lbxd4 "xd4 Gligoric-Geller, YugoslaviaUSSR 1967.

=

02) 8.••dxc4 (D)

0) 8lbd2 (D)

H

Strongly discouraging the simple recapture with the bishop. We consider: Dl: 8 .....xCS 167 D2: 8 ••• dxc4 167

01) 8 .....xCS9:c1 1) 9 a3 dxc4 10 .i.xc4 l:td8 11 :cl lbc6 12 b4 "fS = Benko-Ivkov, TelAviv 1964. 2) 9 lbb3 "b6 10 cxdS lbxdS 11 lbxdS exdS 12 .td3 .tb4+ 13iOd2 (13 c;t>e2!?; 13 c;t>n I?) 13 ...10:6 14 0-0 .i.e6 KlingelhOfer-A.Nikitin, Dortmund 1993. 9 •..lbc6 10 a3?! White's pawn-storm on the queenside is rarely dangerous in the .i.f4 system, while a tempo-gain with ...eS

=

9.i.d6 In another dodgy ECO verdict, this is given as ±. 1) 9 lbxc4 "xeS 10 l:te1 l:td8 11 "e2 lbc6 12 a3 "fS 13 .te2 (13 "xfS!?) l3 .....xe2 14 :xe2 .i.d7 Plachetka-Prandstetter, Czechoslovak Ch 1978. 2) 9 .i.xc4!? :d8 10 "e2 "xeS 11 lbb3 "fS 12 .i.e7 :d7 13 .tg31O:6 14 :dl (14 0-0 lbbS =; 14 h3! ;1;) 14...:xd1+ 15 "xd1 .td7, when, unfortunately for the author's Elo rating, there was not enough to squeeze in Croueh-N.Moloney, Mill Hill 1997 . Even so, with the move 14 improvement, this might be White's best try. 9 •••.i.xd6 10 cxd6 Black has no time for 10... b5?? in view of 11 .f3. l1lbxc4?!

=

.cS

The Queen's Gambit Declined: 5 j.f4!

168

It is tempting to try to keep hold of the extra pawn, but the effect is to retard White's development. 11 .i.xc4 is safe, if deadly dull. 1l••• bS! 11 ...~c6 12 Wb3 :b8 13 a4 b6 14 WbS .i.a6 15 Wxa6 ~b4 16 'it'bS ~2+ 17 ~d2 ~xa118 .te2 WxbS 19 axbS ~b3+ 20 ~c2 ~S 21 b4 ~b7 22 e4 is clearly better for White, Chernenko-Rovner, corr. 1970.

w

12~aS?!

But 12 ~d2 b4 13 ~e4 ~xe4 14 ~xe4 WeS still gives Black a dangerous lead in development. 12••• b4! Improving on the horrible 12 ... a6? 13 ~b3 Wb6 14 Wd4 Wxd4 15 ~xd4 :d8 160-0-0 :xd6 17 .i.e2 ~bd7 18 .i.f3, when Black has fallen a long way behind in development in regaining his pawn, Knezevic-M.Kapelan, Vriac 1977. 13 ~b3 14liJa4 14 Wf3? bxc3 15 Wxa8 cxb2 wrecks White's position. 14•••.i.d7 The obvious continuation 14...:d8 ISWd4! Wxd616'ifxd6lhd617:cl leaves Black with a few weaknesses along the c-file.

"eS

IS~bcS.tc6

But not IS ... ~4? 16 f4!. Black should be at least equal, and possibly a little better, given the tangle of the white knights. Offering the queen exchange with 16 Wd4 is White's most prudent choice.

5.4 The 5 ... c5 system 1 d4 dS 2 c4 e6 31Oc3lM6 4 ~f3 .te7 5 .i.f4 cS (D)

Black may play ... cS before castling, but generally chooses not to. White's natural response is 6 dxcS, when play could transpose to the main lines if both sides are willing, but deviations are possible. 6 ...~6 could be tried, for example, with the idea of bringing pressure to bear on e4, but various games from the early 1970s suggest that what usually happens is that White exchanges on dS and gains comfortable play against the black isolani. 6dxcS 6 e3?!justifies Black's move-order. 6 ...cxd4 7 exd4 dxc4 (simplest; 7...0-0 8 cS ~e4 9 .td3 fS 10 h4 .i.d7 GO Horak-Efler, Czech Republic 1996) 8 .txc4 0-0 9 0-0 gives rise to a thematic IQP position, with White for once having the isolani. Play continued 9 ...~bd7 10 :el ~b6 II .tb3 ~bdS?! (l1...~fdS leaves Black less tucked up on the kingside) 12 .i.gS b6?! 13 ~dS exdS (13 ...~dS?? loses a piece to 14 .i.xdS) 14 ~eS ;t in E.Meduna-Brameyer, Bad Worishofen 1995. Now we have: 169 A: 6•••.txCS B: 6...liJa6 169

Black Avoids the Main line: ... c5 Systems

A) 6•••.i.xc5 This would quickly lose independent significance if after 7 e3 Black were to castle immediately or to play 7...1Oc6 and then castle. Here we consider the deviations.

169

and no worse than 5... 0-0 6 e3 c5, though from force of habit Black usually castles first. B) 6...~a6(D)

7e3

7 cxdS ~xd5! 8 ~xdS exdS is unpromising. Black is threatening ...'iWb6, lind if 9 e3, then 9....i.b4+ =. 9 'ii'c2 'iVb6 10 O-O-O?! ~c6 would be taking great risks with the king.

w

7...~c6 Here there are three examples: 1) 8cxdS~dS9~dSexdSlOa3.

Now 10...0-0 transposes to the Knight Exchange Variation. In ChekhovAnikaev, Moscow 1992, Black varied with 1O..•.i.e7, maybe hoping to establish the bishop on f6 before castling. Play continued 11 ~5!? 0-0 12 .i.d3 ~xe5 13 .i.xe5 .i.f6 14 .i.xf6 'ii'xf6 15 'iVc2 g6 160-0 d4 17 e4 with perhaps a slight edge to White. 2) 8 .i.e2 .i.b4!? (8 ...0-0) 9 0-0 (9 1i'c2!? ~ 10 0-0 ;t) 9 ... .i.xc3 10 bxc3 0-0, Hertneck-Beliavsky, Munich 1994 takes the game out of established theory. Unclear play resulted after 11 ~d2 (11 'ii'c2! Hertneck) 11.....e7 (1l...dxc4! 12 .i.xc4 ~S 13 .i.xd5 'ifxdS 14 c4 Hertneck) 12.i.gS Itd8 13 'ifc2 h6 14 .i.h4 ~5!? 15 Itfd 1 ~g6 16 .i.g3 b6. 3) After 8 a3, 8 ... 0-0 goes straight back to the main lines, while 8 ... a6?! gains nothing for Black: 9 'ii'c2!? dxc4 (9 ... 0-0 10 Itdl ;t) 10 Itdl 'ii'a5 11 .i.xc4 .i.e7 12 0-0 0-0 13 ~e4 ;t Whiteley-McDonald, London 1994. There is nothing particularly startling here. S...cS is probably no better

=

This form of development is familiar from the Nimzo-Indian (1 d4 ~f6 2 c4 e6 31Oc3 .i.b4 4 'ii'c2 c5 5 dxcS 0-0 6 ~f3 ~6), but Black is worse off here through his bishop being passively placed on e7, rather than on b4. This takes all the bite out of the ... ~a6 idea, leaving White free to develop without interruption. He may continue: Bl: 7 e3 169 B2: 7 .i.d6!? 170

B1)

7 e3 ~c5 8 adS 8 .i.e2 0-0 (8 ...dxc4!?) 9 cxdS ~dS 10 ~dS exd5 11 .i.e5 ;t D.GurevichBegovac, Bern 1995.

8...~dS 8 ...exdS and then: 1) 9 .i.e2 0-0 10 0-0 .i.e6 11 .i.eS Itc8 12 Itcl a6 13 h3 bS with unclear play, Petrosian-Spassky, Moscow Wch (5) 1969. The normally alert Petrosian now fell for 14 .i.d3?? d4! IS .i.xd4

170

The Queen's Gambit Declined: 5 1./4!

lilxd3 16 "xd3 .i.c4 winning the exchange. ECO gives 14lild4 and then 14....i.d6 IS .i.xd6 "xd6 16.i.f3 ;t, but this is far from conclusive. Black could, for example, consider 14.....d7, with ... b4 in mind. 2) However, there is a simpler and better plan for White: 9 .i.bS+! .i.d7 10 .i.xd7+ "xd7 11 0-0 0-0 12 .i.eS l:fd8 13 "e2 a6 14 l:fdl ;t and White has effortlessly achieved pressure against the isolani, Gheorghiu-Rubinetti, Mar del Plata 1971. See the Introduction for the complete game. 9 lbxdS exdS 10 .i.bS+ .i.d7 11 .i.xd7+ "xd7 12 0-0 ECO tersely gives this as ;t, quoting Darga-C.Bielicki, Havana 1965, but on the database I am using, Black has 212 from this position! Apparently, a more extensive database gives three extra draws, but White's minus score still stands. 12•••0-0 13 .i.e5 :'c8 13 ...f6 14 .i.d4lile4 IS "d3 b6 16 l:fdl l:fd8 with equality, DanielianNikolac, Balatonbereny ECC 1993. 14 "d4?! Depriving the bishop of an important retreat square. 14 l:cl is about equal. 14•••f6 IS .i.g3lUd8 16lUdl 'ii'bs 17 :ac1 lile4 ; Orlowski-Begovac, Switzerland 1994. Chess technique is still evolving! A generation ago, players felt happier protecting the isolated pawn with the knight, but now there is more sympathy for an approach involving exchanging Black's passive f6-knight for White's active c3-knight. This plan gives Black's other pieces more freedom.

82) 7 .i.d6!? (D)

B

White too can pick a pair of minor pieces to exchange. Black's 'good bishop' is the target. 7•••lbxcs 1) 7 ...dxc4 8 "d4 0-0 9 e4lLld7 10 .i.xc4 (10 eSt? Portisch) 1O...lilaxc5 II l:dl .i.xd6 12 "xd6 "b6 13 .i.bS a6 14 "xb6lbxb6 IS .i.e2 ;!; PortischIvkov, Prague 1970. 2) 7 ...0-0 8 cxdS exdS 9 e3 (9 .i.xe7 "xe7 10 e3lilxcS transposes to the next note) 9....i.xd6 10 cxd6 "xd6 II .i.e2 "b6 12 "b3 "xb3 13 axb3 lilb4 14 0-0 .i.d7 IS :fdl :fc8 16 lilel Wf8 17 .i.f3 ;!; Tukmakov-Savon, Lvov 1978. 8 .i.xe7 ~e7! The text is startling, but it avoids the tedious positions reached after 8.....xe7. for example 9 cxdS exdS 10 e3 0-0 11 .i.e2lLlce4 12 "d4lilxc3 13 "xc3 .i.g4 140-0 :ac8 15 11'b3 :c7 16lLld4.i.xe2 17lilxe2 "eS 18 l:adl ;t Portisch-Bobotsov, Amsterdam 1971. It should be noted that the game cited in the text actually went 7 cxdS exdS 8 .i.d6lilxcs 9 .i.xe7 "xe7, but there seems to be no reason that the

Black Avoids the Main Line: ... c5 Systems king capture can't be tried against the 7 i.d6 move-order. Black's king looks temporarily exposed, but it can soon he secured by ...:e8, ... ~f8, etc. 9cxdS 9 e3 dxc4

=.

171

9•.•exdS 10 e3 .te6 11 .te2 'ii'b6 12 Wd4 :hc8 13 0·0 ltlce4 14lt:la4 Wxd4 15 ltlxd4 .td7 16 .tbS .txbS 17 ltlxbS ltlcs 18 ltlxcS :xcS 19 ltld4 I/z·112 L.Popov-Spassky, Dortmund 1973.

=

6 Systems with ... b6 1 d4 d5 2 c4 e6 3lLlc3lLlf6 4lLlr3 i.e' 5 i.r40-0 For S...b6, see Line F, at the end of the chapter. 6e3b6(D)

In effect, we reach Part 2 of the book. White's basic plan in the i.f4 system is simple and unpretentious development, and the avoidance of unnecessary exchanges of pieces, with his pawn structure providing just that little extra room to manoeuvre to give chances of a lasting edge. We have so far considered what happens if Black challenges White's initiative with an early ... cS; now we examine what happens if Black too is content to play simply and unpretentiously. Few would contend that White's advantage is anywhere near decisive if Black plays quietly, but it would also be difficult to argue that Black, a move behind and with less space, can claim easy equality without making a serious

assault on the logic of White's plan of development. Thus with both the ...b6 system, and the ... c6 and related systems discussed later, we are in the territory of a slight but unquestioned advantage to White. The important theoretical and practical question is whether this is the sort of slight advantage that is likely to be the prelude to a bigger advantage, or whether Black is gradually edging his way towards a playable middlegame with chances for both sides. In the ...b6 system, White's main hope of making progress is either to play for a Pillsbury Bind with cxdS, lLle5, i.d3, Wf3, etc., or to play against the hanging pawns if Black plays a later ...c5. In terms of particular moves from the diagram. position, the main choice is between 7 cxd5, 7 i.d3 and 7 l:[cl, with 7 i.e2 being solid and unambitious. 7 Wc2 has had a big cloud over it, but may still be OK if White takes care; we look at this shortly. Also, a game with 7 lLlbS? i.b4+ 8 lLlc3 :j: has crept on to the database, but since this was in an under-l0 tournament, it would be unfair to cite the exact reference. Our sections are: A:' Wc2 173 B:' i.e2 174 C: 'adS 175 D:' i.d3 187 E:' l:[c1 190 F: 5••• b6 (instead of 5•••0-0) 194

Systems with ... b6 8cxdS~xd5!

A)

7"c2(D)

This routine developing move is not as noncommittal as it looks. Black can now revert to the aggressive plan of attacking White's centre with ...cS, with the bonus of having the bishop occupying the long diagonal. 7 •••.i.b7 I) For 7 ... c6, see 6 ... c6 7 "c2 b6 (Chapter 7.2, Line E). 2) 7...cS?! 8 dxcS bxcS 9 .i.e2 offers White the pleasant prospect of play against the hanging pawns. M.Gurevich-Dolmatov, Marseilles 1988 continued 9 ....i.b7 (9 ...d4 10 exd4 cxd4 11 l:dl ~c6 120-0;t Gurevich) 10 0-0 ll:lbd7 11 l:fdl l:c8 12 l:d2 "b6 13 cxdS exdS?! (l3 ...~xdS 14 .i.gS ;t Gurevich) 14 ~eS :Cd8 IS .i.f3 ~xeS 16 .i.xeS d4 (16 .....e6 17 .i.xf6 .i.xf6 18 l:ad 1 .i.xc3 19 bxc3 ± Gurevich) 17 .i.xb7 dxc3 18 l:xd8+ l:xd8 19 .i.f3 cxb2 20 .i.xb2 and White had the better pawn structure to work with, along with his bishop-pair. 31. 7 ....i.a6 8 ~S dxc4 9 .i.xc4 .i.xc4 10 ~xc4 ~dS 11 ~xdS "xdS 120-0cS 13dxcS"xcS 14l:acl ~6 Khalifman-Pigusov, Sochi 1989.

=

173

The one example of 8...exdS on the database is ancient. 9 .i.d3 and rather than 9 ...~bd7 10 0-0 cS 11 dxcS bxcS 12 l:adl h6 13 .i.fS, when Black's hanging pawns soon fell apart in Blackbume-Taubenhaus, New York 1889, 9 ...~6!? 10 a3 cS 11 dxcS ~cS is worth considering. White could also win a pawn by 9 ~bS?!, but Black has compensatory piece activity after 9 ... .i.h4+ 10 ~dl ~6 11 .i.xc7 "e7, etc. Maybe 9 .i.e2!? 9 ~xdS "xdS 10 .i.d3 10 "xc?? is extremely risky due to 1O....i.h4+ followed by ...l:c8. 10 a3 "a5+ 11 ~2 =. 10•••c5 The check isn't worth it 1O... .i.b4+?! 11 ~e2 .i.a6 12 l:hcl .i.xd3+ 13 "xd3.i.d6 14 .i.xd6 "xd6 IS ~gs;t Bewersdorff-Rybak, Prague 1989. 11 dxc5 11 .i.xh7+ ~h8 12 .i.d3 cxd4 13 exd4 ~6 14.i.e4 ~d4 ;. 11•••h6! This pawn sacrifice, if accepted, allows Black to take the initiative. 12cxb6?! But White shouldn't accept, although one would have to see over ten moves ahead in a line where White wins a pawn and exchanges queens to see why this should not be played. White still has chances for an edge if he ignores the pawn and just develops, for example 12 J:[dl "xcS 13 0-0!1. 12.....a5+ 13 ~e2 13 ~2ll:la6 +. 13•••l:c8 14 "b3 ~6 15 l:hc1 axb6 16 'lfb5 If White doesn't exchange queens, ...~S is going to be a big problem.

174

The Queen's Gambit Declined: 5 ~f4!

16••••xb5 17 J.xb5 J.f6! 18 ~e5 :c5! 19 :Xc5 ~c5 (D)

':dl cxb3 + was Naumkin-Pushkov, Azov 1993. Astonishingly, Pushkov got a second chance to find the correct move!

25 J.xe5 J.xe5 26 .:et

w

26 ~c2 J.xg2 27 f4 gxf4 28 exf4 J.e4+ 29 J.d3 J.d4 30 J.xe4 ~xe4 31 ':dl (31 ~d3 ~c5+) 31...J.c5-+ Oragomaretsky-Koniushkov, Sochi 1996.

26•••J.e4+ 27 ~d2 ~bl+ 28 ~e2 29 J.d3 J.xg2 30 f4 gxf4 31 exf4 J.d4 32 ':c7 J.d5 33 ':a7 J.c5

~xa3

0-IIzkuznykh-Pushkov,~1

20a3?! White made this identical mistake on all three games on the database! We are in the middle of one of the deepest and most subtle opening traps there is. White has won a pawn, and naturally wants to keep a tight pawn structure to make use of his extra pawn in the endgame. It is only with the benefit of hindsight that one can say that 20 a4! is the correct way to play. Naturally the pawn cannot be held in the long run, but in the time that Black takes to win the pawn back, White can get all his pieces into reasonably active play, with a draw being the likely result. The most interesting traps are those which only a strong player would fall into.

20... g5 21 J.g3 ':a5 22 J.c4? It is still not too late for 22 a4.

22••• ~4!! The trap is sprung. Suddenly the black pieces are totally dominant, and White's position is about to collapse.

23 b3 ~+ 24 ~d3 ':xe5 The cleanest kill. 24 ... b5?! 25 ~xc3 bxc4 26 ~b4 ':xe5 27 J.xe5 J.xe5 28

1995.

B)

7 J.e2 (D)

B

Ultra-solid, if slightly unthematic.

7•••i.b7 For 7 ...dxc4 8 J.xc4, see Line 02 (7 i.d3 dxc4). 80-0 As with so many games from the Soviet and ex-Soviet school of chess, one wonders whether A.Nikitin-Baikov, Erevan 1983 took longer than five minutes: 8 cxd5?! (drab) 8 ... ~xd5 ~ ~d5 J.xd5 10 0-0 ~7 II ~5 ~xe:), 12 J.xe5 i.d6 13 J.xd6 cxd6 14 J.f3' Ih-1h.

Systems with ... b6 For 8 :c 1, see Line E2, note to White's 8th (7 :el i..b7 8 i..e2); the normal move with the 7 :c 1 moveorder would be 8 cxd5. 8...li:)bd7 8 ... c5 (T.Georgadze) is a touch premature. With White's development close to completion, it is wiser for Black to get some more pieces out before opening the centre. White can now try 9 dxc5 i..xc5 (9 ... bxc5 10 cxd5 exd5 ;t gives Black the problem of the hanging pawns) 10 cxd5li:)xd5 II li:)xd5 "xd5 12 "c2 with a slight advantage. Symmetry does not guarantee equality. 9h3 White keeps the tension. Instead 9 li:)e5 li:)e4 1O:C 1 li:)xc3 11 :xc3 ll:)f6 12 :el dxc4 13 :xc4 c5 was equal in Tomczak-H-l.Schulz, 2nd Bundesliga 1983/4. 9 .•.0 10 i..h2 a6?1 The pawn becomes a potential target here. Black probably has several routes to a reasonable game, for example 1O... dxc4 11 i..xc4 cxd4 12li:)xd4 (12 exd4!?) 12...:c8. lla4:c8 Again, pawn exchanges could be considered. 12 adS lOxd5 Now Karpov-T.Georgadze, Moscow 1979 continued 13 "bllD7f614:dl cxd4. Karpov in ECO suggests 13 li:)xd5 exd5 (13 ... i..xd5 is impossible in view of the weakness on a6) 14 b3 ;to I was at first sceptical of this judgement in view of 14...c4 creating a passed pawn, but if White avoids the immediate exchange and plays, for example, 15 Wc2, it is difficult to see what Black can do with the pawn.

17S

C) 7cxdS We now move on to the heavy theory. White will generally want to exchange pawns here sooner or later, in order to clarify the position in the centre, and is more likely to be able to force a pawn recapture if the bishop is not yet on b7. That having been said, Black is still able here to recapture either with pawn or knight: Cl: 7 •••li:)xdS 175 C2: 7 •••exdS 178 C1} 7... li:)xd5 (D)

SlOxdS After 8 i..g3, the critical reply is 8...i..b7 9 i..d3 c5. Instead it is diffIcult to see why Black should voluntarily strengthen White's pawn centre by 8 ...li:)xc3?! 9 bxc3 i..b7 10 i..d3 c5 11 O-O;t Fiedler-Pallova, Bradee Kralove 1992. S...exd5 8 .....xdS has also been tried: 1) No one seems to have risked taking the gambit pawn with White, although there is no immediate refutation of9 i..xc7!?, which not only wins

The Queen's Gambit Declined: 5 1./41

176

a pawn, but also cuts out the ...cS break, thus helping keep the centre secure. The onus must be on Black to justify the sacrifice. 2) Instead, the insipid 9 .td3 has generally been chosen, with games mostly tending towards early eqUality. 2a) 9....ta6 10 0-0 cS 11 .txa6 lbxa6 12 "e2 'iVb7 13 :fdl :fdS 14 e4 cxd4 IS lbxd4 lbcs 160 a6 17 :ac1 lIac8 D.zilbershtein-Pigusov, Irkutsk 19S3. 2b) 9.....aS+ 10lbd2.ta611.txa6 "xa6 12 'ii'e2 c5 13 .txbS 'iVxe2+ 14 r,i;>xe2 :axb8 = Yakovich-Gelfand, Minsk 19S6. 2c) 9 ....tb4+?! 10 r,i;>e2 1.d6 11 e4 ~ 'ii'aS 12 .td2 .h5 (12 ....tb4 13 a3 .txd2 14 .xd2 .a4 15 :hcl ;t G.SchrolI-Bradner, Austrian League 1989/90) 13 1Ic1 eS 14 dxe5 .txeS 15 h3.ta6! 16 1.xa6lbxa6 17 'iVa4 (17 g4? is pointless, as 17...'iVg6 threatens ...•xe4+) 17....txb21S 'iVxa6.txc119 :xcl cS 20r,i;>n .g6, Tisdall-Kveinys, Oslo 1992, and now, instead of21 :c4 :feS 22 eS 00, the immediate 21 eS!? possibly slightly favours White. 2d) 9...cS?! 10 .c2 .tb7 11 0-0 (Oreev's notes in Informator 57 suggest that 11 .txh7+! r,i;>hS 12 .td3 cxd4 13 0-0 is a more accurate moveorder, as now Black can confuse the issue with 11...lba6!) l1...cxd4 12 .txh7+ r,i;>hS 13 1.d3 dxe3 14 fxe3 lba6 15 :adl lbb4 16 .c7 .tf6 17 1.bl ± 'ii'c6 18 'iVxc6 lbxc6 19 :d7 i.a6 20 :fdl :acS 21 i.e4 1.xb2 22 .td6 .te2 23 .txfS 1-0 Dreev-Doroshkevich, Rostov on Don 1993. We leave the lack of votes for 9 i.xc7 as an unsolved mystery. 9.td3

=

Thematic in that White places the bishop on its strongest diagonal, but there is the drawback that Black has an awkward bishop check. Two attempts to bypass this problem: 1) 9 .c2!?.tb4+ 10 lbd2 i.d6 11 .txd6.xd6 12 .td3 h6 130-0 .ta6 14 :acl .txd3 15 .xd3 c6 16 lbo lbd7 17 .a6 lUcS 18 :c2 ;t R.Scherbakov-Landa, lurmala 1989. The classic 'minority attack' pawn structure, in which Black's c-pawn is held back by White's d-pawn, and White can prepare to lay siege to the queenside with b4-bS. Black's position is tedious to defend. 2) 9 .te2 .tb4+ 10 lbd2 i.d6 11 1.g3 .tfS 12:c 1 i.xg3 13 hxg3 .d6 14 .a4 c6 IS 0-0 (IS b4!?;t) IS ... aS! 16 "b3 lbd7 17 i.o lbf6 =Aleksandrov-Landa, lurmala 1991. Now 18 .xb6 :fb8 proved perfectly satisfactory for Black, although later in the game White blundered a piece ... and won! 9•••i.b4+ (D) The usual irritant, although 9...cS is also possible: 1) 10 dxcS bxcS 11 0-0 i.e6 (ll...lbc6 12lbeS lbxe5 13 .txeS .te6 14 .c2 g6 15 b3 ;t Donner-Darga, West Germany 1969) 12 .c2 h6 13 :fd 1 .b6 14 e4 lbc6 15 exdS lbb4 16 .e2lbxdS 17.teS lbb4 18 i.c4 :ad8 was equal in Ruderfer-Butnorius, Moscow 1979. 2) 10 0-0, with divergent experiences: 2a) 1O...lbd7 11 :cl (II dxcS!?;t) l1...c4 12 .tbllbf6 13 lbes .tb7 14 . 0 bS 15 b3 .b6 16 .tgS :fe8 17 .tfS h6 18 .txf6?! (winning material is tempting, but it releases the tension

Systems with ... b6 on the kingside; 18 .i.f4! gives dangerous attacking chances against the under-protected kingside) 18 ...•xf6 19 .i.d7 .xf3 20 gxf3 a6! 21 .i.xe8 :'xe8 22 bxc4 dxc4 GO LechtynskyUbilava, Tren~ianske Teplice 1985. Black's queenside pawns outweighed the exchange in the cited game. This is nur first introduction to the Pillsbury Bind in this chapter; there will be several more. 2b) 10....i.e6 and now 11 dxcS hxcS 12 e41Dc6 13 exdS tWxdS 14.e2 :'008 IS .i.e4lM4 =Blagojevic-Hresc, Kirchheim rpd 1990. However, this was only a quickplay, and improvements for White are probably easy to lind, e.g. 11 tWc2!? h6 12lbeS. 2c) 1O...lbc6 11 lbeS lbxeS 12 .i.xeS .i.d6 13 .i.xd6 .xd6 14 dxcS hxcS IS :'cl (1S tWc2 followed by :'fdllooks more accurate) IS ... fS 16 h3 .i.e6 17 .i.e2 :'ac8 18 .d2 :'fd8 with equality, Rezsek-Letay, Hungarian Cht 1993/4. 3) 10 lbeS!? and now 1O...cxd4 11 cxd4 .i.b4+ 12 ~f1 was suggested by Psakhis in Informator 61 and given as unclear. White should be able to play this for an edge; after all, the king gets in the way less on f1 than on e2. One possible idea is to gain space on the kingside with .f3, g4, h4, ~g2, etc. If instead 10....i.b7, 11 0-Olbd712dxcS bxcS 13 .a4lbxeS 14 .i.xeS;t Ibragimov-Radulov, Iraklion 1993. 10~e2

Sometimes the most obvious moves remain untested; there is nothing on the database on 10 lbd2. If White can't improve on the Lalic-McDonald game given below, this might have to be tried. Then:

177

1) After 1O...cS 11 dxcS, 11...bxcS would leave the black bishop misplaced, but 1 l...lM7!? (Burgess) might be a try, for example 12 c6 lDcs 13 .i.c2 .i.a6! (13 ... lbe4 14 .i.xe4 dxe4 IS a3 ±) 14 a3 .i.xd2+ IS .xd2 :'c8 16 c7 tWf6 and White's king is uncomfortably stuck in the centre, while Black can play to recover the pawn with ...lbe6. Maybe White should settle for play against the IQP with 120-0 lbxcs 13 .i.e2; if then 13...d4, 141Dc4! presents Black with a few problems in completing his development; the bishop is still weak! 2) If 10....i.d6 11 .i.xd6 "xd6 White can if he chooses transpose into the Scherbakov-Landa game above (9 .c2) with 12 .c2 ;to 10•••.i.d6 11 hd6 "xd6 12 "c2 12 :'cl cS (12 ....i.a6?! 13 .i.xa6 lbxa6 14 .a4 bS IS .c2 ± Psakhis) 13 dxcS (after 13 .c2?! lba6! 14 .i.xh7+ ~h8 IS .i.fS lbb4 16 .bl lbxa2!, Psakhis, Black is in control) 13 ... bxcS 14 .a4, and now PsakhisWesterinen, Gausdal 1994 continued with the routine 14....i.a6?! IS :'hdl .i.xd3+ 16 :'xd3 and White was a little better. Psakhis notes that 14...c4 IS .i.bl .i.g4 16 :'hdl! .i.xf3+ 17 gxf3 is

178

The Queen's Gambit Declined: 5 1./41

good for White, but 14 .. JWb6 should be playable for Black.

12...cS! Yet another instance where Black's best policy is to let the h7-pawn go. 12...J.a6 13 :hdl J.xd3+ 14 :xd3 ~a6 15 a3 c5 16 dxc5 ~xc5 17 :c3 as 18 b3 gave White a small advantage in Seirawan-Christiansen, Lone Pine 1981. 13:hc1 Black gets plenty of chances for counterplay after 13 J.xh7+ ~h8 14 J.f5 J.a6+ 15 ~d2 ~6 16l:[hcl c4Lalic. 13•••~6 14 J.xh7+ ~h8 15 J.d3 After 15 J.f5 ~b4 16 .bl J.a6+ 17 ~d2 (17 ~el!?) 17... g618 a3liX:6 19 J.d3 J.xd3 20.xd3 c4 followed by ...~a5 White has queenside problems - Lalic. Again, winning the h7pawn has used up a lot of time. 15•••~b4 16 dxcS 16 .c3 c4! - Lalic. 16..:.f6 It is not advisable to snatch the exchange. 16... ~xc2? 17 cxd6 ~xal 18 ~5! (the prisoner can wait) 18...J.e6 19 f4 d4 20 f5 J.d5 (20...J.xa2 21 :xal J.d5 22 J.b5 ±) 21 e4 :ad8 (21...:fe8? 22 d7 :xe5 23 :c8+ +-; 21...J.xe4? 22 J.xe4 :ae8 23 d7! :xe5 24 :c8 :xe4+ 25 ~f2 +-) 22 d7 ± is analysis by Lalic. Black is temporarily a rook ahead, but his piece coordination is miserable, and White will soon start clawing back material. 17 'ire3 ~xd3 18 'irxd3 18 ~xd3? chances his arm too much; after 18.....g6+ followed by ...•xg2 Black's attack should prevail. 18...bxcS 19 ~el liz_liz B.LalicMcDonald, Hastings 199415.

White prudently brings the king back to safety and offers the draw; 19 .xd5? J.a6+ 20 ~el .xb2 would have been unwise. After the text-move, Lalic gives 19...J.a6 20 .a3 :ab8 21 ':c2 "g6 22 :acl "xg2 23 'irxa6 "xO 24 "e2 =. This isn't quite proof that White has nothing after 7...~xd5, as Black still has to demonstrate clear equality after 10 ~d2. Even so, White's edge is far from overwhelming. e2) 7...exd5 (D)

The more traditional, and indeed the more popular response. However, one cannot help feeling that, as with analogous lines after 6... c5 7 dxc5 J.xc5 8 cxd5, the failure to exchange the defensive f6-knight for the attacking c3-knight favours White. 8J.d3 This is White's natural response, aiming for the Pillsbury Bind with ~5, "0, etc. The alternatives mainly anticipate playing against a hanging, pawn formation. I 1) 8 J.e2 with the following possi1, bilities: .

Systems with ... b6 I a) 8 ....tb7 9 0-0 (9 ~5lbbd7 10 II·() transposes) 9 ... lbbd7 (9 ... c5 10 Ic I lbbd7 11 lbe5 is discussed under 7 Icl) 10 lbe5 a6 11 'iVb3 (11 Icl .id6 12lb
179

.teS lbd7 12lbellbhf6 13 .tg3 Ac8 14lbd3 ± Steinitz-Schallop, Nuremberg 1896. 8 ...cS 9 .te2 is examined under 7 Acl c5 8 cxd5 exd5. Line EI below. 4) The only examples I have with 8 lbeS or 8 h3 follow up with 9 .td3. and are considered under 8 .td3. The position after 8 .td3 is a major branching point. We consider: e21: 8 •••cS 9 0-0 .tb7 (or 8 •••.tb7 9 O-OCS) 179 e22: 8 •••cS without 9 0-0 .tb7 184 e23: 8 •••.tb7 without 9 0-0 cS 186 Many of the games listed in C21 and C23 started via 7 .td3 .tb7 8 cxdS exdS. C21) 8•••cS 9 0-0 9lbeS .tb7 10 'iVf3 will almost invariably transpose. 9•••.tb7 (D)

10lbes The most natural, but the position after 10 h3 has an interesting history. In 1978 Tony Miles tried out a new system against the Queen's Indian, scoring wins against leading grandmasters such as Spassky (twice!).

The Queen's Gambit Declined: 5 i,t4!

180

Larsen, Dzhindzhikhashvili and Timman. The system he tried was 1 d4lDf6 2 c4 e6 3 lDf3 b6 4 .i.f4, which could clearly transpose into a .i.f4 Queen's Gambit if Black plays an early ... dS. The innovation was briefly successful, but then antidotes were found. In particular, after 4 ....i.b7 S e3 .i.e7! White must play 6 h3 (D) in order to avoid the exchange of his bishop after ... 1OhS. Instead 61Oc3ll)bS 7 .i.g3 d6 8 .i.d3 lOd7 9 "c2 (9 O-O?! g6 10 h3 lOxg3 Spassky-Karpov, Montreal 1979) 9 ... g6 10 .i.e4 .i.xe4 11 Wxe4 0-0 12 Wc6 a6 13 a411a7 14 aSlOb8 IS "e4 cS proved very solid for Black in Miles-Andersson, Amsterdam 1978.

+

Mter 6 ...cS, I do not even begin to see how to play for an edge for White, whether after 7 dxcS bxcS, or 7 lOc3 cxd4 S lOxd4 0-0 9 lbdbS ~8. In view of this possibility, I can't really add a chapter recommending 4 .i.f4 as a system against the Queen's Indian. However, Spas sky preferred 6 ... 0-0 7 lOc3 dS 8 cxdS exdS?! (8 ... lOxdS! should equalize; White's h3 is of doubtful relevance) 9 .i.d3 cS (9... lObd7 10 0-0 a6 11 lOeS lIe8 12 Wf3 .i.d6 13 lOg4 ;t Miles-Larsen, Tilburg 1975)

10 0-0, and we have transposed to the line being considered (Le. after 10 h3 in the Queen's Gambit move-order of our main line). Now we have: 1) 1O... lOc6 11 ~S, and now 11...c4 12 .i.c2 a613 g4 bS 14 gSlOe8 IS Wg4 ;!; Miles-Spassky, Montilla 1978; or l1...a612"f3 lIe8 13 lIadl cxd4 14 lOxc6 .i.xc6 IS exd4 ;t MilesSpassky, Buenos Aires OL 1978. Looking at this from a Queen's Gambit point of view, if White can establish an edge by playing a quiet move like 10 h3, his prospects must be very promising if he can play something more aggressive! 2) 1O... a6 11 ~S c4 12 .i.fS bS 13 "f3 g6 14 .i.h6 lieS IS .i.c2? (prematurely releasing the pressure; IS lIadl ;!; Florian) IS ...lObd7! with chances for both sides, Florian-Berney, Hungary 1979. 3) 10... lObd7 11 We2?! (the queen proves to be poorly placed here; Stean suggests 11 IIcl!?) 11...c4 12 .i.c2 a6 13 lIadl bS 14 a3 lIe8 IS lOeS 1Of8 and Black stands well, KorchnoiPetrosian, Velden Ct (S) 1980. Black's kingside is secure and he is ready to roll on the queenside; a knight on f8 neutralizes the Pillsbury Bind. From the Queen's Gambit moveorder, 10 IIcl is also possible, sometimes reached by transposition from 7 IIcl. After 10...c4?! 11 .i.bl a6 12 lOeS bS 13 "f3 b4 14 ~2 lIa7 IS lOg3 as 1610fS ± Sydor-Zinn, Lublin 1977, White had achieved the Pillsbury Bind attacking formation against the minimum of resistance. 1O...lObd7 is wiser, and if 11 dxcS then 11...bxcS / (ll...lOxcS 12 .i.bllOfe4 13 .i.eS IIcS\ 14 Wc2 Wd7 IS b4 lOa6 16 a3 is a'

Systems with ... b6 little better for White. Ara.MinasianKashashvili, Erevan 1996) 12.e2 (12 h3!?) 12...lbhSI3.i.g3lbxg314hxg3 "'a5 15 :tfdl tLlf6 16 tLld2, and now instead of 16...':fe8?! 17lbb3 .c7 18 ~a4 c4 19 .i.c2 .i.f8 20 lbd4 ± Gipslis-Andri~, Belgrade 1968, Black can play 16...c417 .i.c2.b4=. Tbehanging pawns are less problematic to defcnd if White's dark-squared bishop has been exchanged for a knight. After 10 lbeS, Black has tried all three ways of developing the queen's knight, and also 1O... a6: 181 C211: 10•••a6?! C212: 10...ttx6 181 C213: 10...lbbd7 182 C214: 10...lba6!? 183 C211) 10...a6?! 11 "(3:a7 ll...lbbd7 or 11...lbc6 would transpose to lines considered shortly. With the text-move, Black hopes to create a hit more flexibility on the queenside hy keeping the knight at home, but White's advantage on the kingside remains unchallenged. 12':adl 12 g4!?; 12 .i.gS!? - Mikhalchishin. 12...c4 13 .i.c2 bS 14 a3 14 g4!? - Mikhalchishin. 14... lbbd7 15 "h3 .i.a8 16 g4 h6 17f3 It is best not to be provoked too quickly. Mikhalchishin notes that 17 .i.xh6? only draws after 17...gxh6 18 "'xh6':e8. 17...lbh718 .i.xh6 gxh619 "xh6 tLldf6 20 gS lbe4 21 fxe4 .i.xgS 22 1I'h3 ± Mikhalchishin-A.Hoffman, Kecskemet 1991.

181

C212) 10...lbc6 (D)

A natural and popular move, yet it is still not enough to challenge White's kingside preponderance. l1"f3 11 ':cl!? might well be just as strong: 1) 11...cxd4 12 lbxc6 .i.xc6 13 lbe2! ± Kramnik. 2) 11...a6 12 'iFf3 .i.d6 13 ':fd1 cxd4 (13 ....c7 14 lbg4! lbxg4 IS lbxdS lbxh2 16 'iWh3 +- Kramnik) 14 lbxc6 .i.xc61Slbe2! .i.a4 16 b3 .i.bS 17lbxd4 ± Kramnik-Yanvariev, USSR 1991. The discovered attack with lbe2 shows that there is a specific point to ':c 1, and that it is not just a random developing move. 3) 11...lbxeS 12 dxeS (12 .i.xeS ;t Kramnik) 12 ... lbe4 13 "c2 fS and now 14 exf6lbxf6 IS .i.eS h6 GO was N.Popov-Klovans, Moscow 1979. One suspects that White has room for improvement, for example 14 :Cdl I?, and if 14 ... tLlxc3 15 bxc3 g51! 16.i.g3 f41, then 17 .i.xh7+~h818 .g6fxg3 19 .i.g8! wins nicely. bS 14 4) l1...c4 12 .i.bl a6 13 a3 ':a7 15 ':cdl!1 (at first glance, a

.f3

182

The Queen's Gambit Declined: 51.f41

strange loss of tempo, but White wants to discourage the defensive idea of ... iLlxe5 followed by ... 1Lle4; Black's structure is passive, and he can do little to stop White's attack) 15 ...1..d6 16 "h3iLle717iLld7! iLlg6 (17 .....xd718 1..xh7+ <;Ph8 19 1..f5+ +-) 18 1..xd6 winning material, Malaniuk-Arbakov, Warsaw 1992. 11•••cxd4 1) 11...a6?! is too slow. 12 lIacl transposes into Kramnik-Yanvariev above, while 12 dxc5 bxc5 13 lIadl leaves Black with a particularly vulnerable pair of hanging pawns. Black's attempts to wriggle out of trouble in the game Dautov-M.ROder, 2nd Bundesliga 1995/6, were unsuccessful after 13 .. :ii'b6 14iLlxd5iLlxd5 15 "xd5 iLlxe5 16 "xe5 "c6 17 f3 1..f6 18 "d6 1..xb2 19 1..e4 "xd6 20 1..xd6, winning the exchange. 2) 11.....c8 and now after 12 lIadl c4131..bliLlxe5141..xe5(14dxe5d4 00) 14 .....g4 Black succeeded in exchanging queens in M.Kobrin-Kalcheim, Petach Tikva 1997. Possibly this would be the time to follow the Miles-Spassky games in the note to White's IOthinC21 and play 12h3!? 12iLlxc6 1..xc6 13 exd4 An illustration of the point that positions with symmetrical pawn structures are not necessarily drawish; indeed the symmetry often enhances the effects of any superiority in piece mobility that one player might have. White is clearly better. 13.....d7 13 ...1..d6 14 1..g5 1..e7 15 lIfel h6 16 1..h4 "d7 17 lIe5 lIfe8 18 lIael ± Dinstuhl-Hedke, Bundesliga 1996/7. 1411rel

14 a3?! (why?) 14... lIfe8 15 lIfel 1..d6 16 1..xd6 "xd6 17 "g3 is only equal, Lirindzakis-Mikhalevski, Khania 1993. 14...:re8 15 1..eS "g4 15 ... b5 16 1..f5 "b7 17 lIe3 b4 18 1Lle2 g6 19iLlg3 ± Keene-O.Jakobsen, Esbjerg 1981. 16iLlb5! 16 "e3 1..d6 17 h3 "h4 18 iLlb5 1..xb5 19 1..xb5 lIe6 does not give White very much, Ruban-Begun, Podolsk 1990. Black's pressure on the e5-bishop keeps him in the game. 16.....xfJ 17 gxf3 1If8 18 lIael 1..d7 19 a3 1..xbS 20 1..xbS I.FaragoSturua, Erevan 1982. White has a dominating bishop-pair and, notwithstanding the doubled f-pawns, the safer pawn structure. C213) 10•••iLlbd7 (D)

Again this allows White too free a hand. 11 "'3 a6 1211adl 12 1..f5!? iLlxe5 13 dxe5 1Lle8 14 lIadl1Llc7 15 Wh3 g6 16 1..h6 lIe8 17 e6 ± is analysis by Nogueiras and Garcia Gonzalez.

Systems with ... b6 12 l:lfdl l:leS 13 l:lacl l:la7 14 g4 g6 I ~ .i.e2 b5 16 .h3 b4 17 llla4 lllxe5 IK dxe5 lOd7 19 .i.f3 ;t NogueirasSpeelman, Taxco IZ 19S5. 12.••b5

12...cxd4 13 exd4 l:leS 14l:[fellOfS 15 .i.g5 l:lcS 16 .tc2 b5 17 .txf6 .i.xf6 IK .i.b3 ± Mikhalchishin-AzmaiparIIshvili, Tbilisi 1980. 13 a3 c4 14 .tc2lllb6?! 15 g4lOe8 16"h3 g617 .th61Og718 f4 A kingside wipe-out is in progress, Mikhalchishin-S.Luce, Budapest 1990. C214)

10...lDa6!? (D)

The only move to cause White any problems. Black intends to swing his knight round to e6, challenging the hishop rather than the knight. 11"a4 The latest try, switching attention to the queenside. 11 'Wf3 ~7 is the older line: 1) 12 l:lfdllOe6 13 dxc51Oxf4 14 cxf4 .txc5 15 .tc4 dxc4 16 'Wxb7 'ii'c8 17 'iff3 'We6 Islllb5 'iff5 19lOd6 .i.xd6 20 l:lxd6 l:lacS was equal in Vaganian-Beliavsky, USSR Ch (Vilnius) 1980.

183

2) 12 l:ladl lOe6 13 dxc5 lOxc5! (Bagirov suggests that both 13 ... bxc5 14 .tc4 and 13 ....i.xc5 14 .tc4 turn out well for White) 14 .tbl .cS 15 "e2 l:le8 16 l:lfel .td6 BagirovLputian, Erevan 1982. It is surprising that both here and in '1' White waited until the knight had already landed on e6 before capturing on c5. 3) 12 dxc5! .txc5 13 l:ladl may well be the way to play the position for an edge. 4) 121Og4 has also been tried. The correct response is not 12...lOe6?! 13 lIIh3 g6 141Oh6+ ~g7 15 .te5 .td6 16 f4 ± Kolesar-Hirsch, Bmo 1991, but rather 12...lOxg4 13 .xg4lOe6. 1l.....CS 11...1Ob4!? 12 .i.g5 If 12 .i.b5, then Cvetkovic gives 12 ... lllc7!? 13 lllc6 .d7 14 lOxe7+ 'Wxe7 15 .i.c6 llle6! =. In V.Mikhalevski-A.Mikhalevski, Beersheba 1996, Black's attempts to complicate misfired: 12...1Ob4?! (amove too late?) 13 a3 cxd4 14 exd4 .f5 15 .tg3 as 16 .td7.h5 17111b3lOd3 ISlIIxb6 .i.a6 19 .tb5 ±. 12...1Ob4 12...c4?! 13 .tbl l:le8 14 .dl! (an attractive switching of flanks; the Pillsbury Bind is powerful, even if a couple of tempi have gone missing) 14...g6 15"f3 "e616 .txf6! winning a pawn (16 ....txf6 17lllxc4; 16...lIIxf6 17 lOxd5), Avrukh-Florath, Groningen 1995. 13.tel 13 .tbl! looks stronger, with a possible battery along the diagonal leading to h7. The bishop is a little misplaced on e2.

=

184

The Queen's Gambit Declined: 5 j.f41

13•••l:tdS 14 l:tael a6 IS a3 15 Wdl!? IS •••c!0c6 16 "'e6! Since 17 Wxb6?! is answered by 17...ltlxd4, Black was OK in the game Avrukh-A.Mikhalevski, Beersheba 1996. However, the suggested improvements in the above notes run in White's favour.

""3

C22)

clear extra pawn, Keene-Kraidman, Haifa OL 1976, and 10 .txa6ltlxa6 11 'ii'a4 cxd4 12 exd4 ltlc7 13 ttles ± Taleb-S.Mohd, Singapore 1995. 3) 9...ltlc6 10 .tbs!? (10 It!es .tb7 leads back to C212) 1O....tb7 11 Wa4 ltlaS 12 dxcs bxcs 13 l:tfdl a6 14 .te2 ~ Forintos-Liptay, Hungary 1969. C222) 9 dxeS bxeS (D)

S•••eS Here we discuss lines without 9 0-0 .tb7 as follows: C221: 90-0 without 9•••.tb7 184 C222: 9 dxeS 184 C223: 9 l:tel 184 C224: 9ltleS 185 C221) 9 0-0 (D)

This exchange is probably premature. To play it as a pure hanging-pawn position, the white bishop is better placed on e2. 10 0-0 .tb7 11 l:tel ltla6 12 ltle2 "'b6 13 ltlg3 g6 14 h4 l:trd8 IS hS ltle4 Keres-V.Simagin, USSR Ch (Tallinn) 1965. In this example White made no progress against the hanging pawns, so diverted his attacks to the kingside, but there was no joy to be gained there either.

=

1) 9...ltlh5?? 10 .txb8 l:txb8 11 ltles wins the exchange - Keene. If Black wants to play ... ltlhs, he must prepare it with ....tb7 rather than ...cs; see Line C23 below. 2) 9....ta6?! leaves White the pleasant choice between 10 dxcs bxcs 11 l:cl WaS?! 12 .txa6ltlxa6 13 ltlxds ltlxds 14 'ii'xdS l:tad8 15 'ii'c4 with a

C223) 9 l:tel (D) There are obvious transpositional possibilities, after, for example, 9 ....tb7 10 0-0 ltlc6 I1ttles. Here we consider the independent lines.

Systems with ...b6

H

9...J.b7 9 ... a6?! allows White to follow the main lines with 10 lLle5, etc. It also gives White a more opportune moment to initiate play against the hanging pawns; even so, success is not guaranteed. 10 dxcS bxcS 11 0-0 lLlbd7, nnd now 12 lLleS J.b7 13 lLla4 lLlxeS 14 J.xeS lLld7 IS J.g3 Wa5 16 J.fS lLlf6 17 'ifc2 g6 18 J.h3 lLle4 19 J.eS .i.c6 20 lbc3 lLlgS 21 J.g4 ':ad8 = Velikov-Abramovit, Pamporovo 1982; or 12 We2 J.b7 13 ':fdl (13 h3!?) 13 ... lLlh5 14 J.xh7+ ~xh7 IS lLlgS+ .ixgS 16 'ifxhS+ .ih6 17 lLlxdS (17 .ixh6 lLlf6) 17 ....ixd5 18 Wxd5 .ixf4 19 exf4 ':a7 20 ':c3 'iff6 21 ':h3+ ~g8 22 'ifhS 'ifh6 23 'ifxh6 gxb6 24 ':xh6 =Doroshkevich-Vrulin, Kstovo 1994. So maybe 10 lDes is clearest. 100-0 1) 10 dxc5 bxc5 11 0-0 and now instead of 11.. ..id6?! 12 .igS lLlbd7 13 .ifS ;t Teichmann-Marco, Hastings 189S, Black should play ll...lLlbd7 =. 2) 10 lLleS lLlbd7 11 Wf3 lLlxeS 12 dxeS (12 .ixeS!?) 12...lLle8 130-0 lbc7 14 :fdl 'ifc8 Lugovoi-Sirotkina, St Petersburg 1996. 10•••lDa6!?

=

185

In comparison with Line C214 above, White has played ':cl instead of lLleS; the difference is in Black's favour, as White is not so aggressively placed in the centre. 11 "e2lDc7 12 :rdl ttJe6 13 .ig3 c4 14.ifS a6 IS e4 dxe4 16 lLlxe4 lLlxe4 17 .ixe4 .ixe4 18 "xe4 bS 19 a4 .id6 with a slight advantage for Black, M.Nemeth-Vujo§evit, Balatonbereny 1996. White should be able to improve on this, but Black's set-up is sound. C224) 9ttJeS (D)

B

Normally just an inversion of move-order, with 100-0 (or 10"f3 and 11 0-0) coming next. 9 •••.ib7 9 ....ie6?! worked out badly for Black in the game Mikhalchishin-Zumtobe, Bern 1995 after 10 0-0 'iFc8 (ugly, but how else does Black develop the knight?) 11 :CI1Wb7 12"f3 :d8 13 ':fdl ± lLla6? 14 .ixa6 "xa6 Islbc6 .ig4 16 lLlxe7+ ~f8 17 "g3 .ixdl 18 .id6 lLle8 19 lLlf5+ ':xd6 20 lLlxd6 and White wins. 10 "nlDa6 11 g4?!

186

The Queen's Gambit Declined: 5 j.f4!

An unconvincing attempt at a pawnstorm with the king still in the centre. 110-0 is discussed under C214. 1l•••cxd4 12 exd4 li:k7 13 g5?! lCle4! 14 lClxe4 dxe4 15 .txe4 .txe4 16 Wxe4 .txg517 l%dl.txr418 Wxf4 Wd5 =1= Dinstuhl-Florath, Bundesliga 199516. At the very least, White should wait for Black to play ...c4 before carrying out this pawn-storm. e23) 8.•..tb7 (D)

Here we discuss lines without 9 0-0 cS.

90-0 An early lCleS loses much of its sting when Black has not committed his queenside pawns: 9 lCJes lClbd7 10 0-0 (10 'iff3 l%e8 11 0-0 and now not 11.. ..td6?! 12lClxd7 Wxd7 13 l%fcl ;!; Damaso-Alexandre, Lisbon 1997, but 11...a6!? followed by ... lClf8) 1O... a6 11 Wf3 l%e8 12 Wh3lClfS 13 .tgS cS 14 1i'f3 lh,_lh, Kharitonov-Shvedchikov, Moscow 1992. 9 h3, motivated by fears of ... lClhS, leads into the Miles Variation of the Queen's Indian, discussed in Line

C21. The one independent, or semiindependent, line is 9 ... a6 100-0 lClbd7 11 l%cl cS 12 l%el l%e8 13 a3 l%c8 A.Nagy-S.Baranyaijr, Hungary 1993. In the Queen's Gambit move-order, 9 h3 is an unnecessary concession.

=

9.•.86 Black's most direct try, assuming he is not content to transpose into the main line with 9...cS, is 9 ...lClhS. Whether it is good is another question. 10 .tg3 ought to be harmless after either 10... lClxg3 11 hxg3 lCJd7, or 1O...lCJd7 immediately. Osipiak-Dervishi, Rimavska Sobota 1992 saw instead 1O...lCJxg3 11 hxg3 cS 12 ~h2!? .tf6 13 l%hl g614 ~gllCJc6 00; one of the quickest uncastlings on record. Instead after 9 ... %5, the move 10 .teS is critical: 1) After 1O... lCJd7, White is doing well on either 11 'ifa4 c6 (l1...lCJxeS 12 dxeS ±) 12 l%acllClxeS 13 dxeS g6 14lCld4 bS IS Wdl a6 16 f4 ± O.Foisor-Bonsch, Warsaw 1983, or 11 l%ct lClxeS 12 lClxeS g6 13 Wa4 .td6 14 .te2lClg7 ISlCJc6 'ifgS 16 f4 ± P.Neuman-J.Groh, Czech Republic 1996. 2) 1O...f6 lIlClgS! 'ife8! (lI ...fxgS? 12 Wi'xhS +-; 11...g6? 12lClxh7 fxeS 13lClxfS followed by .txg6 ±) leads to unexplored complications. However, these would seem to run in White's favour after 12 lClxh7! fxeS 13 dxeS lCJd7 14lClxf8 .txfS IS f4! - White is slightly ahead on material, the knight on h5 has no way back into the game, and White has kingside attacking chances by means of l%f3-h3. Black would need to find a substantial improvement to put 9 ... %5 back in business; his retarded queenside development works against him.

Systems with ... b6 Why though 9 ... a6, a seemingly passive move? The point is that if 9 ... ti)bd7? immediately, White has 10 ti)bS! ti)e8? (1O...cS 11 ti)d6 ±) 11 Wc2 winning a pawn. White missed this trick in Teichmann-Schallop, Nuremherg 1896, playing instead 10 l:lc I?! :t. Therefore ...a6 must come first to cover hS. 10 l:lct ti)bd7 1O... .i.d6 was met by 11 ti)eS l:le8 12Wf3ti)bd713 ti)xd7Wxd714l:lfdl c5 IS .i.xd6 Wxd6 16 dxcS bxcS 17 .tc4 :t in Uhlmann-Bonsch, Leipzig 1983. In his notes, Uhlmann gave 11 J.gS as even better, e.g. 11...ti)bd7 12 c4! dxe4 13 ti)xe4 .i.e7 14 ti)g3 :t, or 1l...l:le8 12 l:le1 ti)bd7 13 .i.fS :t. 11 ti)es cS 12 "13 l:lc8 13 "h3 g6 14.i.h6l:le8 White has the makings of a formidable kingside attack. In SoponyaiH1azsik, Hungary 1993, he played impatiently with IS f4?! ti)xeS 16 fxeS ti)e4 GO, whereas the 'mysterious' 15 l:lcdl, preventing Black's freeing plan. would have kept full control. To summarize, after 7 cxdS, the pawn recapture 7 ...exd5 looks distinctly favourable to White, and Black is probably better advised to recapture with the knight. However, White is not obliged to capture on d5 immediately; he can delay a move with 7 .i.d3.

D) 7.i.d3 (D) Naturally this could easily transpose to variations already given, after, for example, 7 ... .i.b7 8 cxdS exd5, though White must always be careful

187

B

not to capture in a position in which ... ti)xd5 is strong . The most popular moves for Black here are: Dl: 7••..i.b7 188 D2: 7...dxc4 189 First we shall consider a couple of alternatives: 1) 7 ....i.a6?! 8 cxdS exdS 90-0 c5 10 .i.xa6 ± Taleb-S.Mohd, Singapore 1995 has been noted under 7 cxd5 exd5 8 .i.d3 .i.a6?!, Line C221. 2) 7 ...c5 is given as the main line in ECO, but is seldom played: 2a) ECO gives 8 0-0 .i.b7 9 'ire2, transposing to a position we consider under 7 ....i.b7. 2b) 8 dxcS and now: 2bl) 8 ...dxc4 9 .i.xc4 'irxdl+ 10 l:lxdl .i.xc5 11 0-0 ti)c6 and now 12 ti)es ti)xe5 13 .i.xe5 .i.b7 14 h3 1/2- 1/2 was Szymczak-Rybak, Prague 1989. Maybe both players were happy with a draw before the game. We have already cited an Ibragimov-Lputian game (reached via 6...cS 7 dxcS .i.xc5 8 .i.e2 dxc4, etc., Chapter 4.4, Line All), in which White kept a slight edge by playing instead 12 e4 .i.b7 13 e5 ti)a5 14 exf6 ti)xc4 IS b3.

188

The Queen's Gambit Declined: 5 ~f4!

2b2) If instead 8 ...bxcS, then after 9 0-0 1Oc6 10 cxdS (10 "'e2!? keeps the tension; Black will not want to play ... dxc4) 1O...exd5 11 :c1 .i.e6 von Bardeleben-Em.Lasker, Hastings 189S continued 12 tOe2 W'b6 with approximate equality. A century later your author made no progress with 12 ~S l'bxeS 13 .i.xeS .i.d6 =in CrouchVan Voorthuijsen, Guernsey 1997. Yet a subtle shift in the position (placing the bishop on e2 rather than on d3) would change this to a substantial plus for White, which is the motivation for the 7 :cl variation described later. 2c) 8 cxdS is probably the best, transposing to the ... b6 Exchange Variation just considered. Your author would have played it in the Guernsey game, had he seen the simple tactic 8 ...cxd4? 9 iDxd4 iDxdS 10 .i.xb8! .I:Ixb8 111Dc6 winning the exchange.

01) 7•••.i.b7 (D)

w

80-0 It may well be that White should transpose to the ...b6 Exchange Variation with 8 cxdS here, and indeed this is White's most common option. In

this section we consider some independent lines. 8 "'e2 cS 9 0-0 leads back to the main text. 8•..cS 8 ...dxc4 leads to the 7 ...dxc4 variation, Line D2. 8 ... iDbd7 is again most simply answered by 9 cxdS. Instead, 9 :lc1lDhs 10 .i.esiDxeS 11iDxeSiDf6 12 cxdS exdS 13 f4 cS 14"'f3 a61S g4 cxd416 exd4 :c8 17 <.t>hl tOe8 =was Portenschlager-Klovans, Loosdorf 1993.

91i'e2 9 :cl cxd4 10 iDxd4iDbd7 = ECO. 9 dxcS!? bxcS (9 ...dxc4 10 .i.xc4 "'xdl 11 .I:Ifxdl .i.xcs 12 iDeS ± ECO) 10 "'e2 iDbd7 (for 1O... iDc6, see 9 "'e2 iDc6 10 dxcS bxc5, note '2b' to Black's 9th) 11 :fdl and now: 1) l1...lDhS?! 12 "'c2! iDxf4 13 .i.xh7+ \&Ih8 14 exf4 "fIc7 IS cxdSiDf6 and here, instead of 16 d6 .i.xd6 17 iDbS "'b6 ~ Yuferov-Klovans, USSR 1977, Lepeshkin suggests 16 .i.e4 "xf417 g3 "'g418tOeSW'hS191Dc6 with a clear advantage for White. 2) ECO recommends 11.. ....b6, continuing 12 :acl .i.d6 =, but one feels there ought to be scope for White to improve - maybe 12 .i.gS!? One possible point of capturing the pawn on move 9, rather than delaying, is so that Black cannot recapture with the knight, e.g. 9 'iFe2iDbd7 10 dxcS? iDxcS ;. 9•••iDbd7 The most solid response. 1) It is probably a little too early, given that Black is not fully developed, to open the centre with 9 ...cxd4 10 iDxd4 (10 exd4 dxc4 11 .txc4 00 ECO) 1O...iDbd7 11 cxdS iDxdS 12

=

Systems with ... b6 ~xdS .i.xdS. Now ECO gives 13 e4 .tb7 14 lIfdl .i.cS 15 li)fl 'ile7 = Forintos-Donner, Wijk aan Zee 1971, hut it is presumably possible for White In play more incisively, for example 13 .i.bS!? eS?! 14 .i.c6!? 2) 9 ...li)c6 and now: 2a) 10 lIadlli)b4 11 .i.bl .i.a6 12 ~S 'ilcs 13 a3ll)c6 14 .i.gS ± cxd4 15 exd4l:[dSI6l:fel h617 .i.Mdxc41S lLlxc6 'ilxc6 19 dS! lhdS (19 ... lLlxdS 20 li)xdS .i.xM 2Ili)b4 'ilcs 22 'ile4 .tf6 23 'ilh7+ ~f8 24 'ilhS+ g;e7 25 J:[xdS 'ii'xdS 26ll)c6+ +-) 20 :txdS (20 lLlxdSli)xdS ±) 20... exdS (20 ... li)xdS 21 li)xdS .i.xM 22li)b4 'ilcs 23li)xa6 'ii'xa6 24 We4 +-) 21 'ilxe7 lieS 22 liJxdS G;hS 23 WxeS+ li)xeS 24 lLlb4 'ii'bS 2Sli)xa6li)f6 26li)b4 as 27ll)c2 'ii'xb2 28 .i.xf6 gxf6 29 a4 fS 30 g3 bS :~ lli)e3 b4 32 .i.xfS c3 33 J:[bl Wa234 i.c2 and White's material advantage eventually proved decisive, even though it took another thirty moves to convert in Keres-Pelikan, Prague 1937. This entertaining combinative display from Ihe great Estonian did not even make his autobiographical collection - is Ihere a flaw somewhere? 2b) 10 dxcS bxcS 11 :fdl 'ilaS (ll...li)b4 12 .i.bl WeS 13 a3 dxc4 14 axb4 cxb4 ISli)a4 .i.c6 16 Wc2 .i.bS 17 b3±Zaitsev) 12a311fdSI3h3:d7 14 Wc2 ;t Forintos-Bobotsov, Hungary 1969. The general impression is that 9...ll)c6 is a little inflexible; it does not protect the cS-pawn, and gets in the way of the bishop. Hence 9 ...li)bd7. 10J:[fdl 10 h3 is unnecessary. 1O...a6 11 a4 lLle4 12 lIfdl cxd4 13 exd4lLldf6 14 cxd5 exdS 15 Wc2 lIcS 16 Wb3 .i.d6

189

17 .i.e5 lIe8 leads to equality, Tukmakov-T.Georgadze, USSR Ch (Tbilisi) 1975. 10.••a611 adS exdS12 b3 :teS13 J:[actli)h5 14 .i.eS g6 15 g4li)bf6 16 h3 White has some chances of a kingside attack, but his play is clearly slower than in the lines with 7 or 8 cxdS, li)eS, 'ilfl, etc., so the early exchange should be preferred. In TukmakovPetrosian, Vilnius 1978, Black sacrificed a pawn to gain active piece-play: 16... bS!? 17 .i.xf6.i.xf618dxcSd419 li)bl 'ilc7 20 e4 M 21 'ilc2 J:[ac8 22 b4 ll)e6 with unclear play; the game was later drawn. D2) 7.••dxc4 8 .i.xc4 .i.b7 (D)

A standard Queen's Gambit reaction by Black; he waits until White makes a move with his king's bishop, then exchanges on c4, gaining a tempo at the cost of conceding a little ground in the centre. 90-0 9 We2lLld5 (9 ... a6!?) 10 j.g3 c5 11 0-0 cxd4 12 exd4li)xc3 13 bxc3lLld7 14 j.d3 li)f6 IS lIad 1 li)h5 16 j.e5

190

The Queen's Gambit Declined: 5 jJ4!

l:lcS 17 c4 .td6 IS 'ire3 ~ GavrikovCiri~, San Bernardino 1991. It makes a change for White to have the hanging pawns in this system! 9•••a6 Not so much to prevent ~bS, but rather to prepare ... bS and ... cS, after which Black will have no reason to feel inferior on the queenside. While it is true that ... b7-b6-bSloses a tempo, White has also lost a tempo with .tfl-d3xc4, thus leaving the tempocount the same as in the Queen's Gambit Accepted. White's initiative is unchallenged by other methods: 1) 9 ...cS?! 10 dxcS .txcS?! (Black should prefer 1O.....xdl ~) 11 "e2 'ire7 12 .tgS ~bd7 13 e4 ± IvkovCudina, Sombor 1968. 2) 9...~bd7?! lO'ire2~ ll:acl (11 ~xe4 .txe4 12 l:lac1 as 13 .tbS l:la7 14l:lfdl 'ircs IS ~ .taS 16 e4 ± Ivkov-Pietzsch. Sarajevo 1966; horrible contortions by Black to defend his weak queenside) 11...~xc3 12 l:lxc3cSI3l:ldl cxd414~xd4±.tb4 IS ~e6 fxe6 16 .txe6+ ~hS 17 l:lc7 +- Liogky-Pushkov, Ore11994. lO'ire2 lO:Cl bSll.tb3 cS12dxcS ~bd7 13 c6 .txc6 14 ~2 ~S IS ~ed4 .td7 16 .tc2 l:lc8 17 .tbl "b6 18 'ire2 l:lfdS Ih-lh Tunik-Doroshkevich, Anapa 1991. lO••• bS 11 .td3! ~bd7 11...cS?! 12 dxcS, and now after 12...~bd7, 13 l:lfdl 'ireS 14 .tc2 ~xc5 soon led to a draw in TurianPushkov. Orel 1995, but can anyone explain why White didn't play 13 b4 or 14 b4. keeping the extra pawn? If Black plays 12....txc5. all he has

succeeded in doing is entering the line 6 ...c5 7 dxc5 .txc5 S .te2 dxc4 9 .txc4 a6 10 0-0 b5 11 .td3 .tb7 (Chapter 4.4, Line A12) a tempo down. 12 l:lfdl cS 13 dxcS .txcS 14 ~ b6 While Black may be happy on the queenside, White still has the centre and kingside. Now 15 ~xd7!? ~xd7 16 .tc2 ~ is promising (implying also that White could consider 12 l:ladl and 16 .tbl). while Dokhoian-Kveinys• Bad Godesburg 1993 continued more violently with 15 ~xb5!? axb5 16 .txbS .tdS 17 ~xd7 ~xd7 IS e4 .td4 19 .txd7 "f6 ~. White has an extra pawn. but it is difficult to convert. Thus the idea of 7 ... dxc4, S....tb7 and 9...a6 does not solve Black's problems. E)

7l:ld(D)

B

A natural plan in the Orthodox Queen's Gambit (l d4 d5 2 c4 e6 3 ~c3 ~f6 4 .tg5 .te7 5 ~f3 ~bd7 6 e3 0-07 l:lcl). but with even more point here in that there is direct pressure on c7. This is one of the few lines in the .tf4 system where White directs

Systems with ... b6 his main attention to the queenside rather than the kingside. The main lines to be considered, the second of which is perhaps the more accurate, are: 10;1: 7•••cS 191 ":2: 7•••.ib7 193 There are, as usual, a few minor alternatives to be considered, the most implausible being the game cited in BCO as continuing 7 ... lDbd7(?) 8 1i'c2(?) l:le8(??) 9 a3(?) with both players being apparently oblivious to the effect of a white knight reaching hS. Checking sources, in the game cited Black had actually played 6 ...c6 rather than 6 ...b6. Others: 1) 7 ... dxc4?! leaves Black a tempo down on the already unsatisfactory 7 .td3 dxc4 line: 8 .ixc4 cS 9 0-0 .ib7 10 'ii'e2lDbd7 11 dxcS .ixcs 12 l:tfdl and White has a clear advantage, Saltcrs-Maris, Hengelo 1995. 2) 7...lDhS?! is mistimed, with White being so far ahead in development. 8 .ig3 (8 .ieS!?) 8 ....ib7 9 cxdS exdS 10 .id3 lDd7?! 11 lDeS lDxg3?? 12 .ixh7+ +- Paasikangas-Moshin, Kishinev 1995. 3) 7 ... c6?! adopts a plan sometimes seen in ...c6 systems, but generally used only when White has played, as he often does, a waiting move such as h3. In the present position, White has consistently played developing moves, and his pieces are well enough activated exploit the weak square on c6. Or, as in A.Martin-A.Johanessen, Gausdal 1990, to take advantage of Black not having challenged the centre with ...cS: 8 .id3 .ib7 9 cxdS exdS IOO-Ol:le8111DeslDbd712'ii'f3M

191

13 g4lDg6 14 gSlDxeS IS .ixeSlDd7 16 'ii'h3lDfS 17 f4 ±. 4) 7 ... a6 8 cxdS exdS 9 .id3 cS, Doroshkevich-Vrulin, Kstovo 1994, has already been considered under 7 cxdS, Line C223.

E1) 7•••cS (D)

w

Now White really oUght to resolve the pawn structure in the centre, before Black develops his queenside. After 8 .id3 .ib7 9 0-0 lDbd7 10 lDe5, 1O... lDxe5 11 dxe51De8 12 "c2 g6 13 :rdl ± was Schlechter-Janowsky, Karlovy Vary 1902, but 1O...cxd4 11 exd4 dxc4 12 .ixc4 l:tc8 is fully satisfactory for Black. Instead, after 8 .ie2 .ib7 9 dxc5, 9 ... bxcS transposes to note' I' to White's 8th move, below. However, since White is not threatening to isolate the d-pawn, 9 ... .ixcS!? should be safe enough. 8cxdS

8 dxcS, as played in the YusupovLjubojevic game cited below, may ultimately transpose, e.g. 8 ... bxcS and now: 1) 9 .ie2 .ib7 10 0-0 lDbd7 (l0...lDc6?! 11 cxdS lDxdS 12 lDxdS

192

The Queen's Gambit Declined: 5 J.f4!

exd5 13 1IIa4 111b6 14 l:tfdl lOb4 15 ~5 l:tfd8 16 a3 ~c6 17 .tf3 ± Ufimtsev-Konstantinov, Sverdlovsk 1951) 11 cxd5 (11l'tle5lOxe5 {1l...lOb6!? 12 cxd5lObxd5} 12 .txeS dxc4 13 .txc4 111b6 14 1IIe2 l:tfd8 15 e4 a6 16 .tf4 ~d7 17 f3 ;t Obukhov-Zhukovitsky, Krasnodar 1991) and after 11...exd5 12lOe5 we have transposed back into the main line. Since on Dvoretsky's reckoning 1l ...lOxd5! would cut down White's advantage significantly in this particular move-order, there would seem to be a strong case for taking on d5 as early as possible, to make it less enticing for Black to recapture with a knight. 2) The one possible objection to an early cxdS for White in this variation is that it gives Black the chance to defend his centre with ....te6. For example, Guedes-de Asis, Havana 1997 went 9 cxd5 exd5 10 .te2 .te6 (1O....tb7 11 0-0 lObd7 transposes to the main line) 11 0-0 h6 12 ~e5111c8 13111a4 ~fd7 14 ~xd7 ~xd7 15 .tf3 ~f6 16 l:tfdl l:td8 17 h3'ifb7 18 .te5l:td7 19 1IIf4; nevertheless Black has not succeeded in relieving the pressure on his pawn centre. 8...exdS 8... lOxdS 9 lOxd5 1IIxd5 (9 ... exd5 10 .td3 lOd7 11 0-0.tb7 12 1IIc2 g6 13 dxc5 bxc5 14 :fdllOf6 15 .th6 ± Agdestein-Spassky, Gjl/lvik 1983) 10 .tc4 1IId8 11 0-0 .tb7 (11...a6?! 12 dxc5.txcS 13 ~e5111e7 14111g4 f6 15 ~3 ~h8 16 ~xcS bxc5 17111f3 .tb7 18 .td6111d7 19 .txe6 winning a pawn, Kallai-Zo.Varga, Hungarian Ch 1993) 12 dxc5 .txc5 13111c2?! (13 ~eS!?;t) 13 ... .txf3 14 gxf3 ~d7 IS .tbS 'fIe7 16 1IIa4 lOf6 with an equal position in

the game Kallai-T.Bauer, Hungarian Cht 1992/3. 9.te2 9 .td3 .tb7 10 0-0 lObd7 00 is covered under Line C21, 7 cxd5 exdS 8 .td3 c5 9 0-0 .tb7 10 l:tcl. It is more flexible for White, having already played l:tc 1, to keep the bishop off the d-file. 9••..tb7 10 0-0 ~bd7 11 dxc5! In strategic terms, Yusupov's real innovation in his game against Ljubojevic, quoted in full in the Introduction, was not so much the individual move 8 dxc5. Rather it was the appreciation that this set-up could be treated by means of an assault on the hanging pawns, rather than by trying to create a permanent outpost on eS. 1) 11 l'tles is the older try: la) 11...lOxe5 12 .txe5 lOe4 13 ~xe4 (13 dxc5 .txc5 14 ~xe4 dxe4 IS .tc4 "'g5 16111d7111xe5 17111xb7 :ae8 18 :cd 1 with at best a microscopic edge, Antoshin-Donner, Havana 1968) 13 ... dxe4 14 "'a4 .tf6 15 .txf6 1IIxf6 16 dxc5 bxc5 17 1IIbS ;t Ree-Donner, Leeuwarden 1981. 1b) 11.. .:c8 12 dxcS ~xc5 (after 12 ... bxcS?! 13 "'a4 ~xe5 14 .txe5 111b6 15 :fdl :fd8 16 .tf3 Black's hanging pawns cracked in Ibragimov-Galdunts, Kherson 1991), and in Gheorghiu-Ree, Wijk aan Zee 1981, White could find nothing better than the retrograde 13 lOf3 lOfe4 14lOd4 .tf6 15 .tg4lOxc3 16 l:txc3l'tle4, with equality. 2) The quiet 11 h3 has also been tried to no great effect: 11...lOe4 12 dxc5 ~xc3 13 l:txc3 bxc5 = R.Kempinski-Gyzinski, Poland 1993, or l1...a6 12 dxc5 bxc5 13 l'tle5lOxe5 14

Systems with ... b6

=

i.xeS i.d6 Agzamov-Azmaiparashviii, Erevan 1981. 11... bxc5 If 11...lClxcS, White is two tempi ahead of Gheorghiu-Ree above. 12 lLld4 should preserve an advantage. 12 lCleS! lClb6 12... lLlxeS 13 i.xe5lCld7 14 i.g3 ± Dvoretsky. 13a4! Forcing Black to make concessions in his queenside pawn structure, in view of the threat of as and a6. The immediate 13 i.f3 'iVc8 14 lCla4 (14 a4!?) 14"'lClxa4 15 'fIxa4 'fIe6, Tolstikh-Anikaev, Cheliabinsk 1991, is not so incisive. 13... a5 14 i.r3 :e8 IS lLlbS! ± Yusupov-Ljubojevic, Tilburg 1987 (see Historical Introduction). E2) 7...i.b7! (D)

This is a much safer line for Black. By holding back the c-pawn, and by being prepared to recapture on dS with a piece, he minimizes White's chances of achieving a favourable pawn structure. 8cxdS

193

8 i.e2 and then: 1) 8 ...dxc4 9 i.xc4lClbd7 10 h3 cS 11 0-0 a6 12 a4lLldS (l2 ... cxd4!?) 13 i.h2lCl7f6 14 'it'e2 'it'e8 ISlCle5 with a small advantage for White, SemionovMuraviov, Alushta 1994. 2) 8 ...cS 9 0-0 (clearly White could consider 9 cxd5 or 9 dxcS) 9 ... lClbd7 10 cxdS lClxdS 11 i.g3 lCl7f6 Ih_lh Lechtynsky-Chiburdanidze, Banja Luka 1985. 8 ...lClxdS! 9lClxdS 9 i.g3 is less aggressive, but not totally harmless: 1) 9... cS 10 lClxdS (10 i.d3lCld7 11 0-0 lCl7f6 12 :el lClxc3 13 bxc3 i.e4?! {13 ... lCle4 would have been equal} 14 i.a6;1; D.Rajkovic-M.Kapelan, Vdac 1983) 10...i.xdS 11 dxcS bxc5! (11...i.xc5 12 'fIa4 as 13 a3;1; Kapelan) 12 i.c4 'it'b6! 13 i.xdS exd5 14 'it'xdS?! (Petrosian was not normally a man in a hurry; 14 O-O!, hoping to win the pawn thirty moves later, makes it more difficult for Black) 14...:d8! IS 'it'b3 (IS "xaS 'iWb4+ 16 ~f1 'it'xb2 17 :el 'it'bS+ 18 ~gllClc6 + Kapelan) IS .....aS+ 16~e2lClc617 :hdl i.f6 and Black's active pieces make it very difficult for White to utilize his extra pawn, Petrosian-M.Kapelan, Vdac 1981. 2) 9 ... lLlxc3 10 bxc3 cS could be considered. 9..:ibdS 10 i.c4 10 :xc7?! i.b4+ II lCld2 lCla6 12 :el 'iVxa2 ±; 10 i.xc7 i.b4+ 11lCld2 'iVxa2 ;I; Kapelan. 10..:.a5+ 11 ~e2 i.d6?! Slightly limp. Surely Black can't be worried about that c7-pawn? ll...lCld7 should equalize, since if 12 .i.xc7?!, then 12 ...:ac8 13 i.g3 (13 i.b3??

194

The Queen's Gambit Declined: 5 1./4!

7lbbS and 6 ...•d7?! 7 .xd7+ ~xd7 (otherwise White exchanges as necessary on dS, then wins the c-pawn) 8 lOeS+ ~e8 9 cxdS with a secure positional edge for White. After 6....td7 7 'iib3 lLlc6 (7 ... dxc4 8 'ili'xc4 cS 9 dxcS .txcS 10 :dllLlc6 II e3 leaves Black under a lot of presIn conclusion, 7 :cl may not be as sure) 8 cxdS lOaS 9 'ili'c2 exdS 10 e3 strong as its reputation, but 7 .td3 and and White stands better, both Black's 7 cxdS both offer excellent prospects queenside minor pieces being misof a comfortable edge for White. placed. 7cxdS There is just one loose end to be After 7 .txb8 :xb8, 8 .xc6+?? tied up; Black can perhaps delay cas- .td7 traps the queen, while 8 .xa7 .td7 9.a4 cS 10 .b3 dxc4 II .xc4 tling and try 5...b6. bS gives Black compensation for the F) pawn. 8 lLles .td7 is critical, for ex1 d4 dS 2 c4 e6 3 1Oc3 lOr6 4 lOr3 ample 91Oxc6 (otherwise Black is se.te7S .tr4 b6 ~ .. :·.·:cllf.tlL9 ... 'iIi'c7 10 'it'xa7! .xa7 11 It seems a little strange that while . 'lOxa7 dxc4. Black is a pawn down, but S... c6 instead of 6 ... c6 is frequently has the two bishops and prospects of played, S... b6 has been ignored. How- active play. It is perhaps safest to deever, if Black can substitute ....tb7 for scribe the position as unclear. ...0-0, then he is more likely to be able The text-move aims for a more poto make a safe recapture on dS with a sitional solution. piece, and in some lines (for example 7•••exdS 8 g3! Fianchettoing the bishop makes it 6 e3 .tb7 7 .td31OhS!?) he can play ...lLlhs without having to worry about much more difficult for Black to free a later .txh7+ or attacks against the his position with ... cS. White should king along the open h-file (after .tg3 be better, in a very technical sort of polLJxg3). sition. This line reminds me of a sysI do not propose to analyse this in tern in the a3 Queen's Indian (1 d4 lOf6 2 c4 e6 310n b6 4 a3 .tb7 SlLlc3 detail; what follows is just an outline. 6 .a4+!? dS 6 cxdS exdS 7 g3 .i.e7 8 .a4+ c6) 6 cxdSlOxdS 7 .tg3 (71Oxd5 .xdS except that White has been able to 00) 7... c5 00. substitute .tf4, which is the thematic 6•.•c6 developing square for the bishop, for Black should avoid both 6...lObd7?? the by now redundant a3. .ta6+ -+; 13 .td3 .xa2 +) 13 ...:xc4 14 l:r.xc4 .ta6 with a slight advantage for Black. 12 ~S .txeS 13 .txeS lbd7 14 .txc7 :rc8 15 .td3 .txg2 16 :gl 'ii'dS 17 ~e1;t Svirin-Yanvariev, Naberezhnye Chelny 1988.

7 Systems with ... c6 1 d4 dS 2 c4 e6 3 itX3 lbi'6 4ll)f3 JLe7 5 i.r4 0-0 Non-transpositional lines involving S...c6 will be covered at the end of the chapter. 6 e3 c6 (D)

system. Such a reader will presumably want a lot more than 7 'iFc2 ;t or 7 ':c 1 ;t, and so I have resisted the temptation to truncate coverage in line with what I would estimate to be the true theoretical significance of this defensive system.

7.1 7 'ifc2 ~bd7 1 d4 dS 2 c4 e6 3 M lbi'6 4lbi'3 i.e7 5 i.r4 0-0 6 e3 c6 7 'iFcZll)bd7 (D)

Black is not aiming for excitement, and indeed one gets the impression from the database that this is the line that is all too frequently wheeled out if Black is anxious about the opponent's possible superiority of opening preparation. To an even greater extent than with the ... b6Iines, White can achieve a slight advantage just by playing natural moves, without any detailed knowledge of theory. Perhaps partly for that reason, this system is largely ignored by ECO, getting just one line plus footnotes of coverage, compared with nine lines on the Old Main Line. This variation will probably be encountered a lot at club and congress level by the player who tries the JLf4

The first key position. The theoretician's frustration at dealing with this system can be seen from the fact that just about every reasonable move from a3 to h3 has been tried, with most of the lines leading to a slight edge for White. Why though should White play a nondescript pawn move rather than completing his development? The answer lies in what is commonly known as 'the battle of the tempo'. To free his position, Black will want to play

196

The Queen' s Gambit Declined: 5 J.J4!

something like ...dxc4 followed by ... b5 and ... .i.b7. If White plays an early .i.d3, Black will generally be very happy to make this exchange. White having lost a tempo with his bishop. If. however. White can play a useful waiting move like h3. and Black then captures on c4. it is White who has won the battle of the tempo. There are even a few instances in this chapter where White plays a3 and h3. and Black plays ... a6 and ... h6. each player waiting for something to happen. White. with more space to manoeuvre. has nothing to lose from these tempo battles, but Black often has to think of some way to free his rather constricted pieces without opening the position. The usual if convoluted plan is ... l:e8 followed by ... ~f8-g6 and ....i.d6. and suddenly Black is ready for ... e5. Now: 196 A: 8 h3 B: 8l:dl 205 C: 8cxdS 208 D: 8 .i.d3 208 E: 8 .i.e2 209 F: 80-0-0 210 G:8a3 211

A) 8h3 The benefits of this waiting move are readily appreciated; the bishop now has a retreat on h2, while assistance is given for a possible later pawn-storm with g4. Now Black can choose from: AI: 8•••l:e8 196 A2: 8•••dxc4 200 A3: 8•••86 202 A4: 8...b6 203 204 AS: 8.....aS

A6: 8••• g6?! A7: 8...h6

204 204

A1) 8 •••l:e8 (D)

White's main choices here are 9 0-0-0. 9 l:dl and 9 a3. All: 9 i.h2!? 196 AU: 9 .i.d3 196 197 A13: 9.i.e2 A14: 9 cxdS 197 AIS: 90-0-0 197 198 A16: 9l:d1 A17: 9 a3 199 A11) 9 i.h2!? deserves to have been played more than once. White's plan worked smoothly in Conquest-Hogg. Kuala Lumpur 1992. though Black's play can doubtless be improved: 9...00 10 0-0-0 dxc4 11 .i.xc4 b5 12 .i.e2 .tb7 13ltle5"a5 14 Wbl l:ac8 15 f4 .i.b4 16 ltle4 ~S 17 "b3. Maybe 9 ...dxc4 10 .txc4 bS.

A12) 9 i.d3 is too much of a routine developing move; it does nothing to prevent Black's freeing plan. 9 ... ~f8 10

Systems with ... c6 0-0 (10 g4 .i.d6 11 ~5 dxc4 12 .i.xc4

li)d5 13 0-0-0 ~xf4 14 exf4, Seredcnko-Mainapalert, Manila OL 1992, 14... f6!? 00) 1O...~g6 11 .i.h2 .i.d6 12 ~xd6 .xd613 .i.xg6hxg614e4 .d8 R.Kempinski-Asrian, Erevan OL 1996. Black has successfully exchanged two pairs of minor pieces, and although White has an obvious advantage in space, it is difficult to see what damage he can do to the black position; if for example 15 b3 b6, even the bad hishop has a life. In the game, White over-pressed with 15 e5 ~h7 16 Wd2 h6 17 lIfel .i.a6 IS c5? (18 cxd5 =) 18 ... bxc5 19 dxc5 20 We3 lIabS 21 l:.abl ~f8:j:.

.as

A13) 9 .i.e2 ~f8 10 0-0 ~g6 11 .i.h2 i.d6 12 .i.xd6 Wxd6 13 c5 'fIc7 14 i.d3?! merely leaves White a tempo down on the 9 .i.d3line, but in the game Loncar-Ranieri, Rebecchetto 1996, Black obligingly played 14...~f8? 15 li.:le5 ±. A14) 9 cxd5 ~xd5 10 ~xd5 exd5 11 0-0-0 ~f8 12 .td3, with two games from the same tournament diverging: 1) 12....i.e6 13 ~bl :CS 14 g4 (14 h4!?;t Speelman) 14....i.d6 15 Wa4 a6 16 .i.xd6 .xd6 17 lIcl lIc7 with a very slight edge to White which could easily be (and was) overturned, Khalifman-Speelman, Moscow GMA 1990. 2) 12 ... .i.d7 13 ~bl lIcS 14 g4 'fIb6 15 lIc 1 c5 16 dxc5 lIxc5 17 'fIb3 'ii'xb3 18 axb3 ~e6 19 .tg3 lIecS 20 lIcdl .tb5 21 .txb5 lIxb5 22 lId3 ;t Gelfand-Azmaiparashvili, Moscow GMA 1990.

197

A15) 90-0-0 (D)

This is promoted to joint main-line status for subjective reasons: I like it! It is surprising that this ultra-direct move is not tried more often; after all, castling queenside is now an accepted part of White's arsenal against ...c5, and the queenside is being opened up rather more slowly here.

9•••dxc4 1) 9 ...~f8 10 .td3 (10 .i.h2!?) 1O...~g6 11 .te5 ~7 12 .txg6 hxg6 13 h4 ~f8 14 h5 ± Cattomio-Allerheiligen, Bie11993. 2) 9 ...'fIaS (0stenstad-Lein, Gausdal 1990) 10 ~d2! borrows an idea from the New Main Line, yet is apparently an innovation. White must surely be doing very well. 3) 9 ...b6?! 10 cxdS cxdS (lO...exdS? 11 ~b5! wins the exchange) 11 .tb5! (11 ~b5 .ta6! 12.tc7 {12 ~7 .txfl 13 ~xa8 .txg2!} 12...'fIcS 13 ~6 .i.xd6 14 .txd6 .txfl) 11.. ..tb7 12 ~5 a6 13 ~c6 .txc6 14 .txc6 gives White a slight advantage.

10.txc4 bS Not 1O...~d5?? 11 .txdS exdS 12 ~xdS cxdS? 13 .tc7 +-.

198

The Queen's Gambit Declined: 5 i.f4!

11 1.d3 'ii'b6 12 g4 ~f8 13 ~e4

lbds

13 ... ~xe4 14 1.xe4 1.b7 ±. 14 ~eg5 g6 15 h4 ~b4 16 'ii'b1 e5!? Fritz's idea is critical. Crouch-Bielby, Durham 1997 continued 16... cS 17 h5 1.b7?! (17 ... cxd418 hxg6; 17 ...~xd3+ 18 1i'xd3 ±) 18 hxg6 fxg6 19 1.xg6! hxg6 (I was hoping for the chance to make a queen sacrifice with 19... ~xg6 20 1i'xg6+! hxg6 21 1.eS 1.xgS {or 21...~xa2+ 22 ~bl 1.e4+ 23 ~al 1.xgS 24 ~xgS .txhl 2S l:lxhl ~fS 26 .tf6 with mate on h8} 22 l:lh8+! ~f7 23 ~xgS+ ~e7 24 l:lh7+ ~d8 2S dxcS+ and White regains the queen, keeping a decisive attack) 20 .teS 1.xgS 21 ~xgS 1.xhl 22 l:lxhl 1-0. I enjoyed playing this game. 17 ~xeS f6 18 h5! fxeS 19 bxg6 h6 20g7! 20 .txeS?! 1.xgS 21 g7 l:lxeS! 22 gxfS1i'+ ~xfS 23 dxeS 1i'cS+ gives Black a strong counter-attack, and 20 l:lxh6 is slow. 2O•••1.xgS 20...~xg7 (20...hxgS? 21 .txeS +-) 21 1.xeS+ 1.f6 22 1.xf6+ ~xf6 23 l:lxh6+ ~xgS (otherwise White is ahead in material) 24 l:ldhl +-. 21 gxf8'ii'+ ~xf8 22 1.xgS bxgS 23 1.c4!? This position contains a wealth of fascinating tactics. White has no trouble drawing, but a win is elusive. 23••• bxc4 24 'ii'g6 ~a2+ 25 ~d2 :e6 26 'ii'xgS ~e8 and now: I) 27 ~el 1i'b4+ 28 ~fl 1i'e729 l:lh8+ ~d7 301i'fS 1i'd6! 31 l:lh7+ (31 ~e2 ri;c7) 3l...~d8 32 1i'gS+ (32 ~e2 .l:.b8 33 dxeS l:lxb2+ 34 ~el .l:.e2+ and 32 ~e I e4 33 1i'a5+ ~e8 34 1i'xa2

'ii'b4+ 3S ~fl 1i'b3 are bad or lost for White) 32... l:le7 33 l:lxe7 1i'xe7 34 dxeS+ 1.d7 3S 1i'g8+ 1i'e8 36 1i'xc4 ~c7 37 1i'xa2 1i'xeS and Black is better.

2) 27 ~e2 ~d7 28 l:lh7+ ~d6 29 1i'g7 ~dS! 301i'fS c3! opens the a6-fl diagonal for a juicy check and clears c4 for the king to flee to the queenside. After 31 dxeS+ ~c4 32 l:ld4+ 1i'xd4! 33 exd4 l:lxeS+ 34 ~fl cxb2 Black wins easily. 3) 27 l:lh7! 1i'xb2+ 28 ~el 'ifb4+ 29 ~fl ~3 (forced) 30 l:lh8+ 'iffS 31 'ifhS+ ~e7 and now, after 32 l:lxfS ~xfS 33 l:lcl !? (or maybe with checks on h8 and h7 first) the position is very unclear. A16)

9l:ldl (D)

B

A good square for the rook, and the most popular move here. But why this strange reluctance to have the king on cl as well? 9...~f8 9 ...1.fS 10 cS!? b6 II b4 a5 12 a3 axb4 13 axb4 bxcS 14 bxcS 'ifa5 15 1i'd2 1.a6 16 g4 ± Van Wely-Ravia, Antwerp 1995.

Systems with ... c6 9...•aS 10 a3!? ~4 ll.i.d3lDdf6 (l1.. ..i.xa3? 12 l:tal! lDxc3 13 l:txa3 lDa4+ 14lDd2 followed by b3 wins a piece) 120-0 lDxc3 13 bxc3 dxc4 14 i.xc4lDds IS .i.d3 g6 16 c4!? lDxf4 17 exf4 .xa3 18 cS.aS 19 h4 .c7 20 g3 and White's positional bind is well worth the pawn, Lima-de Andrade, Brasilia 1997. 10i.d3 lO.!DeS ~4! ll.i.d3 f6 12lDf3 fS 13 lDeS i.f6 140-0 .i.xeS IS .i.xeS lDg616i.h2.gS17lDe2lDit418lDf4 =Landa-Astian, Russia Cup (Novgorod) 1997. 10•••dxc4 1O... lDg6 II .i.g3 i.d6 12 .i.xd6 .xd6130-0.c714e4dxc4IS.i.xc4 cS 16lDgS l:tf8 17 dS cxdS 18lDxdS lDxdSIgexdS.d820h4i.g421 f3?! (21 l:td2 ±) 21...bS 22 fxg4 (22 .i.b3 23l:tf2 :rc8 followed by ...m) 22... bxc4 23 .fS .c8 00 A.ZakharovAstian, Moscow 1996. 11 i.xc4lDds II...bS 12 .i.d3 i.b7 13 0-0 l:tc8 looks like loose queenside play by Black. Then, rather than 14 a3 .b6 IS b4 as 161i'b2 axb4 17 axb4lDdS 18 lDxdS exdS = Elsness-Kosasih, Djakarta 1997, White can, for example, gain space in the centre with 14 e4 followed perhaps by 'iVe2, when Black will find it hard to generate counterplay. 12.i.g3 12 O-O!? should also be possible. As we shall see at various points in this chapter (Section A2, etc.), capturing on f4 does not guarantee Black easy equality, despite the advantage of the bishop-pair and the doubled f-pawns. 12.••i.d6

.b6+

199

12 ... lDxc3 13 bxc3 (or 13 .xc3) 13 ....i.d6 14 .i.xd6 .xd6 IS 0-0 will transpose. 13 .i.xd6 'iVxd6 140-0 lDxc3 IS bxc3 IS .xc3 b6 16 l:tc1 .i.b7 17 i.e2 lDd7 18 :rdl .e7 19.!De1 eS =Kobaliya-Astian, Russia Cup (Moscow) 1996. IS•••b6 16 .i.d3 'iVc7 A little refinement to cover f7. Lima-A.Hoffman, Matanzas Capablanca mem 1994 continued instead 16... i.b7 17 lDgS g6 (17 ...h6?? 18 i.h7+ lDxh7 19 .xh7+ ~f8 20 lDe4 +-) 18 f4 cS 19.f2;t. 17 c4 cS IslDgS h6 Now this is perfectly possible, and gives Black a very solid position. 19 .i.h7+ ~hS 2O.i.e4 l:tb8 21 .i.b7 =Razuvaev-Astian, Russia Cup (Novgorod) 1997. Ultra-solid defence from the land of Petrosian. The exchange of two pairs of minor pieces ensures that Black is able to manoeuvre freely behind the lines without feeling cramped.

m

A17) 9a3(D)

200

The Queen's Gambit Declined: 5 1./41

White continues to play the waiting game. One possible justification for this move is that White may have in mind, after Black's freeing manoeuvre, b4 and l:acl restricting Black's queenside. 9...lDfS 1) 9 ....tfS 10 .td3 dxc4 11 .txc4 lDdS 12 .th2lDxc3 13 bxc3 b6 140-0 .tb7 IS .td3 g6 16 a4 cS 17 e4 "ii'c8 18 1IM2.ta6 19 as ~ Gelfand-H.Olafsson, Novi Sad OL 1990. 2) 9 ...dxc4 10 .txc4 bS 11 .td3 (11 .ta2!?) and now: 2a) 11.. ..tb7 120-0 a6 13 b4 as 14 l:abl h6 IS e4 axb4 16 axb4 l:a3 17 .tcl l:a6 18 .te3 ~ L.Hansen-Lyrberg, Copenhagen 1994. 2b) 11...a6 12 b4 as 13 l:bl axb4 14 axb4 'iVb6 IS 0-0 .tb7 16 e4lDfS 17 l:fdl l:ec8 18 .tgS 'ifd8 19 eSlDdS 20 lDe4 f6 ~ Hjartarson-G.Thorhallsson, Reykjavik 1994. 10.td3 10 l:d 1 and now, instead of the inferior 10....td6?! lllDeS a612 cS .txeS 13 .txe5 ± A.Ivanov-Korkina, Vladivostok 1994, 10... lDg6 11 .th2.td6 would be more thematic. 10•••dxc4 1O...lDg6 11 .th2 .td6 12 .txd6 'iVxd6 13 0-0 (13l:dl 'ife7 140-0 dxc4 15 .txc4 b6?! 16 ttle5 ± P.Iotti-Doleschal, Salsomaggiore 1997) 13 ...dxc4 14 .txc4 lDd5 15 l:adl lDxc3 16 'ifxc3 b6 17lDe5lDxe5 18 dxeS "ii'e7 ~ Mikhalchishin-Brestian, Tmava 1988. 11 .txc4lDds 12 .tg3lDxc3 12....td6 13 .txd6 'W'xd6 14 lDe4 "ii'd8 150-0 lDg6 16 l:adl b6 17lDe5 .tb7 18 'ife2;t lDf6? 19lDg5 ± SavonGerbich, Vladivostok 1995. 13 Wxc3 .td6 14 O-O!?

14 .txd6!? 'W'xd6 IS b4 b6 160-0 .tb7 17 l:fdl l:ed8 18 l:d2 ~ Beliavsky. 14....txg3 15 fxg3 .td7 16 g4 We7 17 gS l:ec8 It takes a lot of ingenuity for White to create anything unusual in such a stodgy variation. White now made a pawn move too many on the kingside in Beliavsky-Speelman, Linares 1991: 18 h4?! c5 19 d5 exd5 20 .txd5 .tc6 =. Beliavsky suggests instead trying to spread the bind across the whole board with 18 b4 as 19 l:abl axb4 20 axb4 .te8 21 l:f2 b6 22 h4 l:c7 (22 ...c5 23 dxcS bxc5 24 b5 ±) 23lDd2 l:ca7 24

lDe4±. White's best chances of securing an edge seem to be in the ultra-violent but under-explored 9 0-0-0 or in the outwardly ultra-quiet 9 a3. Black can handle 9 l:dl.

A2) 8 •••dxc4 9 .txc4 (D)

B

Black apparently disregards the battle of the tempo. The plan has its logic, however, in that Black can offer a double exchange of minor pieces

Systems with ... c6 rather more economically than after ...l:le8, ...M, etc. 9 •••lOb6 I) 9 ... lOdS?! and now: la) 10 .th2?! and here: lal) 1O...107f6?! 11 0-0 b6 12 l:lfdl (too quiet; 12 e4!?) 12 ....tb7 13 a3 .td6 =B.Reilly-Conrady, Moscow 19S6. la2) 1O... bS?! II.te2107f6120-0 i.b7 13lLlxdSlOxdS 14lLld2;1;, and in Psakhis-Onoprienko, Groningen 1995 Black tried to solve his difficulties by tactical means, but after 14 ... cS IS dxcSlOb4 16 'ifc3 .tf6 17 Wxb4 as 18 'it'xbS .ta6 19 'ifxa6 l:lxa6 20 .txa6 the passed c-pawn soon told. la3) 1O...107b6! 11 .tb3 (11 .td3 h6 12 a3 .td6) 11.. ..td6 is Black's hcst. Ib) 10 .tg3, which seems to put the bishop on a better square, is still well answered by 1O...107b6!, just as in line' la3'. lc) The real problem is that White can play 10 lOxdS exdS 11 .td3 .tb4+ 12 ~e2, reaching Vaganian-Forintos in Section C (8 cxdS lOxdS 9 lOxdS exdS 10 .td3 .tb4+ lllit>e2) a tempo up. Since a possible kingside pawnstorm is part of White's agenda, the extra h3 tempo is useful. 2) 9 ... bS 10 .td3 .tb7 11 e4 b4 12 lLla4 cS 13 dxcS l:lc8 14 eS lOxcs IS lOxc5 l:lxcS 16 'ife2 lOdS 17 'ife4 g6 18 .th6 lOc3 19 'ife3 l:ldS 20 .tc2 ± Mainville-Brumen, Cannes 1997. If Black wishes to prepare for the 'semiMeran' formation, 8...a6 (Line A3) is a more natural way to do so. 10.tb31OfdS Black drifted into complete passivity after 1O... l:le8 11 0-0.td6 12 .tgS

201

(12 .teS!?) 12... lObd7 13 l:lfdl 'ikc7 14 :tacl a6 IS We2 'ifaS 16 e4.tf8 in Doering-Genduso, Nuremberg 1989. This is exactly the type of nightmare that Black must work hard to avoid in the ...c6 variation. 11 0-0 Ironically, the extra move h3 works against White here, the point being that if the pawn had still been on h3 (after 8 .td3 dxc4 9 .txc4, etc.), White could happily drop the bishop back to g3, planning to recapture with the hpawn after ... .td6xg3. Here though, 11 .tg3 .td6 12 e41Oxc3 13 bxc3 (13 .txd6!?) 13....txg3 14 fxg3 cS! should be satisfactory for Black; White's pawns are too badly split to be a real threat. 11...lOxf4 12 exf4 Establishing what might be described as the 'Rubinstein Bind'. White intends to keep a complete grip on the central and kingside dark squares with pawns, with space being created on the kingside by g3 and h4. Black's position is more difficult to defend than it looks, and any plan to open the centre quickly (for example 12... cS 13 dxcS .txcS 14 lOe4 ±) is likely to come unstuck. 12...lLlds 13 g3 Now: 1) 13 ... lOxc3 14 Wxc3 Wd6 IS l:lac1.td7 16lLleS l:lad8 17 :fdl.tc8 18 a3 .tf6 19 We3 'ike7 20 h4 l:ld6 21 hS :fd8 22 :c3 and the bind has been tightened considerably, RubinsteinCanal, RogaSka-Slatina 1929. 2) 13 ... f6 covers both eS and gS, and prepares for possible counterplay later on with .. :ii'e8-hS. If White plays lOxdS, Black will generally want to capture with the d-pawn. However, 14

202

The Queen's Gambit Declined: 5 i.,f4!

l:fel "eS 15 f5 'ifh5 16 ~g2.i.b4 17 fxe6 causes problems for Black, e.g. 17 ...l:eS IS .!l)xd5! cxd5 19 g4! "g6 20 "xg6 hxg6 21 .i.xd5 .i.xel 22 e7+ ~h7 23 l:xel .i.d7 24.i.f7 and White's connected passed pawns should win the endgame, or 17....i.xc3 IS bxc3 %leS 19 c4 .i.xe6 (19 ....!l)b4 20 "d2 .i.xe6 20 l:xe6 is similar) 20 l:txe6 and wins. The whole line with S...dxc4 is so little explored that the theoretically critical game is nearly 70 years old.

A3)

8•.•a6 (D)

This move takes its inspiration from a main line of the Meran Variation of the Semi-Slav (1 d4 d5 2 c4 e6 3 ~3 .!l)f6 4 .!l)f3 c6 5 e3 .!l)bd7 6 .i.d3 dxc4 7 .i.xc4 b5 S .i.d3 a6). Black's plan is to pick a moment to play ...dxc4, then follow up quickly with ... b5, ... .i.b7, ... c5, etc.

9:ctl At least Black knows White is not going to castle queenside. Alternatives: 1) 9 a3 dxc4 10 .i.xc4 bS 11 .i.a2 .i.b7 12 l:dl "b6 (12 ...b4!?) 13 .i.bl l:fdS 14 e4.!l)f8 15 e5 ~5 16 .!l)xdS

exdS ~ Ivkov-Kurajica, Wijk aan Zee 1970. 2) 9 cS b6 10 cxb6 "xb6 11 .i.d3 cS 120-0 .i.b7 13 l:adl :acS 14 "bl h6 15 :fel :fdS (1S ...cxd4 00) 16.i.c2 "as 00 Shulman-Lyrberg, Minsk 1996. 3) 9 .i.d3 dxc4 10 .i.xc4 b5 11 .i.e2 .i.b7 120-0 c5 13 l:fdl 'ilb6 14 .!l)e5 :fdS 15 .!l)xd7 :xd7 16 dxc5 "xeS 17 l:xd7.!l)xd7 IS :dl .!l)f6 =KobaliyaAsrian, Russia Cup (Krasnodar) 1997. There are too many options on both sides to draw any strong conclusions from these lines. 9..•b5 Deciding that it is not worth playing the waiting game any longer. 1) 9 ... h6?! (not one of the most successful of recent grandmaster innovations) 10 a3 dxc4 11 .i.xc4 and now: la) 11...~S 12 O-O! .!l)xf4 13 exf4 "c7 (after 13 ...:eS 14~5 .!l)f815 f5 .i.f6 16 .!l)e4 "e7 17 .!l)g4 .!l)h7 18 l:fel Black's position is close to collapse-Kramnilc) 14~5~615.i.a2 .i.d7 16 .i.bl .i.eS 17 d5 ± KramnikLputian, Debrecen Echt 1992. Ib) 11...b5 12 .i.a2 .i.b7 13 e4?! (13 .!l)e5! looks very strong, and if 13 ...%leS, then 14 .i.xh6 gxh6 15 .i.xe6 with a winning attack) 13 ...:eS 14 .i.bl .!l)f8 15 b4 as 16 "b3 axb4 17 axb4 .!l)6d7 18 e5 00 Psakhis-Pigusov, Novosibirsk 1993. 2) 9 ...l:e8 10 a3 lLlf8 11 cSlLl6d7 (11...b6 12 b4 ±) 12 .i.e2 .!l)g6 13 0-0 b6 14 b4 bxcS 15 bxc5 .i.f6 16.i.d6 17 e4 ± Gheorghiu-Westerinen, Palma de Mallorca 1965. 3) 9 ...dxc4 10 .i.xc4 bS 11 .i.d3 "b6 12 e4 g6 13 eS ~ 14 .!l)xdS cxd5 15 h4 ± Gheorghiu-J.Kristinsson, Reykjavik 1972.

"as

Systems with ... c6 4) 9 .....a5 and now: 4a) 10 ~d2 cS 11 ~b3 "b6 12 cxdS ~xdS 13 lLlxdS exdS 14 i.e2 cxd4 15 ~xd4 "a5+ 16 ~fl ao Korotylev-Deviatkin, Moscow 1996. At last we see a position with some life in it for Black. 4b) 10 cS!? (a more positionally 11mbitious try for White) 10...~e4 (1O... b6 11 cxb6) 11 "a4 i.d8 12 "xa5! (after 12 b4?!, 12... ~xc3?! 13 'irb3! is good for White, but 12.....xa4 13lLlxa4 bS! favours Black) 12 ...i.xaS 13 J:[c 1 ;to If Black tries to break White's spatial advantage by 13 ... b6, White still keeps the initiative with 14 cxb6 ~xc3 15 bxc3 i.xb6 16 i.d3. 10cS 10 cxd5!?; 10 cxbS!? 1O•••aS 11 i.d3 b4 12ltla4 i.a6 13 .i.xa6 :xa6 14 0-0 15 lOel i.d8 16 ~3 i.e717 a3 bxa318 bxa3 J:[e8 White has a definite advantage in space, but needs to find some way to make it tell. White took care to keep cS covered in Bagirov-A.Petrosian, Riga 1981 with 19 "b2 i.xf4 20 exf4 "c7 21 J:[bl J:[b8 22 'ii'c2 ±. In a later game, White used a more tactical approach: 19 i.xc7 'fIxc7 20 J:[bl eS 21 J:[b6! J:[aa8 (21...~xb6 22 cxb6 J:[xb6 23 ~xb6 'fIxb6 24 l:[bl ±; attempts by Black to keep the exchange lead to catastrophe) 22 'ii'b2 exd4 23 exd4 J:[e4 24 J:[b7 'fIc8 25 J:[bl ± Kramnik-Zude, Bundesliga 1994/5. But was Black's plan worth repeating in the first place?

"cS

A4)

8...b6 (D) Encouraged by White's loss of time with h3, Black tries to fit in the fianchetto without ill effect.

203

gexdS The natural response, trying to fix a weakness on c6. 9 i.d3 dxc4 10 i.xc4 00 11 i.h2 (11 O-O!?; 11 ~xdS!?) 11.. ..i.b7 12 0-0 lLlxc3 13 bxc3 c5 =Nazin-Pirozhkov, Kaluga 1996. 9 J:[dl!? i.a6 10 ~S "c8 11 i.e2 ~xeS 12 dxeS ~7 13 0-0 bS 14 cxbS cxbS 15 'ii'd3 ;t Kacheishvili-Dzhandzhgava, Georgian Ch 1996. 9•••cxdS 9 ...exdS?! is to be regarded with some suspicion. At best, Black will be able to reach a hanging pawn position a tempo down after ... cS. For example: 1) 10 i.e2 .i.b7 11 0-0 'fIc8 12 J:[acl i.a6 13 'fIa4 i.xe2 14 ~e2 cS 15 dxcS bxcS 16 J:[fdl "b7 17 'fIc2;t Minzer-Gamarian, Boca 1997. 2) 10 .i.d3 i.b7 11 ~S J:[e8 12 0-0 ~f8 is another way of playing it, treating ...c6 as a loss of tempo in a ... b6 system. 13 'fIa4!? bS 14 'fIdl ;t Muse-Fuhrmann, Bundesliga 1996n. 10 i.d3 i.b7 11 0-0 a6 12 a4 h6 13 i.h2 ~b8 14 ~eS lLle6 15 ~xe6 .i.xe6 16 J:[rct .i.d6 17 i.xd6 "xd6 18 ~b5 i.xb5 19 axb5 axb5 2O:xas J:[xa8 Arduman-Vaganian, Erevan OL 1996.

=

204

The Queen's Gambit Declined: 5 ~f4!

This is fine for Black if you are happy playing the more deadly dull forms of the Exchange Slav. White for his own part could play more aggressively at various stages, for example 13 'ife2!?

A5) 8 ••••aS(D)

We have just a couple of examples of this move, with Black scoring 100% after White tried too hard to refute the queen sortie. 9 lM2 :e8 10 J-d3 J-b4 11 adS After 11 0-0 e5 12 J-g3 e4 13 J-e2 a6 14 a3 J-fS, Karlsson-Hf1ji, Copenhagen 1985 continued 15 b4?! 'ifd816 c5? g6 17 a4 h5, when Black's play on the kingside proved stronger than anything White could muster on the opposite side. However, closing the queenside is entirely anti-positional for White; he needs to expose weaknesses, not conceal them. Thus 15 cxdS cxd5 16 ~b3 'ifd8 17 ~a4 is good for White. 11•••cxdS 12 0-0 eS 13 ~b3 "'b6 14 dxeS ~eS IS i.xb7+!? White's advantage is much more slender than it looks, bearing in mind

such possibilities as 15 i.g5ltJxd3 16 'ifxd3 J-xc3 17 'ifxc3?? ltJe4. Steady play such as 15 :fd 1 !? should give something though. IS...ltJxb716ltJxdS "'d617 ~c7? Burgess makes the interesting suggestion of 17 :fdl I?, keeping a few threats hanging over Black. If then 17...b618ltJxb4 'ifxb419 J-xe5 J-xh3 (19 ....:xe5? 20 :d8+), White continues 20 J-d6 'ifg4 21 J-g3! ltJg5 (after 2l...'iff5 or2l...'ifc8 White has a clear extra pawn in the ending) 22 ltJd4 :ad8 23 gxh3 'ifxh3 24 'ife2 and the attack collapses. 17...J-fS 18ltJxe8 Maybe in Tavadian-Anastasian, Erevan 1982, White's calculations had stopped here. However, after 18......g6! 19 .c7 J-e4 20 f3 (20 J-g3 :xe8!) 20...~xf3+ 21 ~hl ~4 22:n :xe8 Black was winning. Finally a couple of very minor alternatives, although one of these was able to knock over Kramnik in a quickplay.

A6) 8...g61t Not very pretty, so no diagram. 9J-d3 9 :dl!? 9...dxc4 10 i.xc4 bS 11 J-e2 i.b7 12 0-0 a6 13 :fdl "'b6 14 a4 b4 IS as "'a7 16ltJa4 cS 17 ltJeS ;t PovahMerriman, England 1996. There is no immediate refutation, though Black could do without the kingside holes. A7)

8...b6

Systems with ... c6 When one considers that the theoretical recommendation after 1 d4 d5 2 c4 e6 3 lDc3 lDf6 4 cxd5 exd5 5 J.g5 i.e7 6 e3 lDbd7 7 J.d3 h6 is 8 J.f4 ;1;, the wisdom of 7...h6 here is readily doubted. 9l:r.dl 9 cxd5!?; 9 O-O-O!?; 9 g4!? In a game at normal speed, White would have time to think. 9... b610 J.e2 10 cxd5!? 10••• J.a6 11 b3 'irc8 12 0-0 bS 13 cxbS? Walking into a self-made pin. 13 cS might yet give something to play for. Now Kramnik-Ubilava, Oviedo rpd 1992 continued 13 ...cxb5 14 "d3 b4 15 "xa6 bxc3 16 'ira5lDe417 Ilcl a6 18 ':c2 "b7 19 Ilal Ilfc8 20 lDel J.b4 21 "a4 ~5 0-1. The theory on 7 'irc2lDbd7 8 h3 is as yet not very systematic. After 8... lle8, the most popular, the recommendation is 9 0-0-0. 8...dxc4 and 8.....aS do not seem quite sufficient to equalize.

B, 8 Ildl (D)

205

Yet again there is a frustrating lack of consensus as to what Black should play next. White plays a natural developing move instead of a pawn-push, but the move has two potential defects. First, no retreat is provided for the bishop if Black plays 8 ... lDh5. Secondly, if Black can create immediate play on the a5-el diagonal (after 8.....aS), there is a danger that White will have to play ':c 1 at some stage to cover c3. We cover the two aggressive moves first. Our lines are: Bl: 8...lDhS 205 B2: 8....aS 206 B3: 8...1le8 206 B4: 8...a6 207 BS:8•••b6 207 B6: 8...dxc4 207 B1) 8...lDhS 9 .td3!? Aiming for a Rubinstein Bind. The straightforward 9 .tg3 is also possible, and if 9 ...lDxg3?!, 10 hxg3 g6 11 i.d3 dxc4 12 .txc4 b5 13 .td3 14 cRe2.ta6 15 ~ ± Matras-Jirman, Klatovy 1995. Instead, 9 ... f5!? gives good chances of equalizing. 9...lDxf4 10 exf4 g6 After 10... lDf6 11 0-0 dxc4 12 .txc4 lDd5 13 83 86 (13 ... f6!?) 14 .!Lle4lDf6 lSlDe5 (1SlDeg5!?; 15 h4 ;1;) 15 ... lDxe4 16 "xe4 i.f6 17 h4 "c7, White can keep up the pressure with 18 h5;1;, instead of 18 Ilfe1 b619 lDxc6 .tb7 20 d5 .txb2 = MatthiasSchlindwein, Lippstadt 1996. 11 h4 dxc4 12 .txc4 lDb6 13 i.d3?! This is a slightly unfortunate square for the bishop. 13 .tb3! looks very strong, and if 13 ...~, 14 g3 ±.

.as

The Queen's Gambit Declined: 5 iLj4!

206

13•••lbd5 14lbxd5 14 g3?! ~b4 reveals the point behind Black's last move. 14•••exd5 15 b5 i.f6 16 i.g4 17 hxg6 hxg6 18 .i.xg6 .i.xf3 19 .i.h7+ ~g7 20 IId3 IIhS 21 IId3 i.xd4 White's attack has burnt itself out, Van Wely-Radulov, Khania 1993.

wn

=

One can conclude very little from this, except that 8.....aS is a valid try. No real significance should be attached to the statistic, 8 h3 "as, on for White; 8 IIdl "as, 3/3 for White. If anything, 8 .....a5 makes more sense after 8 IIdl.

83) S•••.:r.eS (D)

82)

S.....a5 Here we have three games, and three different moves, even against one of Black's most direct approaches. It is rare to get an opening system where there is such lack of consensus. 1) 9 ~d2 i.b4 10 i.d3 i.xc3 (this can wait for a move; 1O...cS! GO) 11 bxc3 dxc4? (a disastrous surrender of the centre; 11 ...cS 12 ~b3 ~) 12 ~xc4 ~S 13 h3 ~b6 14 ~S +- Buki~­ Mali, BledIRog~ka Slatina 1992. 2) 9 a3 ~410 i.d3 ~df61l 0-0 ~c3 12 bxc3 dxc4 13 i.xc4ltX1S 14 IIcl ~xf4 IS exf4, and in M.Gurevich-Weemaes, Antwerp 1994, Black snatched the pawn with IS .....xa3?! 16 IIfel "d617 ~S .i.f618 :Cdl g6 19 .:r.e3 i.g7 20 h4 and White had a bind. It is more important for Black to free his queenside: IS ...cS GO. 3) 9cSwasmetby9 ... ~41Oi.d3 fS 11 IIbl "d8 12 h3 "e8 13 b4 i.f6 14 ~e2 a6 IS "b2 g6 16 g4 gS 17 i.h2"f7 18 IIgl fxg4 19 IIxg4 i.g7 20 tilegl h6 21 i.xe4 dxe4 22 ~eS ± in Najer-Notkin, Moscow 1995. However, Black doesn't have to playa Stonewall; 9 ... bS!? and 9 ... ~h5!? are both possible, while 9 ... b6!? might be best of all, despite the worrying lack of flight squares for the queen.

The main line here is 9 h3, transposing to Line AI6 above, though White can try a different waiting move: 9a3!?M The standard manoeuvre, preparing ... i.d6,etc. Alternatives tend to be gUilty of mixing plans. I) 9 .....aS (a strange choice with ... i.b4 unavailable) 10 1tX12 i.f8 11 i.g3 eS 12 ~b31t'd8 13 cxd5 cxdS 14 i.e2 e4 15 ~b5 l:e6 16 IIcl ~ P.Neumann-Vorisek, Cesky Brod 1995. 2) 9... aS 10 c5 b6 11 ~S?! (11 cxb6 ~) 11...~xe5 12 dxe5 ~d7 13 cxb6 tilxb614 .i.d3 g61S h4 i.a6 was unclear in the game KlingelhoeferLabahn, Dortmund 1993. 3) 9 ... a6 transposes to 8 ... a6 9 a3 lIe8; see Line B4 below. 10.i.d3 ~g6

Systems with ... c6 1O...dxc4 11 .i.xc4 l2Xis 12 0-0 lLlxf4 13 exf4lLlg6 14 g3 b6 15 h4 .i.f6 16 h5 lLle7 17 lLle5 ± Kacheishviii-H. Schneider, Schwabisch Gmiind 1994. 11 .i.g3 .i.d6 12 0-0 .i.xg3 13 bxg3 "e7 14 l:td2 a6 IS e4 dxe4 16 .i.xe4 eS 17 lLlgS 18 dS ± Nepomnishay-Asrian, St Petersburg 1996.

:rs

84) 8...a6(D)

This has two possible objectives: either to play an early ... b5, or to develop in Meran fashion with ...dxc4, ... bS, etc. 9&3 I) 9 h3!? transposes to the Bag-

irov-A.Petrosian and Kramnik-Zude games, discussed in Line A3 above. 2) 9 .i.d3 dxc4 10 .i.xc4 b5 11 .i.e2 c5?! (l1.. ...b6 =; White lacks the tempo to block the queenside with b4) 12 dxcS .i.xc5 13 0-0 'iWb6 14lLlg5 h6 15 lLlge4 ;t Meulders-G.Winkler, Brussels 1993. 3) 9 .i.e2 b5 and now: 3a) 10 cS?! (thi,~.~aves the position too blocked) 1O...h4 (1O...aS! Gelfand) 11 1&4 as 12 lLleS ~xeS 13

207

~b6! l:ta7 14 ~xcS "xcS IS .i.xe5 ;t Gelfand-King, Haifa Echt 19S9 (the move-order in this game was 6 ... a6 7 "c2 ~bd7 S l:tdl c6). 3b) Gelfand suggests 10 lLle5, giving a slight edge after 1O... .i.b7 11 c5, 10 ... bxc4 11 ~xc6 "eS 12 b3 or 1O... ~xe5 11 dxeSl2Xi7 12 cxd5 cxd5 13.i.d3. 9•••dxe4 9 ...lIeS 10 h3 dxc4 11 .i.xc4 bS 12 .i.a2 .i.b7, and now 13 ~e5 looks slightly better for White. Instead, White over-pressed in Kotov-Grechkin, Leningrad 1938, and after 130-0 'iWb6 14 e4 cS IS dS?! exdS 16 eS d4 17 ~gS c4! 18 exf6 lLlxf6 White, though a piece up, was soon suffocated by Black's advancing pawns. 10 .i.xe4 bS 11 .i.d3 .i.b7 12 b4 as 13 l:tbl l:te8 14 e4 ;t Psakhis-King, London Lloyds Bank 1994.

B5) 8•••b6 Not a move one would expect to be successful. 9cxdS The thematic reply, rather than 9 .i.e2?! .i.a6 10 b3 :CS 11 0-0 ~ 12 .i.eS ~hf6 13 .i.g3 = Avrukh-Hao Yin, Cala Galdana U-18 1996. 9...exdS 10 .i.d3 l:te811 0-0 lLlf8?! The disease of mixing plans. What does Black do about his c6-pawn? 12lLleS .i.b713.i.g3 :c814 .i.fS ± A.Hoffman-Aramayo, Salta 1995. B6) 8 •••dxe4 is not very thematic either. 9 .i.xc4 ~b6 10 .i.e2 ~bd5 l1.i.g3 h6 12 e4 ;t Pogorelov-Ruberg, Munster 1991.

208

The Queen's Gambit Declined: 5 ~f4!

C) S cxdslOxdS (D) We look at the alternative capture 8 ...exd5 under the move-order 7 cxd5 exd5.

w

The knight recapture is known in positions where White has not played .c2, but is all the more tempting after Wc2 (instead of a bishop move) in that White's king is still two moves away from castling, providing chances of play along the a5-el diagonal. 9lOxdS After 9 J.g3 .a5 10 e4 lOxc3 II bxc3, 11...:e8?! 12 J.d3 h6 13 0-0 lOf6 14 J.f4 ;t was H.Miiller-Kajumova, Dresden 1993, but Black has livelier defensive attempts available, such as ll...c5. 9 •••exdS 10 J.d3 White can try 10 a3!1, and if Black replies 1O...•a5+, then 11 Wd2. 10•••J.b4+!? 1O...lOf6 II 0-0 J.g4 1h_ 111 K.BergHf2Ii, Danish Ch (Ringsted) 1995, although White looks a little better. 11 ~e2lOr6 12lOeS 13 g4!1 With ideas of a full-blooded attack. Even so, as with the New Main Line, it would seem to be more prudent to play

13 h4 first, since as played the g-pawn itself becomes a target. 13 J.g5 h6 14 h4 J.d6 ac. 13•••g6?! Unnecessarily weakening his pawn structure. Forintos gives two better defences: 1) 13 ... J.xg4+ 14 lOxg4lOxg4 15 J.xh7+ ~f8 ac. 2) 13 ...c5 14 a3 (14 dxc5 lOe4 15 f3 .h4! 16 fxe4 :xe5 00; 14 gSlOe415 J.xe4 dxe4 16 .xe4 J.e6 00) 14... c4 15 J.xh7+ lOxh7 16 axb4 .d6! 17 .c3 J.xg4+ 18 lOxg4 .xf4 00. 14 :agl J.d6 15 h4 cS 16 h5 White's attack now broke through spectacularly in Vaganian-Forintos, Moscow 1975: 16...cxd4? 17 hxg6 J.xe5 18 g5! J.g4+ 19 :xg4 :c8 20 gxh7+ ~h8 21 J.xe5 :xe5 22 gxf6 :xc2+ 23 J.xc2 d3+ 24 J.xd3 :e8 25 :hgll-O. Forintos suggests as an improvement 16... c4 17 J.xg6 J.xe5, continuing 18 dxe5 d4, but 18 J.xe5 looks strong.

D) SJ.d3 (D)

:e8

The most natural move on the board, aiming to complete development in

209

Systems with ... c6 the simplest fashion possible. The one snag is, of course, the loss of tempo if Black exchanges on c4. 8•••dxc4 There is no particular sense in delaying with S...:eS. Then 9 0-0 dxc4 10 i.xc4 ~dS?? 11 i.xdS exdS 12 ~xdS, winning a pawn, Goldin-Lamoureux, Paris 1994, is just one of several versions of an old trap. 9 lid 1 is another possibility, as after 9... dxc4 10 i.xc4 LDb6 11 i.b3 ~bd5 12 i.eS WaS 13 0-0 ;t Csaba-Vajani, Hungary 1993 (reached via S l:.dl lieS 9 i.d3), White's interpolated move is more useful than Black's. 9 i.xc4 b5 The 'Semi-Slav' treatment is the most effective. Alternatives: 1) 9... ~dS?! leaves White ahead of the S cxdS ~xdS variation. 10 i.g3 ~7f6 11 a3 (unnecessary; 11 0-0 ;t) 11 ... ~xc3 12 bxc3 WaS 13 0-0 ~h5?! (I3 ... cS=) 14i.eSg61Sh3~g716e4 ! Tartakower-Yates, Baden-Baden 1925. This suggests that Marovic's claim, in Play the Queen's Gambit, that the old masters had a better feel for this opening than the moderns, is to be treated with some scepticism. 2) 9...~b6IOi.e2(1Oi.d3!?~bdS II ~xdS exd5 120-0 leaves White ahead of the S cxdS ~xdS 9 ~xdS cxdS line) 1O... ~bdS l1i.g3 ~xc3 12 bxc3 cS 13 ~eS (13 0-0 ;t) 13 ... cxd4 14 exd4 ~d7 15 i.d3 h6 16 0-0 ~xe5 17 i.xe5 i.d7 IS i.h7+ ~hS 19 i.e4 i.c6 = Nedela-P.Bazant, Czech Ch 1996. 3) 9 ... a6 10 0-0 cS tt-dxc5 i.xc5 transposes to the 6 ... c5, S i.e2 variation (Chapter 4.4, line A), except that in that variation White usually plays

his queen to e2 rather than to the exposed c2-square; Kholmov-Dmitrievsky, Moscow 1966 continued 12 lIfdl We713i.d3(13a3!?) 13 ...h614a3e5 15 i.g3 lieS 16 i.f5 i.d6 17 i.h4 WfS IS lId2 ~cS! 19 :adli.xf5 20 Wxf5 lIe6 21 i.xf6 lIxf6 22 Wbl b5 23 b4 ~b7 24 ~e4 lIe6 25 ~xd6 lIxd6 26 lIxd6 ~xd6 27 ~xe5 We7, when White's extra pawn proved insufficient to force the win against active play. 4) 9 ... WaS 10 0-0 ~b6 11 i.e2 (11 i.d3!? ~bd5 12 ~xd5 ;t) 11...~bd5 12i.g5 Wc713 l:.ac1 h6?! (13 ... ~xc3 ;t) 14 i.h4 ~g4? 15 i.g3 i.d616 ~bS ± O.Stefan-J.Groch, Slovakia 1994. 10 i.d3 i.b7 11 0-0 1) 11...h6 (too passive?) 12 lIfdl a6 13 a4 lIcS 14 We2 ~b6 15 e4 lieS 16 h3 i.b4 17 ~a2 i.fS IS as ~bd7 19 b4 ;t Mikhalchishin-Lyrberg, Bern 1995. 2) 11...b4 12 ~e4 ~xe4 13 i.xe4 ;t Korchnoi-Doroshkevich, USSR Ch 1970. 12 ~g5 h6 13 ~ge4 b4 14 ~a4 ~xe4 15 i.xe4 i.a6 16 lIfc1?! 1611fdl =. 16•••eS! 17 ~c5 ~c5 18 dxc5 i.xc5 19 .xeS lIxeS 20 lIxc5 +Galliamova-Kakhiani, Borzhomi wom Ct 1990. Black now contented herself with a draw after 20 ... i.d3?! 21 lIdl i.xe4, but there are several ways of playing for a win, for example 20... f6!? followed by ...e5. If Black plays accurately, White's loss of tempo makes it difficult to keep any ftrm grip on the position.

:c8

E) 8 i.e2(D)

210

The Queen's Gambit Declined: 5 j.f4!

B

The simplest way is 8 ...dxc4, transposing to lines we have just looked at. Black tried to vary, with no great success, in Touzane-Laclau, French Cht 1993: 8 ...~h5?! 9 ~g3 f5 100-0-0 a6 11 ~bl g6 12 ~2 ~xg3 13 hxg3 ~f6 14 g4!? (14 c5!?) 14... ~xg4 15 ~xg4 fxg416 f4 with a positional grip that more than compensated for the sacrificed pawn. F)

8 O-O-O!? (D) Not the sort of move to play if you want a reputation for subtlety! I would suspect that 8 h3 :e8 9 0-0-0 is a slightly more accurate rendering of the plan. Naturally, in either case White is after the black king's throat.

8•••dxc4 Black must urgently start action on the queenside. 8 ...h6? 9 h4 dxc4 10 ~xc4 ~d5 11 g4 ~xf4 12 exf4 .c7 13 ~2 ~d6 14 f5 ~b6 15 ~b3 exf5 16 gxf5 ~d5 17 :dgl ~d7 18 f6 ~f4+ 19 ~bl ~xf6 20 .g6 ~8 21 ~g5 1-0 Doeserich-Krebs, Badenweiler 1995 was a rottweilering in Badenweiler. 8 ....a5?! 9 ~bl?! (unnecessary; 9 ~2 ;1;) 9 ... ~b4 10 ~2 ~ 11 ~g3 f5 12 ~d3 ~xg3 13 hxg3lDffi 14 ~2? (14 f3!?) 14...~xd2 15 :xd2 dxc4 16 ~xc4 ~4 and Black should win, Pein-Martinovsky, Wrexham 1995. Naturally, White's lacklustre play can be improved considerably. 9~xc4bS

The most thematic plan, though 9 ....aS could also be considered. 9 ... ~b6 is perhaps a little inflexible after 10 ~d3. However, Valdes-Riviera, Cuban Ch 1995 continued 10 ~b3?! (the bishop is now a target for Black's advancing pawns) 1O... aS II h4 (11 a3!?) 11 ... a4! 12 ~xa4 ~xa4 13 ~xa4 ~d5 14 ~g5 g6 15 ~b3 ~b416 jlb1 e5! 17 ~xe5 ~f5 18 e4 ~g419:d2c520f3 c421 ~xc4:c8 22 b3 b5 23 fxg4 bxc4 24 bxc4 :xc4+ 25 ~dl .c8 and White was unable to withstand the three-pawn gambit.

10 ~d3 B

.as

11 ~d2

Taking the knight from the kingside is already a minor concession, but as we have seen while examining the New Main Line, White can often follow queenside castling by central play. The point of this move is to discourage ... h4. 11 ~bl ~b7?! (routine play may lead to quick disaster in this type of

Systems with ... c6

211

sharp position; it makes more sense to play 11 ... b4 followed by ...i.a6, taking some of the pressure off h7) 12 ~5 l:lfdS 13 ~?! (13 g4! is much more effective) 13 ...lLId5 14 ~g5lDxe5 15 i.xh7+
Kain-G.Kis, Hungary 1995, White should play 12 i.e5 ±. The queen doesn't really belong on as in this type of Stonewall position. 2) S... as?! misreads White's intentions. In D.McMahon-Tuddenham, Golders Green 1995, the old trap on the c-file was sprung after 9 cxd5 exd5 10 i.d3 l:leS? 11 ~b5. 3) S...h6 9 h3 dxc4 10.i.xc4 b5 11 i.e2 i.b7 120-0 a6 13 l:lfdl .b6 14 a4 l:lacS 15 ~5 l:lfdS 16 .b3 b4 17 as "'a7 IS ~xd7 ~xd7 19 ~e4 and White is taking control, Tisdall-Lyrberg, Oslo 1994. 4) S...a69cxd5(9c5!?c!llli5 1Oi.d3 ~xf4 11 exf4 h6 120-0 i.f6 13 b4 ± Ortega-Giretti, Verona 1997) 9 ...exd5 10 i.d3 l:leS 11 h3 ~fS 12 0-0 ~g6 13 .i.h2.i.d6 14 .i.xd6 .xd6 15 b4 i.e6

G)

Efimov-Barletta, Montecatini Terme 1995. 5) S...l:leS will transpose into lines already considered after 9 h3 or 9 :dl.

"'dB

16~a4~717~5~xc518bxc5;t

Sa3 (D)

H

Another tempo move, leaving yet again the usual bewildering array of choice. 8••• ~bS "I) S....a5 9 ~2 ~h5 10 .i.g3 f5 II i.e2 ~df6, and now instead of playing 12 O-O? ~xg3 13 bxg3 i.d7 ~,

9.teS 9 .td3 ~xf4 10 .txh7+
212

The Queen's Gambit Declined: 5 ~f4!

Gurevich gives 11...l£lxeS 12 dxeS, with the idea of .i.e2, 0-0, etc., as good for White, but one can be sceptical of this assessment. After 12 ... gS, with the idea of ... l£lg7, and possibly ...J.d7e8, Black looks comfortable enough. White's pieces are not working effectively in the centre, which allows Black to take certain liberties with his kingside pawn structure in order to keep his pieces active. 12 J.f4! 1£le4 12 ...111hS 13 J.d3l£lxf4 14 exf4 ± Gurevich. 13 h3 llldf6 14 J.d3 J.d7 15 0·0 J.e816J.h2 Gurevich notes that 16 llleS followed by f3 is also good. 16•••J.hS 17 llle5I£ld7 18 f3111xe5 19 J.xe5l£lxc3 20 bxc3 J.d6 21 J.xd6 'it'xd6 22 c5 'it'c7 23 c4 ;t M.Gurevich-An.Fernandes, Linares 1995. After going through seven alternatives, it is possible to conclude that 8 h3 is at least as promising as anything else in the 'it'c2 system. Before leaving 7 'it'c2, we must note that Black too has seventh-move alternatives.

7.27 'ii'c2: Black delays ... ~bd7 or ... 0-0 1 d4 d5 2 c4 e6 3lllc3 ~ 4 00 .i.e7 5 J.f4 c6 6 e3 lllbd7 (or 6...0·0) 7 'it'c2 Our coverage falls naturally into two sections: A: 6 •••111bd7 7 'it'c2 with •••0-0 delayed 212 B: 6 •••0·0 7 'it'c2 with •••lllbd7 delayed/omitted 213

A) Black delays castling 6...111bd7 7 'it'c2 (D)

7•••111hS 1) 7... h6?! (pointless) 8 lIdl111hS 9 J.d3 lllxf4 10 exf4 lllf6 11 0-0 0-0 121£leS with a distinct advantage for White, Mikhalchishin-Siska, Ljubljana 1994. 2) 7 ...'it'aS?! 8111d2 0-0 and now: 2a) Following 9 0-0-0, Black could find nothing better than 9 ...'it'dS?! in Delemarre-Andrade, Guarapuava U-lS Wch 1995. This is undoubtedly an exaggerated reaction, but the queen is not well placed on as. 2b) 9 .td3 lIe8 10 0-0 J.f8 11 J.g3 g6 12 lIfel a6 13 e4 ± ChenauxAstengo, Geneva 1996. 8 J.g3 1£lxg3 It is usually a positional error to make this capture before White has castled kingside. Kramnik suggests 8 ... g6, S... h6, or simply 8 ...0-0, as possibilities worth exploring. 9hxg3g6 9... h6 10 a3;t Kramnik. 10J.d3 Kramnik notes that 100-0-0 'it'as 11 ~bl is also promising. 10•••0-0?

Systems with ... c6 1O... a6; 1O...dxc4 11 .txc4 b5 12 .td3 .tb7;t Kramnik. 11 0-0-0 dxc4 12 .txc4 bS 13 .td3 "as 14 ~bl .ta6 15 1Oe4 ± Krarnnik-Abrarnovic, Biel IZ 1993. In this game it was not long before White broke through on the h-file. B) Black avoids ...lObd7 6 ...0·07 Wc2 (D)

cxb6;t) 1O..•~h8? (lO...a5 11 b5.tb7! 00) 11 0-0 lOh5 12 :fel ;t Den BerBossuyt, Antwerp 1992. lc) Other 8th move replies by White, such as 8 :dl or 8 h3, will tend to transpose to lines already considered. 2) 7...dxc4? has been played, but it is a simple loss of tempo. 3) 7...h6?! S .td3 (8 h3!?;t) 8...dxc4 9 .txc4 b5 10 .td3 .tb7 11 0-0 a6 12 h31Obd7 13tOe41Oxe4 14 .txe4"b6 Belarnaric-Paraminski, MakarskaTucepi 1995. 4) 7 .....a5?! 8 lDd21Obd7 9 i.d3 :leS 10 0-0 1Of8 11 a3 ;t J .AndersenHellegard, Aalborg 1995. 5) 7...lOa6 and now: 5a) 8 a3 dxc4 (8 ...1Dc7? 9 c5 .td7 10 i.d3 b6 11 b4 and Black is getting squashed, Bobotsov-Shafie, Teheran 1991) 9 .txc4 b5 10 .te2 b4 11 axb4 lOxb4 12 "d2 c5 112-112 KorneevetsBegun, Belarussian Ch (Minsk) 1996. Just occasionally, 'Russian draws' reveal something interesting. 5b) 8 c5 is probably best, and if S...b6 (8 ...1Ob4 9 Wb3; 8 ... b5 91Oe5), 9 Wa4 (9 lOe5? 1Ob4! followed by ... dxc5) 9 ... b5 101Wdl ;t; the knight is misplaced. 8 i.d3 (8 h3 I?) S...dxc4 9 6) 7... .txc41Obd7 100-0 lOb6 11.td31ObdS 12 .tg3 h6 13 a3 lOb5 00 M.VargaSlavk.Marjanovic, Nova Gorica 1997. 8 .td3! (D) 1) Sa3?!i.b79cxd5exd5lOi.d3 c5 111Ob5? 1Dc6 12 .tf5? g6 13 .td3, Em.Lasker-Marco, Hastings 1895, and now Tarrasch gives 13 ... c4 +. The opening never was Lasker's strong point, even by the standafds of a hundred years ago.

=

7 .••b6 While not particularly thematic, this has been tried a few times. Black's thinking is that once White has played 'iic2, the plan of cxd5, tOe5 and "f3, which is so dangerous in the 6 ...b6 line, is no longer available. Others: 1) 7 ... a6 and now: la) 8 cxd5 exd5 9 .td3 :e8 transposes into the ...c6 Exchange Variation, where Black's ...a6 is not looking very relevant. 10 0-0-0 is worth considering, while Doroshkevich-Lobach, Simferopol 1989 continued 10 0-0 lObd7 11 :ablM 12 b4lOb5 13 .te5 f6 14 .tg3 lOxg) 15 hxg3, which is perhaps about equal. Ib) 8 c5!? is also possible, for example 8 ...lObd7 9.td3 b6 10 b4?! (10

213

:e8

The Queen's Gambit Declined: 5 i.f4!

214

2) 8 cxdS cxdS! 9 .te2 .tb7 10 0-0 lObd7 11 h3 %tc8 12 'iWb3 lOe4 13 %tacl a6 Lobron-Beliavsky, Novi Sad OL 1990 is an Exchange Slav with White's queen misplaced. 3) 8 lOes .tb7 9 cxdS, and now 9 ...cxdS is probably safest. 9...lOxdS is tempting, but after 10 lOxdS 'iWxdS 11 .td3! "a5+ 12 ~e2 .ta6 13 lOc4 'iWhS+ 14 f3, Muse-Pfleger, Bundesliga 1993/4, White probably has the better of a messy position.

=

With unclear play, Van der SterrenBrestian, Hilversum 1993. However, White's improvement on move 11 would have kept Black under more pressure. We now leave the 7 'ifc2 system to look at the pawn exchange on dS. First though we make a detour, to consider the 4 ... ~bd7 variation, which often transposes.

7.3 4 ~f3 ~bd7: B

Transpositions and independent lines After 1 d4 d5 2 c4 e6 3 ~c3 ~r6 4 ~f3 ~bd7 the following position is reached (D):

w 8•••.tb7 9 0-0 lObd7 10 cxd5! Good timing. He waits until Black has played ...lObd7 before exchanging on dS. If now 1O..•cxdS White is better as Black cannot play ... lOe4; usually it is only the possibility of creating this knight outpost that justifies ... lObd7 rather than ...1Oc6 in the Exchange Slav. 10...exd5 11 ~e5?! This is a little impatient. 11 %tfdl ~ makes it very difficult for Black to make any freeing pawn move; White can then build up in relative leisure. 1l...h6 12 h3 lOxe5 13 dxe5 1Oci7 14 %tfdl l:r.e8 15 e4 d4 16 .tc4 .tcS! Avoiding 16...dxc3? 17 e6. 17 b4 .txb4 18 l:r.xd4 .tcS 19 l:r.d2 .tb4

It would be stretching the bounds of this book to attempt a full survey of this system, but players with White should beware that S .tf4?! dxc4 is reputed to be fine for Black: 6 e4?! .tb4 7 'ifc2 bS 8 eS (8 a4?! .i.b7 9 eS .txf3 Sokolov) 8 ... ~dS 9 .i.gS f6 ~ Van Wely-I.Sokolov, Tilburg 1994; or 6 e3 ~b6!? (6 ... ~dS 7 .txc4 ~xf4 8 exf4 .i.e7 9 dS? ~b6 10 .tbS+ ~f8 ~

Systems with ... c6 Portisch-R.Byrne, Biel IZ 1976, but Byrne and Mednis note that White keeps his edge after 9 'iVc2) 7 .i.xc4 ~xc4 8 'iVa4+ c6 9 'iVxc4 and now Byrne and Mednis give 9 ... ~d5 as equal. 5 "'c2 has also been tried, with Kasparov-Seirawan, Amsterdam 1996 continuing 5 ... dxc4 6 e4 c5 7 dxc5 (7 d5 exd5 and now Kasparov gives 8 exd5 as unclear, and 8 e5 ~xe5! 9 ~xe5 .i.d6 10 "'a4+ with unresolved complications) 7 ... .i.xc5 8 .i.xc4 a6 9 a4 "fIc7 10 0-0 ~g4 11 h3 ~xf2 (or 1l...~ge5 12 ~xe5 ~xe5 13 .i.e2 0-0 Kasparov) 12 :lxf2 .i.xf2+ 13 "'xf2 "fIxc4 14 Wg3 f6 15 "'xg7 Wc5+ 16 ~hl "'fS 17 "'g4 and White won the game after 17......f7 18 e5 :lg8 19 "c4 f5 20 .i.g5 ±, though I'll have to rub my eyes a few more times before I can believe that White has enough for the exchange after 17......g8. Earlier, White could vary with, for example, 9 a3, but it is far from clear that this brings anything much. What this is all leading up to is that White could consider entering a Classical Exchange Variation with .i.f4 instead of .i.g5. After 5 cxd5 exd5, White has played 6 .i.f4 several times, but this leaves Black the option of developing more actively with 6 ....i.b4 rather than 6 ....i.e7. White can of course play 6 .i.g5 instead, entering the main lines of the .i.g5 Exchange Variation, but if he wants to stick with the .i.f4 Variation, then 6 'iVc2!? c6 7 .i.f4 (D) could be considered. 7 ....i.e7 8 e3 0-"U9-.i.d3 leads to standard positions which will be discussed shortly in Section 7.4. In recent grandmaster play, Black has attempted

=

215

B

to inject a bit more life into the position by playing 7 ... lDb6 or 7 ... lDh5. The following key references come from recent Informators (D35) rather than from computer databases; thus, partly for reasons of space, a full survey is not given. Our coverage divides as follows: A: 7 ••• lDb6 215 B: 7•••lDh5 216

A) 7..•lDb6 Black intends to offer the exchange of dark-squared bishops without wasting a tempo on ... .i.e7, and without playing such elaborate manoeuvres as ...0-0, ...:le8, ...lDfS, ....i.d6. Even so, the knight is not well placed on b6, which gives White a bit of time to catch up later. 8e3 8 h3 .i.d6 9 .i.xd6 "'xd6 10 e3 'iVe7 (1O....i.e6 11 .i.d3 lDbd7 12 O-O-O!? 0-0-0 13 ~bl ~b8 14 ~al ! :lhe8 15 :lbl :lc8 16 :lhcllDb6 17 b4;!; 1\1kmakov-l.Ivanov, Ashkhabad 1978) 11 U3lDc4?! (Smagin suggests 1l ....i.e6 intending ...lDc8-d6 as equal, but maybe White can try 120-0-0 again) 12 .i.xc4 dxc4 130-00-0 14 e4 b5 15

The Queen's Gambit Declined: 5 Af4!

216

':fel ~ Epishin-Smagin, Russian Ch (Elista) 1995. 8••.£g4 9 ~ .*.h5 10 h3 10 .*.d3 .*.e7 II.*.g5 .*.g6 12.*.xg6 hxg6 13 0-0 lDfd7 14 .*.xe7 Wxe7 15 lDxd7 Wxd7 16 a4 as ShabalovKamsky, USA Ch 1993. 10....*.d6 1O.•..*.g6 11 Wb3 (11 .*.d3 .*.xd3 12 lDxd3 .*.e7 Ki.Georgiev) I1...a5?! (weakening and ineffective; I1.. ..*.e7 12 .*.e2 0-0 13 0-0 lDfd7 14 lDxg6 hxg6 is given by Georgiev, without venturing an assessment - probably White is a little better after 15 .*.f3, angling for a well-timed e4) 12 a3 a4 13 WdllDe4 14 .*.e2 .*.d6 15 lDxg6 hxg6 16 lDxe4 dxe4 17 Wc2 f5 18 .ixd6 Wxd6 and now Georgiev gives 19 b4! ±, instead of 19 b3 lDd5 20 bxa4 g5 =as played in Ki.GeorgievPiket, Moscow OL 1994. l1g4 Probably not the most accurate. White can think in terms of a Pillsbury Bind after 11 .*.g5!? and if 11 ....*.g6, 12lDxg6~. There is also 11 .*.d3!? 11 ....*.g6 12 1Wb3 'fIe7 13 .*.e2 lDfd7 Ki.Georgiev-Smagin, Yugoslavia 1995.

=

=

=

B) 7 •••lDh5 (D) This seems very solid here. Black gains considerably from not having committed his bishop to e7. 8.*.13 8 g3 .*.e7 9 h4lDxf4 10 gxf4lDf6 110-0-0 g6 12 e3 1/2- 1/2 Bareev-Kramnik, Biel IZ 1993. 8 .*.d2 lDb6 9 e4 (9 h3!? Bareev) 9 ...dxe4 10 lDxe4 .*.e7 =Bareev-Short, Linares 1992.

8.••g6 8 ....*.e7 9 e3lDxg3 10 hxg3lDf6 11 .*.d3 h6 120-0-0 .*.e6 13 ~bl WaS 14 lDe5lDg4 15lDxg4 .*.xg4 16 f3 .*.e6 17 .*.f5 and White keeps a strong grip in the centre, Piket-Timman, Wijk aan Zee rpd (3) 1995. 9 e3lDxg3 10 bxg3 .*.d6 1O....*.g7 11 .*.d3 0-0 120-0-0 %:le8 13 ~bl (13 %:lh4lDfS 14 %:ldhl b6 15 'i1'd2.*.f6 16 %:l4h2 .*.g4 17 ~bl c5 ; S.Martinovic-Cvetkovic, Yugoslavia 1995) 13 ...lDfS and now Cvetkovic gives 14lDe2!'? .*.e6 15 lDf4 %:lc8 16 %:lcl b6 as unclear; 17 lDe5 looks critical. Instead, D.Gurevich-Csom, Hastings 1983/4 continued 14 e4 dxe4 15 lDxe4 .*.e6 16 lDeg5?! (16 1Dc5 = Csom) 16... .*.d5 17 .*.c4 h6 +. 11 .*.d3lDf6 12 0-0-0 .*.e6 13 ~bl 1We7 14 ~al 0-0.0 15 %:let Seirawan recommends 15 %:lbl followed by %:lhcl, b4, etc., the same plan as in Thkmakov-Ivanov after 7...lDb6. IS••. ~b8 16 c!ba4 h5 with equality, Piket-Seirawan, Amsterdam Donner mem 1995. There is still much to be explored in the 4 ...lDbd7 line. and Black can make the play more lively than in the

Systems with ... c6 4 ... i.e7 and ...c6 line. White's edge is slender, particularly if Black plays

7...tOh5. If White wants to try something a little different, then after 5 cxd5 exd5 6 'ifc2 c6, instead of 7 i.g5 or 7 i.f4, he can try 7 g3!?, not mentioned in ECO. Since this is a book on i.f4, and since this book is already much longer than originally envisaged, I shall not try to examine this further! Back to the key position we left earlier.

7.4 The Exchange Variation, 7 cxd5

217

position too, Black has alternative recaptures: A: 7•••lOxdS 217 B: 7•••cxdS 218 C: 7•••exdS 218

A) 7 •••lOxdS does not appear on the database, but this is perhaps more to do with White delaying the pawn exchange until move 8, than as the result of any specific defect in the knight recapture. This capture has been seen where Black delays castling: S•••c6 6 e3 tObd7 7 cxdS lOxdS (D) and for comparison purposes we shall examine this position.

1 d4 dS 2 c4 e6 3 M ~6 4tOf3 i.e7 5 i.f4 0-0 6 e3 c6 7 cxdS (D)

w

Naturally, in a system in which transpositions are rife, this exchange may take place a move or so later, most notably after 7 i.d3 tObd7 8 cxd5. The general policy in this book is to cover recaptur~s with the e-pawn in this section, and to'cover other recaptures according to the move-order in which White initiated the exchange. It should not be forgotten that in this

8 tOxdS exdS 9 a3 Avoiding the irritating check. After 9 i.d3, P6cksteiner-F.Schuh, Austrian League 1988/9 continued 9 ... i[}f6 10 0-0 i.g4 11 h3 i.h5 12 l:lbl 0-0 13 b4 a6 14 a4 ;t, but Black should of course have taken the opportunity to play 9 ...i.b4+ 10 lOd2tOf6, with White at best having a tiny edge. This sort of possibility might encourage White to delay cxd5 until he has developed his bishop, allowing the king a quick

218

The Queen's Gambit Declined: 5 J.f4!

getaway, for example 7 .i.d3 0-0 8 cxdS tLlxdS 9 tLlxdS exdS 10 0-0 !. 9...tLlf6 10 .i.d3 10 "c2 0-0 transposes into Tisdall-0stenstad, Gausdal Eikrem mem 1997 (reached via S...O-O 6 e3 tLlbd7 7 cxdS tLlxdS 8 tLlxdS exdS 9 a3 tLlf6 10 "c2 c6), play continuing 11 .i.d3 .i.g4 12 .i.eS ~h8 13 h3 .i.hS 14 .i.xf6 .*.xf6 IS g4 ~g6 16 ~xg6 hxg6 17 0-0~e7 18 tLles ~d619 f4 gS 20~g2 =. Of course, in the move-order with an early ...c6 rather than ...0-0, Black could try lO .....aS+!? 10...~g4 11 h3 ~h5 12 0-0 as! Stopping the minority attack. In Crisan-Cifuentes, Amsterdam 1989, White played 13 "c2 ~xfJ 14 gxfJ, though there is no particular reason why this should give White anything. On other moves, Black is comfortable. B)

7...adS (D)

w

more modestly placed bishop might be useful in assisting queenside expansion. For example, Smuk-Zugaj, Ljubljana 1992 went 5 ~f4 c6 6 cxdS cxdS 7 e3 a6 (7 ...0-0 leads to the diagram position) 8 .*.d3 tLlc6 9 a3 0-0 10 0-0 bS 11 :cl ~b7 12 tLles :c8 13 "c2 tLlaS 14 fJ tLlc4, reaching Black's standard queenside formation. White tried rushing forward on the kingside with IS g4 h6 16 gS hxgS 17.*.xgS %S 18 .*.xe7 'ilxe7 19 .i.xc4 dxc4 20 "g2 'ilf6!? 21 tLld7 'ilh6 22 tLlxfB "xe3+ 23 'ilf2 "g5+ 24 ~hl :xfB 2S :cel tLlf4, but Black had excellent compensation for the exchange. It is not being claimed that the above game is necessarily important; it just happens to be the only game on the database with this direct transposition. AJull theoretical survey of this line is beyond the scope of this book, but it is important to recognize that Black's chances of counterplay are considerably reduced if he plays ... tLlbd7 instead of ...tLlc6. This provides White with yet .another reason for waiting until Black plays ...tLlbd7 before exchanging on dS.

C) 7...exdS (D) Again it must be emphasized that most of the examples below were reached with a different timing of the pawn exchange. 8~d3

This is perfectly playable, transposing into one of the main lines of the Exchange Slav. At first sight, the fact that White's bad bishop is outside the pawn chain, while Black's isn't, is a major plus for White, but in fact Black's

8 ~e2 tLlbd7 9 0-0 a6 10 illeS tLlxeS 11 ~xeS 1/2- 1/2 Umanskaya-Emeste, Dresden 1993 (actually reached via 7 .*.e2 and 8 0-0) does not advance theorymuch. 8...tLlbd7 9 h3

Systems with ... c6

219

lLlg6 13 i.h2 i.d6 Ih-Ih SzekelyA.Petrosian, Erevan 1984. A standard equalizing plan. 2) IllDeS lLl6d7 12 :bllLlxeS 13 i.xeS lLlg6 14 i.g3 i.d6 TunikA.Petrosian, Sverdlovsk 1984. The same basic idea. 10•••lOrs 11 0-0-0 This is probably the most accurate. 1) 11 g4 and now: la) 11...i.e6 gives White good chances of an edge after either 12 0-0-0 (see main line) or 12lLlgS i.d6 (12 ...cS!? 13 i.eS h6 14 M {14 lLlxe6!?} 14... lLl6d7 IS 0-0-0 :c8 16 ~bl a6 with a finely balanced position, Timman-Hubner, Wijk aan Zee 1992) 13lLle2 .as+ 14 ~fl 'il'c7 IS i.xd6 .xd6 16lLlf4 i.d7 17 a3 :e7 18 b4;!; M.Gurevich-Andersson, Reggio Emilia 1991. Ib) However, 11...i.d6! 12 i.xd6 .xd6 improves, e.g. 13 gS lLlhS 14 0-0-0 g6! IS lLles lLle6 16 h4lLleg7 = Bareev-Andersson, Biel 1991, or 13 0-0-0 i.e6 14 ~bl :ac8 IS :cl .b8 16 gSlLl6d7 17 .d2 :c7 18 h4lLlb6 19 .c2 lLlc4 20 lLld2lLld6 =M.Gurevich-Andersson, Reggio Emilia 1988/9. 2) After 11 O-O?!, 11...lLlg6 followed by ...i.d6 equalizes easily. Instead 11...lDe6?! 12 i.eS g6 13 a3 lLlg7 14 b4 i.fS IS :fcl :c8 16 i.xfS lLlxfS 17 .d3 ;!; was KozlovskayaSkripova, St Petersburg 1996. 3) 11 lLles and now, rather than 11...lLlhS?! 12 i.h2 g6 130-0 lLlg7 ;!; Lesecq-Caron. Torey 1991, 11...lLlg6 should be safe enough. 11•••i.e6 1l...i.d6 12 .tgS. possibly with e4 to follow. makes it difficult for Black to untangle his kingside.

=

9 .c2 :e8 10 0-0I? (10 h3 transposes to the main line) 1O... lLlfS 11 :abl and now: 1) ll...aSI2a3lLlh513i.g3lbxg3 14 hxg3 i.d6 IS b4 .e7 16 bS! ± Browne-Hulak, Wijk aan Zee 1983, with the idea that if 16...i.xa3, then White replies 17 lLles i.d6 (17 ...cxb5? 18 lLlxbS; 17...•d6 18 bxc6 bxc6 19 lLlxc6!) 18 bxc6 i.xeS 19 lLlxdS ± Browne. 2) 11...lLlhs 12 i.eS f6 13 i.g3 i.e614 b4lLlxg3 IS hxg3 i.n 16 bS;!; Taimanov-Klovans, USSR 1966. 3) ll...lLlg6 12 i.g3 as 13 a3lLlh5 14 b4 lLlxg3 IS hxg3 axb4 16 axb4 i.f6 17 bS ;!; Hemdl-Thomsen, Manila OL 1992. The minority attack has more bite when Black 'wins' bishop for knight than when there is a straight exchange of bishops with ...i.d6. 9•••:e810'il'c2 White is now almost obliged to play aggressively with queenside castling, as h3 is not a constructive move in conjunction with~ngside castling. Black is solid after 10 0-0 lLlfS, for example: 1) 11:c1 a6 (reached via S...c6 6 e31Obd7 7 h3 0-0 8 :cl a6) 12.c2

220

The Queen's Gambit Declined: 5 ~f4!

12~bl

12 g4!? ~4 (12 ...l:c8!?) 13 1.xe4 dxe4 14 liJd2 b5 15 ~bl ± PsalchisYusupov, USSR Ch 1981. 12•••:c8 Black gets carved up if he plays anything remotely slow: 1) 12 ... a6? 13 g4.aS 14liJg5 c5 IS dxc5 1.xc5 16 1.e5 prises open the kingside, Ornstein-Vilar, Osterskars 1995. 2) 12... aS 13liJg5 b5 14 i.e5 h6 IS liJxe6 liJxe6 16 g4 a4 17 i.f5 and White's pawn-storm is much better backed by pieces than Black's, Korchnoi-Eslon, Bie11984. 13liJg5! 1) 13 g4 bS 14liJg5 .•b6 15 1.eS h6 (15 ... g6 16 f4 ± Miles) 16 liJxe6 liJxe6 17 i.f5 liJd7 18 i.xe6 fxe6 19 .g6liJxe5 20 dxeS ;!; Miles-O.Jakobsen, Esbjerg 1984. White may be able to do even better by delaying the pawn-push and getting pieces into play more quickly. 2) There is also the quiet positional approach: 13 l:c1!? c5 14 dxc5 1.xcs IS .a4 a6 16 l:hdl ;!; Tal-Vooremaa. Tallinn 1979. 13...b5 Timman gives instead 13 ...i.d7, continuing 14 i.e5 h6 15 liJf3 cS ;!;, but 14 g4! keeps more of a grip on the position; if 14... c5??, IS liJxd5liJxd5 16 i.xh7+. 14 i.e5 h6 15 liJxe6 liJxe6 16 g4 liJd717 h4! White's kingside attack rolls on nicely. Tmunan-Karpov, Bugojno 1978 continued 17... b4 18 ~2 1.xb4 19 f4 c5 20 1.a6 i.e7 21 i.xc8 .xc8 22 liJg3 f6 23 l:xh6! llJef8 24 :th3 c4 25 liJfS fxe5 26 fxeS .c6 27 :dhlliJg6

28llJd6 (28liJxg7, Tmunan, is a cleaner kill) 28 ... llJdf8 29liJxe8 "xe8 30 l:h5 "c6 31 "fS as 32 e6 'ilxe6 33 'ilxd5 a4 34 l:cl, and White mopped up comfortably enough in the endgame. It is noteworthy that Black has been avoiding this variation in grandmaster play in recent years. For example. in the 4 ... liJbd7 move-order, attention has switched to lines involving an early ...liJb6 or ...liJh5.

7.57.id3 1 d4 d5 2 c4 e6 3 liJc3 liJf6 4liJf3 i.e7 5 1.f4 0-0 6 e3 c6 7 i.d3 (D)

B

Simple and unpretentious development, though it does raise the question of whether Black can gain a tempo with 7 ... dxc4. The main alternative, 7 ... liJbd7, will generally lead straight to the Exchange Variation, which as we have seen is promising for White, but there are also good alternatives. Our lines are: A: 7•••dxc4 221 B: 7...liJbd7 222 Before considering these variations, we may briefly note the possibility

Systems with ... c6 7...:e8!? 8 h3 (8 cxd5!?) 8...dxc4 9 i.xc400 1Oi.g3?! (100-0) 1O...i.d6 11 i.xd6 'ii'xd6 12 'ii'b3 ~d7-= F.Naumann-M.Esser, Aachen 1988. A)

7•••dxc4 8 i.xc4 ~bd7 (D) 8... ~d5 9 i.g3 ~7 10 0-0 ~xc3 11 bxc3 b6 and now, rather than 12 'ii'c2?! i.b7 13 i.d3 h6 00 Milla-Sprotte, Moscow OL 1994, 12 e4! leaves White with control of the centre.

9h3 1) 9 e4 is no refutation of Black's choice of move-order: 9... b5 10 i.d3 b4 11 ~a4 i.b7 with a comfortable Semi-Slav type position for Black, who will shortly play ...c5, for example 12 'ii'c2 h6 13 :cl :c8 with an unclear position. 2) 9 :cl (via 7 :c1) 9...~d5 10 i.g3 ~xc3 11 bxc3 b6 120-0 i.b7 13 ~5?! (13 e4 ~) 13 ...~xe5 14 i.xe5 i.d6 with equality, Plotnikov-Zuev, OreI1996._. 3) 90-0 and now:", 3a) 9... ~h5?! 10 i.e5 ~xe5 11 dxe5 and now 11...'ii'xdl 12 :txdl g6 13 g4 ~g7 14 ~ b5 15 i.e2 c5 16 :acl c417 ~4 gave White complete

221

central control in Ye Rongguang-Handoko, Beijing 1992. The exchange of queens is extremely time-consuming, but Black's position is also bad after l1...g6 12 'ii'e2 b5 13 i.b3 as 14 a4 b4 15 ~4, A.Hammond-Sage, England 1996. The developing moves in lines '2' and '3' both seem quite promising for White, perhaps more so than the main text. Black may do best to copy the Meran formation with 9... b5 followed by ... a6. In such position it is not necessarily an unmixed blessing for White to have the bishop outside the pawn chain. A recent example: 3b) 9... b5 10 i.d3 i.b7 11 :cl a6 12 ~e4 ~xe4 13 i.xe4 :c8 14 ~5 ~xe5 15 i.xe5 "d7 followed by ... f6 is equal, P.Novak-Scblecht, Bundesliga 1996n. 9•••~dS 1) Black neglected the ... c5 break in Rubinstein-Sartori, Barmen 1905, and was soon suffering: 9.....b6?! 10 i.b3 :d8 11 0-0 ~f8 12 'ii'e2 ~g6 13 i.h2 i.d6 14 ~5 ±. 2) Black adopted the Meran formation in A.Moran-Tseleng, Parana 1993: 9... a6 10 0-0 b5 11 i.d3 c5 (l1. ..i.b7) 12 a4 b4 13 ~, but was getting dangerously behind in development. Play continued 13 ... aS?! 14 dxc5 ~xc5 15 ~xc5 i.xc5 16 :cl i.e7 17 'ii'c2 ±. 3) 9...~b6!? is a perfectly acceptable alternative; the other knight swings to d5. 10 i.d3 ~bd5 11 i.e5 ~xc312bxc3c5(l2 ...'ii'aSI3"c2h6 14 e4 c5 150-0 i.d7 = R.SiegmundF.Holzke, Nettetall991) 130-0 b6 14 ~g5 h6 15 i.xf6 i.xf6 16 ~7 :eS = Skembris-Kofidis, Greek Ch 1994.

The "Queen's Gambit Declined: 5 ~f4!

222

4) 9 ... bS!? 10 .te2 .*.b7 11 a3 a6 12 b4 as E.Meduna-Jirovsky, Czech Ch 1997. 10.*.g3 I) 10 .*.h2 WaS (1O ...bS?! 11 .te2 WaS?! 120-0 lOxc3 13 bxc3 .tb7 14 "'c2 ± Barkhagen-Rylander, Swedish Ch 1997, is an inappropriate mixture of plans) and now: la) 11 :cl?! lO7b612 .tb3 (Black should be able to get away with snatching the a-pawn after 12 .td3 lOxc3 13 bxc3 Wxa2 ee) 12....tb4 13 Wc2lOxc3 14 bxc3 .*.a3 IS :dllOds 16 :d3? (16 .*.xdS =) 16... eS! 17 e4 lOxc3 + Steinitz-Chigorin, St Petersburg 1896. Ib) 11 Wc2lO7b6 12.tb3lOxc3 13 bxc3 lOds 14 .txdS 112_1/2 Van Parreren-Palatkova, Prague 1992. The bishop is a little more actively placed on g3 than on h2, provided it is out of reach of Black's minor pieces, and Wc2 seems here a more flexible way of protecting c3 than :cl. 2) Ibragimov gives 10 lOxdS exdS 11 .td3, continuing 11...:e8 12 0-0 lOfS 13 :bl ;t, but Black surely should grasp the chance to play 11.. ..*.b4+, taking advantage of White's loss of tempo with the bishop. 10......85 11 bS 1) By analogy with Van ParrerenPalatkova (' Ib' in the previous note), Black could also consider 11 ...107b6. If then 12.*.d3, Ibragimov gives 12...g6 13 Wb3 cS 14 WbS ;t, but 12 ...h6 13 "'b3 lOxc3 14 bxc3 Wa3!? should be satisfactory for Black. 2) 11...b6 12 0-0 lOxc3 13 Wxc3 Wxc3 14 bxc3 .*.b7 Ju.BolbochanJa.Bolbochan, Mar del Plata 19S1, but 13 bxc3 is more challenging.

=

"'c2

=

3) 11...cS 12 .txdS exdS 130-( Ibragimov. 12 .*.d3lO7f6?! Ibragimov notes 12 ... lOb4?! .txh7+ ~h8 14 "'bl g6 (14 ... fS .tg6 :f6 16 .*.e8 ±) IS .*.xg6 fxg6 "'xg6 with a strong attack. The simple 12 ...h6 equalizes. 13'irb3.ta6 13 ... lOxc3 14 bxc3 lOds IS J:I lOxc3? 16 :xc3 .tb4 17 ~d2 Ibragimov. 14 0-0 b4 15 lOe2! ;t Ibragim( Vaganian, Berlin 1991. B)

7 ...lObd7 (D)

w

This would be a good moment j White to enter the Exchange Variati with 8 cxdS. We consider the alten tives here. 80-0 1) 8 lOeS!? is examined under t move-order 7 lOes lObd7 8 .*.d3, Section 7.8, Line Al below. 2) 8 h3 is too quiet for White hope for much; he has already 'sum dered' the battle of the tempo witt: .*.d3, so waiting pawn moves shot generally be avoided.

Systems with ... c6 2a) S... dxc4 transposes into the 7 ... dxc4line (Line A above). 2b) S... a6 (this and '2c' are less convincing) 9 cxdS!? (9 0-0 dxc4 10 .i.xc4 bS II .i.d3 cS 12 lLleS .i.b7 = Bemei-J.Toth, Hungary 1992) 9...exdS 10 0-0 :leS II :lbl ;t Dottling-Schulzke, Lowenstein 1997. 2c) S...:leS 9 0-0 lLlfS 10 cSt? (10 :lcl dxc4 11 .i.xc4 lLldS 12 .i.g3 lLlxc3 13 :lxc3 .i.d6 14 .i.xd6 Wxd6 IS e4 ;t Lempert-Shur, Naberezhnye Chelny 1993) 1O...1L16d7 11 b4 i.f6?! ( 11.. .lLlg6 fIrst cuts White' s options) 12 e4 (12 i.eS i.xeS 13 lLlxeS ;t Zatulovskaya-Kozlovskaya, Moscow 1979) 12...eS (otherwise Black suffocates) 13 dxeS lLlxeS 14 .i.xeS lIxeS IS :lcl (even stronger than taking the exchange) and Black is in deep trouble, Malaniuk-U.Lau, Dortmund 1993. 8•••1L1bS Naturally, it is to avoid this that White sometimes plays S h3. S...dxc4 is still possible, but other lines tend to leave Black's pieces in contortions: 1) S... b6 91L1eS (9 cxd5!?) 9 ...1L1xeS 10 dxeS (10 .i.xeS ;t) 1O...1L1d7 11 cxdS exdS 121i'hS g6 13 'iff31L1cS 14 i.c2 fS IS :lfdl ;t Kotov-Kasparian, Parnu 1947. 2) S...h6?! (slow) 9 e4 (White should probably prepare e4 more gradually) 9 ...dxe4 10 lLlxe41L1xe4 II .i.xe4 lLlf6 12 i.c2 lLlhS? (Black must play for ... cS) 13 .i.eS ± Strikovic-Maas, Haarlem 1995. 3) S...:eS can easily transpose into other headings, after, fOr example, 9 cxdS, 9 Wc2 or 9 h3 (this is '2c' in the previous note). Also 9 We2!? lLlfS 10 lLleS1Ll6d7 II :ladllLlxeS 12 dxeS ± Bum-Sterling, Paris 1900; or 9 :lc I

223

lLlfSlOlL1eslLl6d7111L1e2?! (11 e4±) 11...f6 00 Mikuli~ic-Puljek, Croatian Cht (Tucepi) 1996. 9.i.e5 1) 91L1eslLlxf4 10 exf4 fS 11 cS gS (11 ...1L1xeS!?) 121L1xd7 i.xd7 13 fxgS i.xgS 14 f4 .i.f6 IS i.e2 b6 is equal, Danielian-I.Saric. Leningrad USSRYugoslavia jr 1991. 2) 9 'ifc2 is a more thematic move. 9 ...1L1xf4 10 exf4 dxc4 II .i.xc41L1b6 (ll...cS 12 :adl cxd4 13 lLlxd4 ;t Dorfman) 12 i.d3 g6 13 a3 lLldS 14 g3 i.d7 IS lLle4 ;t Dorfman-Ivkov, French Cht 1993. 3) Dorfman notes that 9 cS is less effective after 9 ...1L1xf4 10 exf4 b6 II b4 bxcS 12 bxcS WaS followed by ....i.a6. 9••. 1L1xe5 I) 9 ... f6? falls into the thematic trap 10 lLlgS!, although in one of the junior games on the database this was missed. 2) 9 ... fS?! 10 'ifc2lLlhf6?! II cxdS lLlxeS 12 dxeS lLlxdS 13 .i.c4 ± i.d7 14 :ladl bS IS i.b3 Wb6 16 e4 fxe4 17 Wxe4 :lf4?? Is1L1xdS 1-0 Bogoljubow-Rosselli, Baden-Baden 1925. lOdxe5 g6 1O...dxc4?! II .i.xc4 "xdl 12 :lfxdl g6 131L1e4 ± Vaganian. ll1i'c2 f6!? 1l...i.d7 12 :lfdl :lcS, and now instead of 13 :lac1?! bS! 14 cxbS cxbS IS a3 as, Beliavsky-Vaganian, Reggio Emilia 1991, White should, according to Beliavsky, play 13 a3 bS 14 cxbS cxbS IS b4 as 16 'ifb3 ±. The text-move looks weakening at fIrst, but there is no effective way to get at the black king. 12:ladl

224

The Queen's Gambit Declined: 5 ~f4!

12 e4!? is a try for an edge, for example 12... fxe5 13 exdS exd5 14 cxdS, and if 14... cxd5?!, 15 /t)xd5; or 12...~f4 13 exf6.*.xf6 14 eS .*.g7 15 :lfeI. 12... dxc4!? 13 .*.xc4 f5 00. 12•••fxeS 13lbxeS .td614 f4 .txeS 15 (xeS .*.d7 16 .*.xg6 bxg617 "xg6+~g718:lf6

White has compensation for the piece but is not necessarily better, Stefansson-Danielian, Buenos Aires 1993.

7.67 flc1 1 d4 d5 2 c4 e6 3 ~ ~f6 4 ~f3 .te7 5 .tf4 0-0 6 e3 c6 7 :tel (D)

The standard development of the rook in the .tg5 system makes less of an impression here. White does not yet know whether the rook is best placed at cl, dl or even bI. Still, in an opening in which waiting moves such as a3 and h3 are rife, this quiet developing move can hardly be bad. 7...~bd7 7 ... a6 8 cxd5 (8 cS!?) 8...exd5 9 .td3 ~bd7 10 h3 :le8 11 0-0 ~f8 12 ~e5 ~ Hribar-Banic, Ljubljana 1997. 8h3

Yet again, 8 cxd5 is to be considered. Also: 1) 8 .*.d3 dxc4 9 .*.xc4 ~dS 10 .tg3 (10 O-O!?) 10...WaS 11 0-0 ~xc3 12 :lxc3 (12 bxc3 b6 {12 ... bS!?} 13 We2 .*.b7 14 ~d2 :lfd8 15 a4 ~ Ivkov-Pfleger, Palma de Mallorca 1966) 12 ... bS 13 .tb3 .tb714 'irbl b4 15 :lc4 ~f6 00 Znamenacek-Chmiel, Karvina 1989. 2) 8 'iVc2 :le8 9 a3 ~f8 10 .tg3 ~g6 11 cxd5 exd5 12 .td3 .*.d6 13 0-0.txg3 14 hxg3 'ire7 15 ~2 .*.d7 16 e4 dxe4 17 ~xe4 :lad8 18 :lfel ~xe4 19 .txe4 ~ Nemet-Rukavina, Portoroz 1971. In this type of variation, White's h-pawn is better placed on g3 (after hxg3), from where it covers f4, than on h3.

8..•~!? Trying to take advantage of White's half-heartedness in covering e4. The usual stodge has also been tried: 1) 8... a6 9 a3 dxc4 10 .txc4 c5 11 0-0 b5 12 .i.a2 .*.b7 13 dS exdS 14 ~xd5 /t)xdS IS .*.xd5 .txdS 16 WxdS ~f6 17 'irb7 :le8 18 :lfdl "c8 = Malaniuk-Emeste, Katowice 1993. 2) 8...:le8 9 .*.h2 ~f8 10 .td3 ~g6 11 Wc2 dxc4 12 .txc4 ~5 13 0-0 .tgS? ! (difficult to understand) 14 ~xg5 Wxg5 15 ~e4 Wh4 16 .tg3 ± H.Abramson-A. Villegas, Villa Ballester 1996. 9lbxe4 Striking before the wall is fully built. 9 .td3 f5 10 'irb3 ~h8 11 .*.e2 g5 12 .*.e5+ ~xe513 ~xe5 'iraS 14 Wc2 .td6 15 0-0 ~xc3 16 'irxc3 Wxc3 17 :lxc3 f4 18 .tg4 ~g7 19 :ldl ~f6 = Gavrikov-Vaganian, USSR Ch (Moscow) 1988.

Systems with ... c6 9 •••dxe4 10 ~ .as?! Decentralizing. 1O... f5 is natural, with maybe even ... .1g5 to follow. The position is not yet clear. 11 83 rs 12 .1e2 cS 13 dxeS eS 14 .1h2 l:ldS IS 0-0 .xeS 16 b4 .c7 17 cS Adianto-Irwanto, Genting Highlands Z 1995. White's pressure on the queenside counts for more than the blockage of the bishop on h2.

7.77 h3 1 d4 dS 2 c4 e6 3 ~c3lDf6 4 M .1e7 S .1r4 0-0 6 e3 c6 7 h3 (D)

B

The reader should not be misled by the relative brevity of this section. 7 h3 is, in the opinion of the author, the most accurate of the moves at White's disposal, planning to meet 7 ... lLlbd7 with 8 'ifc2. The point is, however, that most games starting with 7 h3 quickly transpose into lines -giVen earlier. Here we consider only the odds and ends. 7...~bd7 The alternatives seem slightly unthematic. 1) 7 ... dxc4?! surrenders the battle of the tempo too easily. 8 .1xc4 ~5 9

225

.1h2 ~d7? (now Black is a whole tempo down on 7 ...~bd7 8 .1d3 dxc4, etc.; 9 ....1d6 10 .1xd6 'ifxd6 11 0-0 ~7 12 e4 ~c3 13 bxc3 is better, but not really satisfactory) 10 0-0 f5 11 l:lc1 ~7f6 12 ~5 and White's advantage is already close to decisive, Seirawan-Miolo, Indonesia 1983. 2) 7 ... b6?! is OK against 7 'ifc2, but not against 7 h3. White plays it as a ...b6 system in which Black has played an irrelevant ... c6. The plan is cxd5, .1d3, lLle5, 'iff3, etc. For example, 8 .1d3 (8 ~e5 is more accurate, while 8 cxd5 ~xd5! 9 ~xd5 .xd5, with ....1a6 to follow, is less rewarding) 8 ....1b7?! (8 ....1a6!?; 8 ...dxc4!?) 9 cxd5 exd5 10 ~5lLlbd7 11 .f3lLlxe5 12 .1xe5 ~7 13 .1f4 l:le8 14 0-0-0 lLlfS 15 h4 .1d6 16 ~2 c5 17 dxc5 bxc5 18 .1c4 ± Lechtynsky-Pfleger, Bundesliga 199112. S83 Another tempo move. 8 .1d3, 8:C 1 and 8 .c2 lead to lines already considered, likewise 8 .1e2 dxc4 9 .1xc4. 8 cxd5 ~xd5 (8 ...exd5 has already been examined) 9 lLlxd5 exd5 10 a3 h6?! (10...~f6) 11 .1d3 ~f6 12 0-0 .1d613 ~5 'ifc714 'iff3;t SopanyiA.Good, Bratislava U-14 Wch 1993. S•••b6 For 8 ... a6 9 c5, compare 6 ... a6, Chapter 8.1. 9 cxdS lLlxd5 10 ~dS exd5 11 .1d3 eS 12 dxeS 12 O-O!? leads back to the 6 ... b6 variation, with White having the extra tempo a3, but also having played h3. On the whole, the difference perhaps modestly favours White. After the text, the hanging pawns are defensible.

226

The Queen's Gambit Declined: 5 1./4!

12••• bxc5 130-0 'ii'b6 14 b4!? 14 "c2 =A.Petrosian. 14•••cxb4 15 axb4 .txb4 16 clOci2 tbcs 17 .tel .ta6 18 lba6!? 'ii'xa6 19 "bl ~ 20 .txh7+ ~h8 Lputian-A.Petrosian, Dortmund 1992. Now 21 .tc2, Petrosian, keeps the position unclear.

7.8 Odds and ends 1 d4 dS 2 c4 e6 3 W ~6 4lDn .te7 S.tf4c66e3 In this rounding-off section we consider two types of idea: A: 6 •••0-0: Odds and ends for White on move 7 226 B: Black delays castling 227

A) Odds and ends for White on move 7 These are: AI: 7lDeS!? 226 226 A2: 7 a3 226 A3: 7 c5 A4: 7 .te2 227 AS: 7 l:[bl 227 A1) 7lDeS!? (D)

A recent attempt to enliven the p0sition. 7•••lDbd7 8 .tdllDxeS S...dxc4 9lDxc4lDb6 10 lDes cS 11 dxcS .txcS 12 0-0 lDbdS 13 lDxdS exdS (13 ...lDxdS 14 .txh7+ ~xh7 IS .c2+) 14 l:[cl .td6 IS .tbl with excellent play against the isolani, Gheorghiu-Rainberg, Moscow OL 1994. 9.txeS dxc4 9 ... b6 ~ Tregubov. 10 .txc4 b6 11 0-0 .tb7 12 'iWe2 .td6 12...cS?! 13 l:[fdl cxd4 14 .txd4 ±; 12...lOd5 13 e4;t Tregubov. 13hdllDdS 13...•e7?! 14 f4 ± Tregubov. 14lDe4 .te71S "g4 ~616 "g3;t Tregubov-Pigusov, St Petersburg 1993. Definitely the most promising of the miscellaneous tries. A2) 7 a3 b6 8 .td3 .tb7 (S ....ta6!?) 9 0-0 cS (Black has unnecessarily lost a tempo) 10 .e2 (10 dxcS!?) 10...cxd4 11 lDxd4lDbd7 12 cxdS eS 13 lDfS exf4 14 d6 .txd6 IslDxd6 f3 16 gxf3 lDeS 17 lDxb7 .xd3 18 .xd3 lDxd3 19 b4 as 20 bxaS l:[tb8 21 a6 lba6 22 :tdl ± Forintos-Dely, Hungarian Cht

199213. B

A3) 7 c5 and now 7 ...lDbd7 S .td3 h6 9 g4 eS 10 dxeSlDe4 11 .txe4 dxe4 12 lDxe4lDxcS gave Black compensation for the pawn in A.Juarez-G.Mendez, Buenos Aires 1993. Black has simpler ways of handling the position, notably 7 ... b6 8 b4 as 9 a3 lDhS!? .... In general, White should wait for ...a6 or ... lDbd7 before playing cS.

Systems with ... c6 A4) 7 .te2 may of course be answered by 7...dxc4 8 .txc4, or 7... ~bd7 8 0-0 dxc4, both transposing into the 7 .td3 variation. However, if7 ...~e4, then 8 0-0 ~d7 9 ~xe4 dxe4 10 ~d2 f5 11 h3 (11 f3 g5 12 .tg3 f4 13 .tf2 is also strong) 11...b6 12 f3 exf3 13 .txf3 ± R.Antoniowski-G.Kuba, Szombathely 1993. AS)

7 :lbl ~bd7 8 b4?! (not an inspired idea) 8...dxc4 9 .txc4 ~b6; N.MoyseE.Crea, Bozen 1992, is one of the few ways White has of getting a poor position from this opening. B) Black delays castling 5•.•c66e3 (D)

la) 7 ... ~?! 8 ~xe4 dxe4 9 ~2 .tb4 10 a3 .txd2+ 11 "xd2 0-0 12 "c2 f5 13.td6 ± Tarrasch-Von Scheve, Leipzig 1894. Ib) 7...~f8 8 c5 ~g6 9 .th2 10 a3 ~4 11 .td3 ~xc3 (11...f5 ±) 12 "d2 ~h4 13 ~xh4 .txh4 14 b4 ± Tarrasch-Walbrodt, Vienna 1898. lc) 7 .....aS 8 ~2 0-0 9 .td3 c5 10 ~b3?! (10 O-O~) 10..... b6 11 0-0 cxd4 12 exd4 dxc4 ; Sarosi-Pasztor, Hungarian Cht 199112. White must be prepared to recapture on c4 with a knight. 2) 7 .td3 and then: 2a) 7 ...dxc4 8 .txc4 ~b6 9 .td3 ~bd5 10 .tg3 11 "b3 is quite reasonable for Black, but 11...~b4? 12 .tbl ~fd5 13 0-0 ~xc3 14 "xc3 "b6 15 e4 ± Schlechter-F.Lee, London 1899, is a poor follow-up. 2b) 7...~f8 8 ~e5?! (8 cxd5!? ~xd5 9 .tg3 ~) 8...~g6 9 ~xg6 hxg6 Gunkel-Lilienbeck, Dortmund 1987. 3) 7 :lcl ~5 8 .td3 ~xf4 9 exf4 dxc4 10 .txc4 ~b6 11 .td3 ~5 12 g3 ~xc3 13 bxc3 0-0 14 h4 ~ G.Goldberg-Savitsky, Leningrad 1932. After 6 .....aS 7 ~d2 .tb4 8 "c2 cS, instead of 9 dxcS ~bd7 ~ SulikPlanas Garcia, Buenos Aires OL 1939, 9 ~b3 ± leaves White well ahead of the Nimzo-Indian. 7 .te5~78"c2(D) 8 ~d2 ~xe5 9 "xh5?! (9 dxe5) 9 ...~7 10 .td3 ~f6 with an equal position, Borsi-Pasztor, Hungarian Cht 199112. 8 h3 ~xe5 9 dxe5 g6 10 Wc2 0-0 (1O.....aS 11 .te2.td7 120-0 b5 =Ye Rongguang-Li Zunian, Beijing 1995) 11 0-0-0 "as 12 h4 b5 13 cxb5 cxb5 14 .txb5 .tb4 15 ~4 .tb7 16 :ld3

"as

"as

=

.... ........,

6•••~h5

The most important independent try.

There have been a few games where Black plays 6...~bd7 without castling, but these usually involve slow and passive plans of development. For example: 1) 7 h3 is best met by 7...0-0. Others:

227

228

The Queen's Gambit Declined: 5 j"f4!

:'fc8 17 ~bl a6 was unclear in Van Wely-Posazhennikov, Vlissingen 1997. Sharp play, but note that White wasted a tempo with h2-h3-h4.

8 •••lOxe5 9 dxe5 g6?! Stereotyped.9 ...f5!? 10 0·0·0 .td7 11 h4 (5 12 ext'6 .txt6 13 g4 with a clear advantage for White, Kallai-A.Schneider, Wiesbaden 1990.

B

This was a complicated chapter to write, with White often having around seven choices on a particular turn, and Black having three or four reasonable replies to each of these. As a result, theory does not go deep. As in the ... c5 lines, however, White does well to bear in mind the possibility of queenside castling.

8 Miscellaneous Systems on Moves 5 and 6 8.1 Systems with ... a6 1 d4 d5 2 c4 e6 3 ~c3 ~6 4 ~f3 i.e7 5 i.f4

Now: 229 A: 5•••a6 B: 5 •••0-06 e3 a6 231 A) 5•••a6 (D)

w

'" We start by looking at systems that do not transpose into 5...0-0 6 e3 a6. Already from the previous chapter (Chapter 7), on ... c6, we will be familiar with the idea of Black playing a string of quiet moves, trying to cover all his weaknesses and to rearrange his pieces behind the lines until he is ready to make some break to free his position. Plans with ... a6 follow similar types of idea, but with Black trying

to avoid spending a tempo on ...c6; sometimes ...dxc4 followed by ... b5 is a reasonable plan, with Black trying to get in ...c5 with one move rather than two. After 5... a6, quick transpositions to the 6 ... a6 line usually follow on either 6 c5 0-0 or 6 e3. The exceptions are as follows: 6e3 6 cxd5 ~xd5 (6... exd5!?) 7 i.g3 (7 i.d2!? - S.lvanov) 7 ...f5?! (an unconvincing way of preventing e4; 7...c5 GO) 8 e3 0-09 i.d3 c5 10 i.e5 ~6 II 0-0 ~xe5 12 dxe5 ~b4 13 i.e2 "xdl 14 lIfxdl and Black proved to have too many queenside weaknesses in BurnJ.Mortimer, Paris 1900. 6 ..•dxc4 6 ... ~5 7 lIcl (7 i.e5!?) 7 ...~f4 8 exf4 0-0 9 i.d3 dxc4 10 i.xc4 b5 11 i.d3 i.b7 120-0 lOd7 13 'it'e2 lIc8 14 lIfdl ~f615 f5 ± Yangel-Stom, Moscow 1997. Yet again, simply taking White's bishop is no guarantee of eqUality. An unconvincing alternative where Black delays castling is 6...~bd7 7 h3 c6 8 c5 b5 9 i.d3 ~5 10 i.h2 g6 11 g4 ~g7 12 'it'c2 b4 13 ~a4 a5 14 'it'e2 f6 150-0-0 ± Van Wely-Sehner, Bundesliga 199617 (the sequence of Black's moves in this game was 5...c6, 6...~bd7 and 7 ... a6).

230

The Queen's Gambit Declined: 51./4!

7.*.xc4 7 a4!? - S.lvanov. 7••• b5 8.*.d3 .*.b7 9 0-0 (D) Ivanov suggests that 9 "e2 is more accurate, and if 9 ... lLlbd7 (9 ...c5? 10 dxc5 .*.xcS 11 .*.xbS+) then 10 e4 cS 11 d5!? exd5 12 exd5, when he continues 12... lLlxd5 13 lLlxd5 .*.xdS 14 0-0-0 with an attack. However, Black can try 12 ...0-0!, meeting 13 d6?! with 13 ....*.xd6 14 .*.xd6 l:r.e8 15lLle5lLlxe5 (15 ...l:r.e6? 160-0-0) 16 .*.xe5 l:r.xe5 17 .*.xh7+ ~xh7 18 "xe5 "d3 19 l:r.dl "c4 20 "f5+ ~g8 with good attacking chances for Black.

8

9 •••lLlbd7!? Trying to show that delaying castling is not merely a question of transposition of move-order; Black seeks to play ...c5 painlessly. 9 ... 0-0 tends to favour White: 10 l:r.cllLlbd711 e4c5 (l1...b4?! 121&4 hits c7) 12 e5lLld5 13lLlxd5 .*.xd5 14 .*.bl c4 15 "e2 l:r.c8 16 "e3 and White's kingside pressure gives him a slight advantage, Kolesar-P.Kramarik, Slovakia 1995 (reached via S...dxc4 6 e3 O-O?! 7 .*.xc4 a6, etc.; 6 ... lLlds would be better in this line); or 10 "e2 c5?! (l0 ...lLlbd7) 11 dxc5 .*.xcS 12

l:r.fdl Wb6 13 a4 b4 14 a5 "a7 15lLla4 Matamoros-Martinovsky, Ubeda 1996. 10e4!? 10 "e2 c5 = Ivanov. 10...b4 10... c5 11 d5 with an attack for White - Ivanov.

±

lllLla4 White is more or less forced to sacrifice; 11 e5?! lLlhS 12 lLle2lLlxf4 13 lLlxf4 c5 and Black is slightly betterIvanov. ll...lLlxe4 Ivanov suggests that the simplest route to safety is 11.. ..*.xe4 12 l:r.c1 .*.xf3 (12 ....*.d6 13 .*.xd6 cxd6 14 .i.xe4lLlxe4 IS "el ;t) 13 "xf3 .i.d6 (13 ... lLld5? 14 .*.e4 lLl7f6 15 .*.xc7 lLlxc7 16 .*.c6+ +-) 14 .*.xd6 cxd6 15 "g3=. 12 'i'd lLld6 13 lbes lLlxe5 According to Ivanov, 13 ... lLlb5 improves, the main lines being 14 l:r.ac 1 .i.d6 15 lLlc5 lLlxd4 16 "a4 .*.xe5 17 lLlxb7 "f6 18 .i.xe5 "xeS 19lLlc5 c6 00 and 14lLlc6 .*.xc6 15 "xc6lLlxd4 16 "xc7 00. There are other messy possibilities the reader might like to explore; I haven't found a clear edge for White. 14 .i.xe5lLlb5 15 l:r.rel! .i.d6 15...lLlxd4 16 .*.xd4 "xd4 17 l:r.adl (In/ormator has 17 l:r.ed 1; a rare misprint?) 17 .....a7 is given by Ivanov as unclear, but 18 .i.xh7 should favour White.

16lLlc5 White has good compensation for the pawn, S.lvanov-Nikolaev, Russia 1992. So S... a6 is a possible independent try for Black, but he must bear in mind

Miscellaneous Systems on Moves 5 and 6 that White can reply 6 cS!?, when 6 ...0-0 7 e3 reaches a position discussed in Line B 1.

231

7 ... bS is discussed in connection with B14. B11)

7.•.tooo (D)

B) 5•••0-06 e3 86 (D)

w

Various developing moves have been tried here, and also the exchange on dS. However, the most popular choice has been the gain of space with 7 cS. The lines are: Bl: 7 c5!? 231 B2: 7:ct 234 B3: 7.c2 234 84: 7 adS 237 B1)

7 c5!? The positional justification for this is that ... b6, usually the thematic reply, is of limited effectiveness, in that Black is not going to be able to recapture with the a-pawn. However, White must be careful not to enter an Exchange Slav a tempo down if he captures on b6. Now Black has: B11: 7 •••~ 231 232 B12: 7•••b6 B13: 7...~?! 233 B14: 7•••00 233

For the time being Black ignores White's queenside advances, and simply develops. The b4-square is now covered, however, and ...b6 might follow without fear of b4. 8.i.d3!? 1) 8 a3 lOhS?! 9 .i.e2 .i.d7 100-0 fS 11 h3 lOxf4 12 exf4 gS 13 fxgS .i.xgS 14 ti'd3 .i.f6 IS :ael ti'e7 16 .i.dl .g7 17 .i.a4 ± Sarosi-Kaspersen, Copenhagen 1988. Not model play by Black; the exchange of the bishop on f4 achieves less than nothing in a Stonewall formation, and the knight would be better placed on e4. 2) 8 h3 and now: 2a) Black replied with 8 ... b6 9 cxb6 cxb6 10:c 1 .i.b7 11 .i.d3 .i.d6 = in Muse-Ruf, Kecskemet 1990. Those who worry that Exchange Slav positions are always drawn will take comfort from Black's victory in the game cited. 2b) However, Black can play more aggressively. Cs.Horvath-I.Sokolov,

232

The Queen s Gambit Declined: 5 1.f41 J

Nildic 1991 showed the correct way for Black to essay the Stonewall: S... ~4 9 .i.d3 f5 10 lbe2 (10 .!DeS .i.d7 II.!Dxd7 'it'xd7 12 f3 .!Dg5 Ih-lh Servat-Soppe, Buenos Aires 1993) 1O....i.f6 11 .i.h2 .i.d7 12 a3 .i.eS 13 b4 'ire7 14 .i.c2 ~hS 15 lbf4 g5 16 lbd3 f4 17 exf4lbxd4 Islbxd4 .i.xd4 19 0-0 .i.b5 and Black's attack proved decisive. Waiting moves are sometimes an unaffordable luxury; it is simpler to develop, and to cover e4. 8 .••b6 9 cxb6 cxb6 10 0-0 1) 10 lbe5 is an aggressive gesture, but in the end it leads only to simplification: 10... .i.b7 11 0-0 lbb4 12 .i.bl lbd7 13 .!Dxd7 'it'xd7 14 a3 lbc6 15 .!Da4.i.dS 16 'it'b3 lbaS 17 'it'b4 'irb5 G.Jacoby-Kordsachia, 2nd Bundesliga 19S3/4. 2) 10 %lcl .i.d7?! (10....i.b7) 11 h3 lba5 12 0-0 lbc4 13 'ire2 b5 14 b3 lbd6 15 .!DeS ± Bogoljubow-Spielmann, Triberg 1921. 10•••.i.b7 Black should have no problems with this position; he has duly reached the Exchange Slav a tempo ahead, White having spent an extra move with the c-pawn before exchanging. Admittedly the tempo disappears if Black plays the thematic ...b5, but Black's formation is still satisfactory. Practical experience from this position: 1) 11 h3lbb4?! (11...b5!?; 11...:CS) 12 .i.bl %lcS 13 .!Dd2lbc6 14 .i.g5 ;!; Petrosian-Pietzsch, Varna OL 1962, but White's edge is solely due to Black's loss of time. 2) 11 a3 h6?! (why? 11. .. %lc8!?; l1...b5!?) 12 %lcl.i.d6 13 .i.xd6 'it'xd6

14 b4 b5 15lbe2 %lacS 16.!Df4 e5 17 %lxc6 'it'xc6 IS dxe5lDe4 19lDd4 'it'c3 20 'irg4 with an imposing attack, E.Kovacs-l.Kadlicsko, Hungarian Cht 1993/4. Black must be careful to avoid being psyched out by symmetricality. 812) 7 ...b6(D)

=

Another principled move, with White achieving nothing after S cxb6 cxb6 (see above). Black's only potential worry is an immediate h4. 8 b4lbe4 1) Black gains nothing by temporizing with S... .i.b7?! 9 .i.d3 a5 10 a3 lbe4 11 .!Db5 lba6 12 cxb6 cxb6 13 bxaS bw 140-0 and White is slightly better, G.TItov-AHashim, Manila OL 1992. 2) S... lbc6!? seems a valid try, for example 9 %lbl bxc5 10 bxc5lbe4 11 lbxe4 dxe4 12 lbd2 f5 13 h4.i.f6 14 lbc4 'ire7 15 h5 %ldS 16 'ircl e5 17 dxe5lbxe5 ISlbxe5 .i.xe5 19 .i.c4+ ~f8 20 h6 g6 21 .i.xe5 .xe5 220-0 ~e7 23 .a3 .i.e6 24 .i.xe6 ~xe6 25 %lb7 ~ and play continues to be sharp even as the endgame approaches,

Miscellaneous Systems on Moves 5 and 6 A.Rustem-S.Begun, Belarussian Ch (Minsk) 1994. 9 ~e4 dxe4 10 tiJd2 .i.b7 1O... fS 11 Wb3 l&6 12 .i.c4 Wd7 13 ':dl gS 14 .i.g3 bS IS .i.e2 f4 16 lLixe4 fxg3 17 bxg3 .dS 18 .bl ':17 19 .i.f3 .fS 20 g4 .g6 21 lLixgS .xbl 22 ':xbl .i.xgS 23 .i.xc6 picks up a third pawn for the piece and keeps a promising position, A.SerebrjanikVeinger, Rishon Ie Zion 1992. 11 Wc2 f5 12 .i.c4?! Preparing to exchange the good bishop for a bad bishop, not generally advisable. White has several tries for an advantage: 12 .i.e2!?; 12 a3!? (intending cxb6); 12l&4!?, etc. 12•••.i.dS 13 .i.xdS exd5 14 lLib3 lLic6 15 a3 b5 16 0-0 .i.f6 lb._lIz Hamann-Hvenekilde, Danish Ch (Aalborg) 1965. 813) 7".lLihS?! (D)

The main effect of this is to increase White's grip on e5. 8 .i.d3 lLixf4 9 exf4 lLIc6 9... f5 10 0-0 lLic6 II ':el .i.f6 12 .i.c2tile7 13 .d2 with a lifeless position for Black, Gosti~a- Vebif, Kladovo

233

1990. The next few moves were painful: 13 ... lLig6 14 g3lLih8 15 h4lLi17 16 lLigS ±. 10a3f5 1O... aS 11 ':cl b6 12 cxb6 cxb6 13 lLia4.i.d7 140-0 g6 15 .i.b5 tile5 16 .i.xd7 lLixd7 17 .b3 ± LechtynskyKiss, Linz 1993. Black's light squares on the queenside are weak. 11 h4 .i.f6 12 lLie2 .i.d7 13 ':c1 'fIe7 14 hSlLid8 15lL1e5 .i.b5 16 Wb3 .i.xe5 17 fxe5 WgS 18 g3 lLic6 19 14 lLixd4 Otherwise Black loses a pawn. 20 Wb4 ~e2 21 fxg5 +- Kasparov-Speelman, Moscow rpd 1995. 814) 7".c6(D) 'Thus far and no farther.' Black tried a similar philosophy in V.Bukal jr-Keglevif, Croatian Cht (Tucepi) 1996, but after 7...bS 8 b4 c6 9 a4.i.b7 10 'ii'b3lLie4 11 axbS cxbS 12.i.xb8 .xb8 13 .i.d3 .i.c6 14 0-0 .i.d8 15 .i.xe4 dxe4 16 lLieS .b7 17 lLixc6 .xc6 18 d5 Black's queenside was leaking seriously. 8.i.d3 8 h3lLibd7 9 .i.d3 ':e8 10 0-0 .i.xc5 11 dxcS e5 12 lLixeS lLixeS 13 .i.c2 lLifd7 14 e4 lLixc5 15 exdS lLig6 16 .i.g3 ± I.Sokolov-Markos, K~ge 1997. The move order in this game was S...c6 6 e3lLibd7 7 h3 a68 c5 0-0. 8".lLibd7 9 b4 9 0-0 lLih5 10 ':el lLixf4 11 exf4 .i.f6 12 .c2 g6 13 b4 .i.g7 14 ':abl .c7 15 g3 f6 16 h4 f5 17lLia4 with a clear advantage for White, V.RagozinGoglidze, Moscow 1935 . 9...':e8 10 0-0 .i.f8 11 "'a4 bS 12 Wb3 lLih5 13 a4 lLixl4 14 exf4 bxa4

234

The Queen's Gambit Declined: 5 i.f4!

15 llxa4 ± Belamaric - Bazaj-Bockai, Croatian Cht (Makarska) 1995. The impression is that 7 ...llk6 is the only good reply to 7 cS (except perhaps for 7... b6 8 b4li)c6), but that it is good enough to equalize. White must look for something different. 82) 7 l:ct (D)

~xdS!? ~xdS 14 "g4 ~f6

IS l:fd1 16 e4 ± Blackburne-Burille, New York 1899, but the more modem plan of9 ...cxd41Oexd4bS 11 ~d3~b712 l:e1 ~b413 ~gS li)bd7 14 We2 Wb6 Ghinda-Ciocaltea, Baile Herculane 1982, is satisfactory for Black. 9.....xdl+ The 'gain' of tempo with 9 ...~xcS 10 "xd8 l:xd8 is illusory, since the rook soon becomes a target. 11 ~e2 ll)c6 12ll)a4 ~a7 13 ~c7 l:e8 (13 ...bS 14 ~b3 ±) 14li)b6 ± Van der VlietBartels, Netherlands tt 1996. ~e7

=

10l:xdl~xcS

B

With yet another transposition to the Malaniuk-Arlandi game given in Chapter 4.4, Line All, where White has llli)eS! li)bd7 12 ~e2;t. White is certainly squeezing more out of this than out of the 7 cS variations. 83)

7•••dxc4 7 ...c6 transposes to the ... c6 line. 7 ...li)bd7?! 8ll)eS c6 9 ~d3 l:e81O 0-0 ± Schlechter-Showalter, London 1899.

7"c2(D)

B

8~xc4cS S... bS 9 ~d3 ~b7 10 a4!? b4 11 li)bl ~d6 12 ~g3 li)bd7 13 li)bd2

l:eS 14llk4li)f8 15 li)aS ± Van der Sterren-Hort, Bern 1993. This illustrates one of the potential perils of an early ...bS; if a4 by White forces the pawn to advance, a lot of weak squares are created. 9dxcS Yet another central pawn clearance and queen exchange. 90-0 is effective after 9...bS 10 ~e2 c4?! 11 li)eS ~b7 12 ~f3 /Ods 13

White makes sure the queens stay on the board. Most of the positions that follow may also be reached via 6 ...li)bd7 7 Wc2 a6, or 6 ...li)bd7 7 "c2 dxc4 8 ~xc4 a6. Black can try:

Miscellaneous Systems on Moves 5 and 6 B31: 7•••llk6 B32: 7•••cS B33: 7•••M B34: 7•••lLlbd7 B35: 7•••dxc4

235 235 235 235 236

235

Wb2 i.c6 15 dxc5 ~xc5 16 0-0 ~6 =; White's a-pawn is just as weak as

Black's d-pawn. Preparing to besiege the hanging pawns with 10 0-0 ~bd7 11 :acl c5 12 dxc5 is more appropriate.

831)

7•..llk6 An independent possibility, but unconvincing. 8 a3 i.d7 9 h3 i.e8 10 i.d3 dxc4 11 i.xc4 h6 12 0-0 i.d6 Black's manoeuvring behind the lines has not been smooth. FioritoSoppe, San Martin 1994 went 13 i.e5?! lLld7 14 i.xd6 GO, but White should simply have completed development with 13 :adl i.xf4 14 exf4 ± with another Rubinstein Bind.

834) 7••• ~bd7(D)

832)

7•••cS This is a ...c5 Main Line, with ...a6 having been played unnecessarily early. 8 dxc5 9 a3 9 i.d6!? 9••.i.xcS 10 0-0-0 This can lead directly back to the New Main Line. There must surely be a case too for trying to take advantage of the omission of ...lbc6, with for example 10 lLlci2 ;to 10•••dxc411 i.xc4 ~c6 As played i\l the Yepez-Pau game cited in Chaptel3.5. 12 ~g5 is the recommended move.

.as

833) 7 ••• b6 8 i.e2 i.b7 9 cxdS exdS In Y.Rahman-Elsayed, Tanta 1997, White now played 10 a3 ~bd7 11 b4?! c5 12 bxc5 bxc5 13 :bl Wc8 14

8 adS A simple and very natural move, but there are several alternatives. 1) It is surprising that 8 c5 is not tried more often. In A.Hoffman-Varela, VIlla Ballester 1996, Black could find nothing better than the truly miserable 8 ...~b8?, leaving White two clear tempi ahead of the 7 c5 line. After 9 i.d3 ~c6 10 a3 h6 11 h3 lLld7 12 b4 Black was soon overrun. Stonewalling with 8 ... ~5 and ... f5 is much too slow, given Black's inability to shift a piece to e4 quickly. This leaves 8...c6, when 9 h3 would transpose into the Shulman-Lyrberg game given in Chapter 7.1, Line A3. White can of course substitute a developing move, possibly 9 i.e2 b6 10 cxb6 Wxb6 11 0-0 ;to 2) 8 O-O-O!? is an untried suggestion of Gelfand's.

236

The Queen's Gambit Declined: 5 i/4!

3) A few years after making this aggressive suggestion, Gelfand played the more restrained 8 a3 in GelfandNogueiras, Biel IZ 1993, play continuing 8 ... dxc4! 9 .i.xc4 bS 10 .i.a2 cS II :dl (11 dS exdS 12lLlxdS is acceptably met by either 12 ... lLlxdS Gelfand, or 12....i.b7 Nogueiras) 11...c4 12 e4 .i.b7 13 .i.bl g6 14 0-0 .b6 with unclear play. White now overpressed with 15 dS? exdS 16 eS lLlhS 17 .i.h6 d4!, running into a strong exchange sacrifice. 4) 8 :dl would transpose into the ... c6 system after S...c6. Alternatively, 8 ... b6?! 9 cxd5 exd5 10 lLles .i.b7 11 .i.e2 ± Gelfand, or S...h6 9 a3 dxc4 10 .i.xc4 bS 11 .i.a2 c6 12 e4 :eS 13 .i.bllLli8 14lLleS ± S.Ivanov-Talanin, Kstovo 1994. 8•••exdS 9 .i.d3 Good enough for an edge, but the uncompromising 9 0-0-0 may also be considered. 9...cS is probably the best reply, but 0stenstad-Mirza, Novi Sad OL 1990 continued 9 ...bS?! 10 lLles .i.b7 11 .i.d3 :c8 12 lLlxd7lLlxd7 13 .i.xh7+ *hS 14 .i.fS ±. 9•••cS 10 lLles 1) 10 g4 c4 11 .i.fS? (11 .i.e2 ;!; Hort) 11...g6 12 .i.xd7 .i.xd7 13lLleS lLlxg4! 14lLlxdS .i.h4 150-0 .i.e6 16 lLlxg4?! (16 e4 00 Hort) 16...•xdS =+= McCambridge-Hort, Dortmund 1982. 2) 10 dxc5!? lLlxcs 11 0-0.i.e6 12 lLld4 :cS 13 .i.fS bS 14 a3 .b615 b4! is also good for White, Porper-Lelchuk, Groningen 1995 (and if IS ... lLlce4, 16 lLlxdS!). 10•••:e8 11 0-0 lOf8 12l:adl .i.d6 13 .i.gS c4 14 .i.e2 .i.xeS 15 dxeS :XeS 16 f4 .i.f5 17 Wet :e818lLlxdS ± Kishnev-Steinbacher, Bern 1992.

No doubt Black's defensive play in this game could have been improved, but this does not change the general verdict that White, not surprisingly, has several ways to achieve a slight edge after 7...lLlbd7. 835) 7 •••dxc4 8 .i.xc4 (D)

B

8...lbbd7 8...bS 9 .i.d3 (9 .i.e2!?) 9 ....i.b7 10 0-0 lLlbd7 11 :fdl, and now Korchnoi-Short, Monaco rpd 1993 continued 11...l:cS? 12lLleS 'ire8 13lLle4 ±. A touch of quickplayitis perhaps? 11 ....i.xf3 12 gxf3 lbhs is much more challenging, and a good reason for White to have retreated his bishop to e2. After 13 .i.xh7+ *hS, 14.i.e4lbxf4 IS .i.xa8? (15 exf4 00) IS ...lbh3+ 16 *g2 .xa8 17 *xh3 'irxf3# would have been a particularly humiliating finish. Of course this mating line can be avoided, but if 14 .i.eS, Black has good compensation for the pawn after 14... fS 15 .i.g6 .i.d6, e.g. 16 .i.xhS .gS+ 17 *hl .xhS IS f4lLlf6. 90-0 1) 9 .i.e2!? sets the positional trap 9...bS? 10 a4 b4 I1lLle4lbds 12 :cl

Miscellaneous Systems on Moves 5 and 6 ~xf4 13 exf4 i.b7 14 0-0 ± Maiwald-Danielian, Vejen jr Ech 1993. Black should reply 9 ...c5!? 10 dxc5 ~xc5 lU[dl WaS. 2) 9:ctl and now: 2a) 9 ... i.d6 10 i.e5?! (10 0-0 1) 10.. :.e7 11 0-0 b5 12 i.d3 i.b7 ; 13 a3 i.xf3 wins a pawn, R.David-Malich, 2nd Bundesliga 199516. 2b) 9 ... b5 10 i.d3 (10 i.e2!?) 1O...i.b7 11 ~g5 h6 12 ~ge4 ~5 13 i.g3 f5 14 i.e2 ~xg3 15 ~xg3 b4 16 'IIb3 ~h8 GO Sturua-Danielian, Protvino 1993. 9••• b5 10 i.e2 i.b7 11 :fdl We8 12c!M2c5 12...:teS!? 13 dxc5 ~c5 14 b4 lOcd7 15 a3

:c8 16 Wd3 ~b6 17 :ad :d8 18 f t l :d7 19 e4 1 Gavrilov-Moroze-

vich, Russian Ch (Orel) 1992. This is quite characteristic of the i.f4 system; White keeps an edge in piece mobility in a position with a symmetrical pawn structure.

84) 7 adS (D)

237

7•••~dS 7 ...exd5 and now: 1) 8 i.d3 ~bd7 90-0 :eS 1O:tel ~f8 11 'IIb3 (11 ~e5 i.d6 12 i.g3 c5 13 i.bI 112- 112 Lechtynsky-Mozny, Prague 1985) 11...~e6 12 i.e5 c6 13 h3 i.d6 14 ~4 and White has a slight advantage, Lugovoi-Margolin, Kstovo 1994. 2) 8 h3 i.f5 9 i.d3 i.xd3 10 'IIxd3 i.d6 11 i.xd6 'IIxd6 12 0-0 ~bd7 112-117. Naumkin-Dydyshko, St Petersburg 1992, was presumably a 'Russian draw'. 8~dSWxdS

8 ...exd5 9 a3 c5 10 dxc5 i.xc5 transposes to the ...c5 Exchange Variation (6...c5 7 dxc5 i.xc5 S cxd5 ~xdS 9 ~xd5 exdS 10 a3), but with Black having played the unusuall0 ... a6. 11 i.d3, threatening i.xh7+, should give White a slight edge, but in KnefevicMcCambridge, Neskaupsstadur 1984, White was content with an early handshake after 11 :cl i.b6 12 .i.e2 ~6 130-0 :e8, etc. 9 i.xc7 i.b4+ 10 ~2 i.d7 11 a3

:c8

Black seems to have adequate compensation for the pawn, for example 12 i.xb8 i.xd2+ 13 "xd2 :axb8 14 f3 e5! Agzamov-Sr.Cvetkovic, Belgrade 1982, or 12 i.g3 i.xd2+ 13 'IIxd2 i.a4 14 :cl ~d7. Black's chances are clearly better than in the similar but unexplored line with 6 ...b6 7 cxd5 ~xd5 8 ~xd5 'IIxd5 9 i.xc7, in that no weakness has been created along the long diagonal. Solid enough, but it makes sense to insert 7 'IIc2 first.

White's best chances for an edge in the 6 ...a6 line are with the continuation 7"c2.

238

The Queen's Gambit Declined: 5 1./41

8.2 6...lLlbd7: Lines not already examined 1 d4 d5 2 c4 e6 3 M lM64lDfJ ~e7 5 ~f4 0-0 6 e3lDbd7 (D)

w

Generally Black will play an early ... c6 or ... a6 here, leading to lines considered elsewhere. This section considers the independent alternatives after the following moves: A:71i'c2 238 B: 7 cS 240 C: 7 adS 241 D: 7 h3 242 E: 7 ~d3 243

F: 7 a3 G: 7:c1

244 244

A) 71i'c2 This is the most common and transpositional of all the possibilities, and is well suited to give an edge against either 7 ... c6 (Chapter 7.1) or 7 ... a6 (Chapter 8.1). 7••.cS!? This is an interesting defensive try for Black. In comparison with the 6 ... c5 variations, Black has the added

option of recapturing on c5 with the knight, leaving the bishop on its best defensive square. In comparison with 5 ... c5 6 dxc5lDa6 7 cxd5 exd5 8 e3. White has committed his queen to c2 a little too early. Other moves: 1) 7 ...b6? leads to disaster on c7: 8 cxd5 exd5 9 lDb5 ~b4+ (9 ...c5 10 ~c7 "e8 IIlDd6 wins the exchange) 10 ~dl ~b7 (1O...c5 II ~c7 again picks up the exchange) II a3 ~e7 12 ~xc7 Alvarez-Bello, San Sebastian 1994. 2) 7 ...dxc4 8 ~xc4 and now: 2a) 8 ...lDb69~b3lDbd510lDxd5 lDxd5 11 0-0 (offering the Rubinstein Bind is unnecessary here, as the straightforward II -*.g3, with e4 in mind, keeps White in control in the centre) 11...lDxf4 12 exf4 b6 13 :adl ~b7 14 lDe5 -*.d6 15 g3 g6 16 :fel l:cS 17 h4 c5 IslDxf7 ~xf7 19 -*.xe6+ ~g7 20 "c3 "f6 was Zaichik-Azmaiparashvili, Tbilisi 1979. ECO ends the game sequence here with an enigmatic 'with compensation for the material' symbol, without making it clear who is supposed to be down in material after 21 ~xc8 or the possibly stronger 21 dxc5. 2b) S... c5 9 dxc5 -*.xc5 is unconvincing, if only because it leaves White a tempo ahead of 6 ... c5 7 dxc5 ~xc5 S -*.e2 dxc4. See Chapter 5.2, Line A52. White's most dangerous plan would probably be 10 :dl, and if 10...We7, 11 a3 (threateningb4) 11...a5 12 -*.a2, etc. Instead, after 100-0 "e7 II e4 e5 12 -*.g5lDb6 Black is back in the game, Kramnik-Andersson, Monte Carlo blindfold 1997. Returning to the position after 7...c5 (D):

Miscellaneous Systems on Moves 5 and 6

Scb:eS Or 8 cxdS: I) 8...exdS 9.i.e2 a6 10 dxcS ~cS II 0-0.i.g4 12 ~4 ~ 13 .i.e5 .i.xe2 14 "xe2 1017 IS .i.g3?! (1S ~xe6 fxe6 16 .i.g3 ;1;) IS ... ~xd4 16 exd4 ~f6 17 'Wf3 .i.b4 = Parker-H.Hunt, British Ch 1993. 2) 8 ... cxd4!? and now 9 exd4 ~xd5 10 ~xdS exdS II .i.e2 ~f6 12 0-0 .i.g4 13 'iVb3 .b6 was Kaidanov-Campora, Tilburg 1993. At flI'St glance 9 d6100ks strong, but 9 ...dxc3 10 dxe7 .xe7 11 bxc3 (11 .xc3lO1s =) 11...eS 12 .i.gS ~S gives Black a very solid, perhaps even advantageous, position. 8•••lDxeS 9 .i.e2 White can probably do better than this. I) 9 a3 .i.d7 10 .i.e2 dxc4 II .i.xc4 :c8 12 0-0 ~a4? (an over-elaborate continuation; 12... a6 keeps a perfectly playable game) 13 ~xa4 bS 14 ~S .i.e8 IS :fdl 16 b4 'l'xa4 17 .xa4 bxa4 18 .i.a6 with a distinct advantage for White, S.lvanov-Grabarczyk, Lubniewice 1995. 2) At the risk of being accused, jocularly or otherwise, of advocating queenside castling as the universal

=

.as

239

panacea, I draw attention to the possibility of 9 O-O-O!? After 9 ...•aS: 2a) 10 1012 .i.d7 II g4! ~xg4 12 cxdS bS (12...~xf2?? 13 ~4) 13 :gl!? with the sort of messy and double-edged position we saw in Chapter 3. With a mixture of relief and regret, at this stage of the book, I leave it to the reader to analyse. 2b) Indeed, as I make the final corrections to this book, I note that 10 ~gS!? has appeared in practice: 2bl) 1O... g6?! II h4 bS 12 cxbS ~hS 13 .i.e2 ~xf4 14 exf4 .i.b7 IS c,tbl .i.f6 16 hS .i.xc3 17 hxg6 hxg6 18 bxc3 ~e4 (18 ...~a4 19 :h3 :fc8 20 :dhl ~xc3+ 21 c,tal {21 'iVxc3?? .xc3 and h8 is covered by the queen} 21...c,tfS 22 ~xr7 c,txr7 23 :h7+ c,tf6 {23 ...c,tfS 24 :xb7 +-} 24.xg6+! c,txg6 2S .i.d3+ ~e4 26 :lh6+ c,tfS 27 :r7+ c,tg4 28 .i.e2# would have been a far more memorable finish) 19 ~xe4 dxe4 20 'W'b2 eS 21 c4 :re8 22 fxeS 1-0 Notkin-Shur, Moscow Ch 1997. 2b2) 1O...h6 seems critical; then perhaps II h4 hxgS 12 hxgS ~fe4 13 f3. Now 13 ...f6 may be met by 14 fxe4 fxgS IS .i.eS dxe4 16 c,tbl! and the broken kingside pawns are more important than Black's temporary extra pawn, e.g. 16...00 17 .i.g3 or 16...:t5 17 .i.d4 eS 18 lOIs .d8 19 .i.xc5 .i.xcs 20 g4 followed by 21 'iVxe4. 13 ....i.xgS! gives unclear play though. 9•••cb:c4 10 .i.xc4 86 11 84 b6 11....i.d7 120-0"aS 13 :fdl :fd8 14.i.fl .i.e8 with equality, Mirall~s­ Spassky, French Ch 1991. 1Z 0-0 .i.b7 13 'iVe2 'iVe8 14 :fdl 'iVc6 with equality, Stohl-Gavrilov, Pardubice 1995.

240

The Queen's Gambit Declined: 5 j"f4!

B) 7 c5 (D)

2b) 8 lt1xe4 dxe4 9 lt1d2 is a better option for White. 3) 7 ...lt1hS 8 .i.d3 lt1xf4 (8 ... g6 9 0-0 c6 10 b4 b6 11 .i.h6 l:le8 12 e4 'ilc7 13 :el ± Portisch-Donner, Halle 1963; 8... c6 leads back to the main line) 9 exf4 b610b4j,b7 11 0-0c612 'ilc2 g6 13 g3;!; Llanos-Cordero, Buenos Aires 1993. 8.i.d3~

With the disadvantage for White, when compared with 6... a6 7 cS, that Black has not played the weakening ...a6, but with the advantage that Black no longer has ...ltX6. This is one of the few variations where relatively little, apart from a few analytical corrections, needs to be added to ECO. 7..•c6 1) 7 ...b6 8 b4 (but not 8 c6?! lt1b8 9 lt1eSlt1e4 10 lt1xe4 dxe4 11 .i.c4.i.b4+ 12 f1 .i.d6 13 dS 'ile7 =+= PortischForintos, Hungarian Ch 1965) 8... as 9 a3 bxc5 10 bxc5 ~ 11 .i.d3 ;!; ECO. 2) 7...~!? and then: 2a) 8 'ilc2 g5 9 j,g3 f5 10 h3· lt1xg3 11 fxg3 and now: 2al) 11...lt1f6 12 .i.d3 lt1e4 13 g4 lt1g314:h2(14:gl!?;!;ECO) 14...c6 ISlt1e2'i1aS+ 16~f2~ 17~gl;!; Forintos-Csom, Hungary 1969. 2a2) ll...b6!, not mentioned by ECO, is promising. If 12 b4, Black has the choice between the enterprising piece sacrifice 12...a5 13 c6 axb4 14 cxd7 .i.xd7 15 lt1dl :a3 co, and the simple and safe 12...bxcS 13 bxc5 c6, with ideas of ...e5 at some stage.

1) 8 ...l:le8 and then not 9 h3? .i.xc5! 10 dxc5 e5 ; Steinitz-Napier, New York 1897, but 9 j,g3 j,fS 10 O-O;!;ECO. 2) 8...b6 9 b4 bxcS 10 bxcS 'ilaS 11 'ild2 ;!; .i.a6? 121t1xd5! won a pawn in Antoshin-Uusi, Moscow 1967. 9 ()..O This by now should be a familiar idea to readers of this book. Whether or not it is historically 100% correct to talk of a 'Rubinstein Bind' (and Janowsky at least partially anticipated the concept against Pillsbury at Hastings 1895), Steinitz was apparently unfamiliar with the idea Steinitz-Chigorin, Nuremberg 1896 continued 9 j,e5?! f6?! 10 .i.g3?! f5?! 111t1g5 lt1df6 12 .i.eS ;!;. This sequence passes without comment in ECO, but there is much that is open to question. 10 lt1g5!, an obvious continuation of the attack, is not even given as a possibility, and yet it poses Black some difficult questions. The attack is extremely strong after 1O... fxe5 11 'ilxh5! (11 .i.xh7+? ~h8 12 'ilh5 j,xgS) 11...j,xgS (or 11...lt1f6 12 .i.xh7+ ~h8 12 'ilh4) 12 'ilxh7+ ~f7 13 j,g6+ ~f6 14 .i.h5 .i.h6 IS g4! with the idea of g5+. White's attack does not win outright, as I had first assumed, as Nunn points out the idea of 15 ...~e7 16 g5 :h8 16

Miscellaneous Systems on Moves 5 and 6 Wg6 WfB. White is about to regain his material though, and ought to have excellent prospects of keeping a substantial positional advantage. For example, one line suggested by Nunn, 17 :gl!? exd4 18 exd4 ~d8 19 gxh6 :xh6 20 Wgs+ ~f6, claimed as only a slight advantage for White, actually looks extremely difficult for Black to hold after 21 .i.e2. It is the type of position that French Defence players have nightmares about; bad bishop, weak dark squares, backward development, and open kingside lines for White. 1O...We8! (Burgess) is the one realistic attempt to hold the balance, or prove that White has over-pressed. After 11 .i.xh7+ ~h8 12 Wc2 fs (12 ... fxes? 13 .i.g6 Wd8 14 ~xe6; 12... fxgs 13 .i.g6 Wd8 14 .i.xh5 ±), the critical move is 13 ~xe6!, based on the tactical theme that if 13 ... tLlxe5 14 tLlxf8 .i.xfB 15 dxe5 ~xh7 16 g4! the black knight is trapped. After 13 ...~xh714tLlxfB+tLlxfBI5~2the position is unbalanced, with White having rook and two pawns for two minor pieces. White has the straightforward plan available of castling queenside and rolling his kingside pawns; it is not so clear where Black gets his play. So 10 ~gs is not quite as massive a missed opportunity as I first thought, but it is still promising, and ought to be noted. Of the other queried move~, ECO notes that 9 ...fs (instead ot; 9 ... f6) would be equal, while a move) later 1O...~xg3 makes more sense. An' interesting historical and analytical digression, but really 9 O-O! is the logical and thematic move. 9...f5

241

9 ... tLlxf4 10 exf4 b6 11 b4 bxc5 12 bxcs Wa5 13 Wc2 g6 14 :fel .i.f6 15 tLles ;!; Kne!evic-D.Jano§evic, Vdac 1977. 10b4! 10 h3 ~6 11 ~s tLlxes 12.i.xes ~7 13 .i.h2 es 14 dxes tLlxcs 15 .i.c2 ~4 Chistiakov-Neishtadt, USSR 1956. 10...8611 84 11 h3!?;!; Gulko. 11••• h612 b5 12 ~es ;!; Gulko. 12...axb5 12... gs? 13 bxc6 bxc6 14 ~es ± Gulko. 13 axb5 :0114 1Wxal gS 15 bxc6

=

~xcS

Is ... bxc6 16 .i.d6 .i.xd6 17 cxd6 tLldf6 18 Wa3 ~e8 19 ~5 Wxd6 20 'ii'cs! ± Gulko. 16 c7 1We8 17 .i.b5 .i.d7 18 .i.e5 ± Gulko-Lautier, Dos Hermanas 1994. Maybe this is White's most promising try against 6 ... tLlbd7.

C) 7 adS Keeping to strategically welltrodden paths. 7...lOxdS (D)

242

The Queen's Gambit Declined: 5 JJ.f4!

For once, the recapture with the knight might be slightly questionable, as the other knight has committed itself to d7, depriving Black of the thematic plan of ... c5 followed by ...ttk6. Despite this, we shall take the knight recapture as our main line, rather than 7 ...exd5 since in that case Black almost invariably plays ...c6 or ... a6 at the earliest opportune moment, transposing into lines already considered in Chapter 7 or S.I. SlBxd5 S Jld3 c5 9 Jlg3 cxd4 10 exd4 lD7f6?! (1O....a5 =) 11 0-0 lDh5?! 12 Jle5lDdf4 13 Jlc2lDg6 14 'ifd3 with a clear advantage for White, JessenStokstad, Copenhagen 1991. S•.•exdS 9 Jld3 The cautious 9 a3 has been tried: 1) 9...c6 10 Jld3 :eS 11 .c2lDfB 12 b4 a6 13 0-0 Jld6 14 Jlxd6 .xd6 15 :fc1 :e6 16 .c5 .dS 17 Jlf5 :h6 IS JlxcS :xcS 19 a4 ;I;; Tisdall0stenstad, Norwegian Ch (Namsos) 1995. 2) 9 ...c5 (via 7 a3 c5 S cxd5lDxd5 9lDxd5 exd5) 10 Jld3 cxd4 11lDxd4 ttk5 120-0 Jlf6 13lDb5 Jle614 Jld6 lDxd3 15 .xd3 :eS 16 Jlg3 :fB 17 :tac1 .b6 IS ttk7 :tacS 19 b4 and White is slightly better, Ehlvest-Krogius, St Petersburg 1994. 3) 9 ... Jld6!? 10 Jlxd6 cxd6leaves Black considerably better off than in the Meduna-Prandstetter game below. With this particular structure, a3 comes across as a weakening move. 9•••cS Others: 1) The most solid move is perhaps 9 ... c6, which we have considered under ... c6 systems.

2) 9...lDf6 10 .c2 c6 is a similar transposition. 3) 9... Jld6 10 Jlxd6 cxd6 11 0-0 lDf6 12.b3 .e7 13lDd2 Jle6 14 f3 lDh515:acl f516:c3l:acSI7:fcl ;I;; E.Meduna-Prandstetter, Trnava 19S1. 4) 9 ... Jlb4+ 10 lDd2 c5 11 0-0 (11 Jld6 :eS 12 dxc5 ~xc5 {but not 12...lDxc5? 13 ~xc5 ~xc5 14 Jlxh7+} 13 ~xh7+? ~xh7 14 .c2+ ~g8 15 ~xc5 .c7) 11...c4 12 ~c2 .a5 13 e4?! (13 lDf3 ;1;;) 13 ... ~xd2 14.xd2 "xd2 15 Jlxd2 dxe4 16 .txe4lDf6 17 .tf3 :d8 18 .tc3 lDd5 = TormaReiter, Hungarian Cht 199213. 100-0 10 dxc5 lDxc5 =Ftai:nik. 10•••c4 1O... b6, Ftai:nik, leads to the 6 ... b6 variation, Chapter 6, Line C. White then has 11 dxc5!? ;1;;. 11 .tel lDf6 12 lDe5 .te6 13 b3 cxb3 13 ... b5 14 bxc4 bxc4 15 ttk6.d7 16lDxe7+ .xe7 17 l:bl ± Ftacnik. 14 axb3lOd7 15ftl fS 16lM3 16 b4! ± Ftai:nik. 16•••85 17 l:c1 ;t Ftai:nik-Prandstetter, Czechoslovak Ch 1980.

D) 7h3 (D) Hoping that Black will do the gentlemanly thing and play 7 ...c6. 7•..cS 1) 7 ...lDe4!? 8 cxd5lDxc3 9 bxc3 exd5 10 .b3 c6 11 ~d3 lDf6 120-0 ~d6 13 .txd6 .xd6 14 c4 :bS 15 cxd5 lDxd5 16 :acl .h6 17 :fel .te6 112_112 Z.Ilic-Lein, Saint John 1988. White may have improvements, but it is unlikely that there will be anything leading to more than a tiny edge.

Miscellaneous Systems on Moves 5 and 6

243

=

'iVb6 13 'iVb3lDf6 Agrest-Gavrilov, St Petersburg 1993. 7 h3 does not seem particularly dangerous; the same goes for the various odds and ends described below.

E)

With 7 h3, White has discouraged the move ... lDhS, but done nothing to prevent ...lDe4. 2) 7 ... b6 S cxdS exdS will tend to leave Black defending hanging pawns. 9 .te2 .tb7 lO 0-0 lDe4 11 :cl cS 12 dxcS lDxc3 13 l:txc3 bxcS 14 "ii'a4 lDb61S .bS .d716 .xd7lDxd7 17 l:tdllDb6 18 .teS;t Ibragimov-Kozhevin, Podolsk 1993. 3) 7 ...dxc4 S .txc4 and now: 3a) S....td6 9 .txd6 cxd6 100-0;t K.Len-Denisov, Podolsk 1993. Black has, in the event of ... dS, the wrong bishop and a misplaced knight. Without ...dS, Black is passive. 3b) S...cS!? 9 dxcS (9 O-O!? cxd4 lO exd4lDb6 11 .td3lDbdS 12 .teS Psakhis) 9 ....txcS 10 0-0 (10 a3!? would discourage Black's next move) 10...•e7 11 e4 eS 12 .tgS lDb6 (or 12... h6!? 13 lDdS "ii'dS 14 lDxf6+ lDxf6 IS "ii'xdS :xdS 16 .txf6 gxf6 17 :acl .tb6 IS .tdS :b8 Psakhis) 13 .tb3 :d8 14 "ii'e2 h6 IS .txf6 .xf6 with equality, Psakhis-Saidy, Waikiki 1997. 8 cxd5 lDxdS 9lDxd5 exd5 10 .te2 lO dxcSlDxcs 11 .te2 1/Z-1f2 HenleyAndersson, Indonesia 1983. 10•••cxd4 11 exd4 'iV85+ 12 .td2 CC)

7.td3 7 .te2 dxc4 8 .txc4 is equivalent. 7•••dxc4 There is also 7 ...cS 8 cxdS exdS 9 0-0 a6 lO lDeS lDb6 11 1 .te6 12 .tbl :c8 13 .tgS Ill-liz Ivkov-Spassky, French Cht 1993. 8 .txc4 (D)

:c

8•••86 8 ...cS 90-0 a6 10 a4 transposes. 984 90-0 bS 10 .td3 (lO .tb3?! .tb7 11 'iVe2 cS 12 .tc2?!"ii'b6 was slightly better for Black in P.Johner-Janowsky, Karlovy Vary 1907) 1O....tb7 11 :cl cS 12 .e2 .b6 13 :fdl :acS 14lDeS ):fdS Verbaere-Lepelletier, Aixles-Bains 1991. 9.••cS 100-0 (D) lO dxcS lDxcS 11 "ii'xdS :xdS 12 ~e2 .td7 13lDeS .teS M.Roiz-Grlgoriants, Sao Laurenco jr 1995.

=

=

244

The Queen's Gambit Declined: 5 J.j4!

=

10 a4 cxd4 ECO) 8 ...~xc5 9 .i.e2 (9 cxd5 ~xd5! 10 ~xd5 exd5 00) 9...dxc4 (9 ... a6 10 cxd5 exd5 11 0-0 b5?! 12 ~d4 ;t and now 12 ... .i.d6? loses a pawn to 13 ~c6 ti'd7 14 .i.xd6 ti'xd6 15 ~xd5, D.Chevalier-Jojic, Paris 1992) 10 .i.xc4 a6 11 ~5 ti'xdl+ 12 IIxdl b5 13 .i.e2 .i.b7 Larsen-Portisch, Palma de Mallorca 1967.

=

10•••~b6?! This is slightly casual. Black should play 10... cxd4 first, as White cannot recapture with the knight. Then 11 exd4 is equal. 11 .i.b3 cxd4 12 ~xd4! .i.b4 13 "f3 .i.xc3 14 bxc3 ~bd5 15 .i.e5 ~xc316I1ac1

8.3 Miscellaneous 6th moves for Black 1 d4 d5 2 c4 e6 3 ~c3 ~r6 4 ~f3 .i.e7 5 .i.f4 0-0 6 e3 (D)

B

With a massive initiative in return for the pawn, Relange-Bricard, French Ch 1997. F)

7 a3 may quite legitimately be met by 7 ...dxc4 8 .i.xc4 a6, going into Line E a waiting move down; White's extra tempo would of course disappear completely after 9 a4. H.Loebler-C.Singer, Austria 1995 continued 9 ti'c2 (9 0-0 b5 10 .i.a2 00) 9...c5 10 dxc5 .i.xc5 11 b4?! .i.e7 12 0-0 b5 13 .i.d3 .i.b7 14 e4 IIc8 :;:. If 7 ...c5, 8 cxd5 exd5 (8 ... ~xd5 9 ~xd5 exd5 10 .i.d3;t; see Ehlvest-Krogius in Line C above) 9 .i.e2 a6 10 dxc5 ~xc5 11 0-0 .i.e6 12 .i.e5 ;t Benko-A.Medina, Palma de Mallorca 1968.

It is doubtful whether any of the lines below will catch on, but Black's battle honours include a draw against an admittedly very youthful Kasparov, and a win against Dreev. The lines are: A: 6...h6 244 B: 6 •••1Ie8 245 c: 6•••~c6 245 245 D: 6 ••• ~h5 E: 6•••dxc4 246

G)

A)

7 1Ic1 c5 (7 ...c6!? is perhaps the safest) 8 dxc5 (8 .i.d3 dxc4 9 .i.xc4 a6

6 ••• h6 is probably most simply answered by 7 "c2, and waiting to see

Miscellaneous Systems on Moves 5 and 6 what Black does next. Instead, Kasparov-Klovans, Minsk 1978 continued 7 cxdS lrurdS 8/t)xdS exdS 9 .i.d3 cS 10 liJes cxd4 11 exd4 .i.b4+ 12..tn ti'f6 13 g3 /t)c6 14 lrurc6 ti'xc6 IS ..tg2 .i.e6 Ih-1h. In the system that White chose, Black's ... h6 was actually useful, side-stepping any threats to the h7-pawn.

245

3b) If 7 ... a5, then 8 cS!, and if Black plays ... b6 at some stage, then his queenside light squares will become weak after White exchanges on b6.

D) 6•••/t)hS (D)

B)

6 •••:le8 gained Black about half a tempo on the main lines after 7 h3?! cS 8 dxcS .i.xcs 9 ti'c2 /t)c6 10 a3 in J .Annas-de Annas, Sagua la Grande 1989. Again, White should choose 7 "c2!?, and see what Black plays.

C) 6 •••/t)c6 is directed against 7 ti'c2, when 7 .../t)b4 is known to be satisfactory for Black (see discussion of 6 ti'c2/t)c67 e3 in Section 8.4 below). Other moves: I) 7 h3 (quiet play allows Black to equalize) 7 ...dxc4 8 .i.xc4 .i.d6 9 .i.xd6 cxd6 10 dS /t)eS 11 /t)xeS dxeS 12 dxe6 ti'xdl+ 13 :lxdl .i.xe6 14 .i.xe6 fxe6 = Blagojevic-Z.Vukovic, Yugoslav Ch 1990. 2) 7 /t)eS /t)xeS 8 dxeS liJe4! 9 /t)xe4 dxe4 is satisfactory for Black. 3) 7 a3 is probably best, preparing cover for the queen to come to c2: 3a) 7 ... a6 was followed by 8 :lel /t)hS 9 .i.g3 g6 10 .i.e2/t)xg3 11 hxg3 .i.f6 120-0 liJe7 13 ti'c2.i.g7 14 :lfdl c6 IS cS fS =in Ljubinkovic-Z.Vukovic, Yugoslav Ch 1992. However, it seems odd for White to prepare ti'c2, and then not play it for several moves. 8 ti'c2 is more thematic, when White should be able to play for an edge.

If this move is effective, White can give up on .i.f4 immediately! Of course, it's not so simple, as Black is falling behind in development. 7.i.e5! 7 .i.g3 g6 (7 ...c6?! should be met by 8 It)eS! ± rather than 8 cxdS exdS 9 .i.d3 g6 ao Saraiva-H.Mira, Erevan worn OL 1996) 8 ti'c2 b6 9 :ldl .i.b7 10 cxdS exdS 11 .i.e2 /t)d7 12 0-0 :le8 = S.lvanov-Bezgodov, Russian Ch rpd (Elista) 1997. 7•••f6 7 .../t)d7 8 cxdS Ih-Ih Yuferov-Varavin, Krasnodar 1991, is another 'Russian draw' . It would be interesting to know how Black was intending to pr0ceed, as he is losing a pawn. 8 .i.g3 lrurg3 9 bxg3 f5 The Stonewall formation might just about be justified if White had already castled. The prospect of White playing

246

The Queen's Gambit Declined: 5 i.J4!

g4 (twice if necessary) and attacking along the h-file is a little bit harder to bear. 10 tLle5 c6 11 g4lDd7 Creating a dreadful impression, but Black's position has already gone, for example 11...g6 12 gxfS exfS 13 cxdS cxdS 14 tLlxdS winning a pawn. 12 gxf5lM6 13 adS adS 14 fxe6 .txe6 15 .td3 +- Nasybullin-Momagadze, Leningrad Cht 1989. Perhaps more catastrophic than it need have been, but this variation is undeniably difficult for Black. E)

6 •••dxc4 7 .txc4 (D)

B

advantage for White, SavchenkoCepon, Bled 1996, or 9 ... lLlbd7? 10 lLlbS tLlb6 11 .tc7 "e8 12 .txb6 axb6 13 tLlc7 +- Dubinka-Skripova, St Petersburg 1996. SlLleS! The most vigorous move, although there are reasonable alternatives. 1) 8 .td3! tLlbd7 (8 ...bS?! 9 a4 bxa4 10 0-0 a3 11 bxa3 tLlbd7 12 e4 ± Psakhis-Pelikian, New York 1997 too much time-wasting by Black) 9 e4 cS 10 eS tLlhs 11 .i.e3 cxd4 12 tLlxd4 g6 13 .te2?! (this looks like a miscalculation, whereas 13 f4100ks like a Sicilian that has gone badly wrong for Black; then Nenashev gives 13 ....tcS 14 .te2 "h4+ as + in Informator 55, without offering any suggestion as to how to meet IS g3lLlxg3 16.i.n winning a piece) 13 ...tLlxeS 14 f4 (14 .txhS gxhS IS 'ii'xhslLld3+) 14...tLlc6 IS .txhS gxhS 16 'ii'xhS tLlxd4 17 0-0-0.tf6 18 :d3 lbb3+ 19 ~c2 20 "g4+ ~h8 21 axb3 with unclear play, Dreev-Nenashev, Manila OL 1992. 2) 8 a4 cS 9 0-0 cxd4 10 tLlxd4 .td7 11 e4 tLlc6 12 eS tLlxd4 13 'ii'xd4 lLIhS 14 .i.e3 .tc6 IS .xd8 :axd8 16 .te2 (16 g4? tLlf4 17 .i.xf4 :d4 +) 16 ... g6 and the game is heading for a draw, Aleksandrov-Danielian, Voskresensk 1993. 3) 8 .e2 lbbd7?! (S ...bS) 9 a4 cS 10 dxcS lbxcs 11 0-0 .td7 12 lbes .e8 13 e4 ± Llanos-A.Hoffman, Buenos Aires 1993. S•••b5?! S... cS?! 9 .f3 cxd4 10 O-O-O! tLlbd7 (lO.....c?? 11 tLlg6 .xc4 12 lbxe7+ ~hS 13 :xd4 .cS 14 .i.d6 is winning for White) 11 exd4 leaves

"as

Swprisingly, 6 ...dxc4 is more popular for Black than the more interesting S...dxc4, discussed in Section 8.S below. After this loss of tempo, White's lead in development is obvious. 7 ...a6 If 7 ...cS, White's simplest is 8 dxcS, with the idea of entering the 6 ... cS 7 dxcS .txcS 8 .te2 dxc4 variation a tempo ahead. 8 0-0 cxd4 9 exd4 is also eossible, for example 9 ... tLlc6 10 a3 llldS 11 .txdS exdS 12 lLleS lLlxeS 13 .txeS .te6 14 'ii'f3 leads to a slight

Miscellaneous Systems on Moves 5 and 6 White several moves ahead in development. 8...cUs 9 .tg3 ;t is probably Black's best. 9 1Wf3 :87 10 .td3 .tb7 11 1Wh3 ll:)bd7 12 :dl g6 130-0 1WaS 14.tgS :e8 IS 1Wh4 and White is clearly better, Kharitonov-Danielian, Voskresensk 1993.

8.4 Miscellaneous 6th moves for White 1 d4 dS 2 c4 e6 3 ll:)c3 ll:)f6 4 00 .te7 S .*.f4 0-0 (D)

w

We can quickly dispose of 6 IOb5? .*.b4+ 7 lD
247

A)

6cS?! This has been played by a reigning World Champion, but is best described as an interesting move played a move too early; it should wait until Black has played ...ll:)bd7 or ...a6. 6",b6 6 ... lOe4?! 7 'iVc2 f5 8 e3 a69 .*.d3 lOc6 10 a3 .*.d7 (10 ... g5 first is more accurate) 11 h3.*.e8 12 g4;t PlaskettA.Gal, Natanya 1987. 7b4aS Black must break up White's pawns as quickly as possible. 8 allOe4 9 :c1 9 ll:)xe4? dxe4 10 lOe5 f6 11 lOc4 axb4 12 axb4 :xal 13111xallOc6 + Em.Lasker-Schallop, Nuremberg 1896; another dodgy position that the great Emanuel managed to win. 9"..*.86 9...axb4!? 10 axb41O:6 11 b51Da7! -King. 10 1Wc2 ll:)xc3 11 "xc3 axb4 12 axb4lOc6 13 e3 .txf114 ~ bS IS :al Plaskett-King, British Ch (Plymouth) 1989. King now suggests 15 ... f6, when Black is starting to take the initiative. The game continued 15 ...1IId7 16 ~e2?! (16ll:)e5 King) 16...f6 :j:.

=

B)

6 adS Another Plaskett try, which maybe deserves a second glance. 6",lbxdS 7 lbxdS exdS (D) 8"c2 8 e3 leads back towards standard lines. There is, however, the unexplored possibility of 8 g3!?, which is suitable

248

The Queen's Gambit Declined: 5 1../41

C) 61lct (D)

w B

if White wants to try an ultra-technical sort of game. With the bishop fianchettoed, Black's d-pawn is going to come under a lot of pressure if he ever plays ...c5, while if Black avoids ...c5, White can put pressure on the b- and c-pawns, and aim for a minority attack on the queenside. Black can try manoeuvring the bishop to e4, but it doesn't really affect White's intention of pressurizing the queenside. This would seem to be the only try for White in this section that offers any realistic prospects for an opening advantage.

8.••~a6! This, however, equalizes instantly; Black aims for the simple but effective ...~b4 and ....tf5 attack, while White cannot get his king away in time.

ge3 9 a3 c5, Chiburdanidze, gives Black excellent play. 9...~b4 10 .trs 11 ~dl Ilc8 12 a3 ~c6 Plaskett-Chiburdanidze, Banja Luka 1985. ECO ungenerously gives this as clearly better for Black, but 13 .td3 keeps White well in the game; it would be difficult to claim any advantage, however.

"'c3

This has occasionally been used as a move-order trick, with the intention of steering the game towards a Pawn Exchange Variation. As Black's reply shows, this particular move-order is inaccurate, but 1 d4 d5 2 c4 e6 3 ~c3 .te7 4 .tf4!? ~f6 5 e3 0-06 Ilcl (e3 being substituted for ~f3) is perfectly playable, and indeed succeeded in luring Kasparov into a Pawn Exchange Variation (after 6 ... c5 7 dxc5 .txc5 8 cxd5 exd5) in Korchnoi-Kasparov, Brussels 1986. Regrettably there is no space in this volume for detailed discussion of the implications of the various move-orders after 3 ....te7; it is clear, however, that Ilc 1 is not a wasted move if Black avoids the ... c5 variations. 6...dxc4! This immediately gives Black a satisfactory game. 6 ...c6 and 6 ...c5 have also been tried, transposing to more familiar lines.

7e4 The position now reached is the same as after 5 ... dxc4!? 6 e4 (Section

Miscellaneous Systems on Moves 5 and 6 S.S below), but with :lc1 inserted for White, and ...0-0 inserted for Black. The difference is for the most part in Black's favour (see the note to Black's 9th for the one exception). Instead 7 e3 lOdS is equal. 7 ••• bS!? As in the s ...dxc4line. 7 ...cs S .txc4 cxd4 9 lOxd4 lObd7 10 lOdbs lOb6 11 .td3 lOe8 12 es a6 13 lOd6 Miralles-Kharitonov, Royan 1988. SlOxbS Otherwise White is just a pawn down. S....tb4+ 9 M! lOxe4 Without:lc1 and ... 0-0, Black would have 9 ....txc3+ 10 bxc3lOxe4, and if 11 .txc4, then 11...lOxc3. 10 .txc4 gS!? Black must keep the initiative, since if he retreats, his weaknesses on the queenside will tell against him. The position reached is perhaps not critical enough for the theory of 5 .tf4 as a whole to justify detailed analysis, but the possibilities of 11 .te3 fs, 11 .tg3 fs, and 11 .td2 .txc3 12 bxc3 Wf6 all give Black chances of dynamic play.

249

B

=

D) 6'iVc2!? (D) An interesting move to dabble with. There is the possibility of returning to the main lines, but White hopes to benefit from having delayed e3, either through the possibility of e4, or ~ keeping a retreat for the bishop. '

6...lOc6! This, however, is recognized as the antidote. Others: 1) 6 ...c5 7 dxcs and now:

la) 7 ....txcs 8 a3 lOc6 9 0-0-0 WaS allows White to return to the New Main Line with 10 e3, which is to be regarded as a satisfactory outcome for him. White has tried a couple of embellishments, with only limited success: 10 .td2 Wc7 11 cxds lOe7 12 .tg5 (12 e4!?) 12 ... lOexds 13 lOxd5 lOxd5 14 e4lOb61S c,tbl.td7 16:lel :lfcS with unclear play, Bareev-Van derSterren, Biel IZ 1993; or IOWbl?! d4! 11 lOa4 (11 lOb5 a6! 12 lObxd4 lOxd4 13lOxd4 .txd414 :lxd4 eS 15 .td2 Wcs =1= Kotronias) 11.. ..te7 12 .td2 Wc7 13 e4 e5 +Conquest-Kotronias, Reykjavik 1992. Ib) 7 .....aS!? S lOd2?! (S 0-0-0, aiming to transpose to Line 'la', is more accurate) S...dxc4 9 e4 "xcS 10 .te3 "c7 11 .te2 (11 .txc4lOg4:;:) 11.. ..td7 12 0-0 lOa6 :;: Pein-Anka, Budapest 1989. 2) 6 ...dxc4 7 e3?! (7 e4!) 7...lOds 8 .txc4lOxf4 9 exf4 c5 (the Rubinstein Bind is not so effective for White when Black has not played ... c6) 10 dxc5 Wc7 11 0-0 'iVxf4 12 .td3 h6 13 :laellOc6 14 a3 .td7 15 b4 :lfdS Savchenko-Polovodin, St Petersburg 1993.

=

7e3

The Queen's Gambit Declined: 5 1./4!

250

7 :d 1 dxc4 S e4 ltlaS is an unclear and untested gambit. 7 •••ltlb4 7...a6Sh3 (SlIdl!?) S... ltlb49"cl c5 10 a3lOc6 11 dxc5 J..xc5 12 J..g5? (12 cxd5 ;1;) 12...d4 13 ltle4 J..e7 :;: Crouch-S.Dumitrache, Cappelle la Grande 1993. S.b3 S "cl dxc4 (S ...c5 9 a3 lOc6 10 dxc5 J..xc5 11 cxd5 ;1;) 9 J..xc4 c5 =. S•••b69 a3 9 lIcl c5 =; 9 lIdl J..b7 10 a3 dxc4 11 J..xc4 ltlbd5 12 J..e5 ltlxc3 13 "xc3 J..d6 Bareev. 9•••dxc4 10 J..xc4ltlbdS 11 J..g3?! 11 J..e5 =Bareev. 11•••J..b7 12 0-0 ltlxc3 13 bxc3 cS 14 a4ltle4 and Black has a slight advantage, Bareev-Vaganian, BledIRog~ka Slatina 1991. One feels that there is scope for further investigation, but that ultimately 6 "c2 is less threatening than 6 e3.

=

8.5 Miscellaneous 5th moves for Black 1 d4 dS 2 c4 e6 3ltlc3ltlr6 4ltlr3 J..e7 S J..r4 (D)

B

Black normally castles here, while 5 ... c5, 5... c6 and 5... a6 have already been considered under other headings. 5 ... dxc4!? is a significant but very much under-researched alternative. We consider first the minor possibilities: A: S•••ltlbd7? 250 B: S•••ltle4?! 250 C: S•••ltlbS 251 D: S...dxc4!? 251 A)

S•••ltlbd7? 6ltlbS! An outright refutation. 6 •••J..b4+ 6... e5 7 dxe5 c6 S exf6 TolozaR.David, Parnaiba 1995, is equivalent to resignation. 7ltld2J..aS 7... ltle4 S ltlxc7+ 'fIxc7 9 J..xc7 J..xd2+ 10 'fIxd2ltlxd2 11 ~d2 dxc4 12 e3 b5 13 a4 will soon leave White a pawn up. 8.a4! 8 b4 a6°o. 8•••J..xd2+ 8 ...c6 is Black's last chance, although 9ltld6+ ~ 10 b4 J..c7 11 c5 is dreadful. 9 xd2 eS 10 dxeS ltle4+ 11 el ltlecS 12 1Fb4 a6 13 e6 nbS 14 exd7+ ltlxd7 15 adS White has a sound extra pawn, Van Wely-J.Armas, Matanzas Capablanca mem 1994. B)

S•••ltle4?! This is more likely to be fruitful in the J..g5 variations, where it offers an exchange of bishops. Here it is a bit lazy with tempi, and straightforward

Miscellaneous Systems on Moves 5 and 6 development gives White a slight but very comfortable advantage. 6 1It'c2lbxc3 7 bxc3 0-0 8 e3 tM7 9 cxdS exdS 10 .td3lbf6 11 0-0 and White is slightly better, RobatschBaumgartner, Austrian League 1988/9.

C) S•••lbhS 6 .td2 6 .te5 f6?! (6 ... ~6!? 00) 7 .tg3 ~xg3 8 hxg3 f5 9 ~5 0-0 10 e3 ± Umanskaya-Dergachova, Svetlogorsk 1997. 6...lMi (D)

This challenges White to find some useful way of avoiding the repetition of moves with 7 .tf4. The usual approach has been 7 e3, for example: 1) 7 ...0-0 and then: la) 8 lIeI ~ 9 .td3 ~xd2 (9 ... fS 10 cxd5 exd5? 11 ~xd5) 10 1Wxd2 dxc4 11 .txc4 with an imposing lead in development, Gheorghiu-Mohring, Romania 1966. 1b) 8 .td3 dxc4 9 .txc4 c5 10 0-0 b6 11 "'e2 .tb7 12 lIadl a6 L3dx~5 .txc5 14 e4 b5 15 .tb3 1Wc7 16 e5 ~8 17 ~4 with a distinct advantage for White, Sarosi-Hartung, Dortmund 1988.

251

2) Black may consider the immediate 7 ...~!? followed by ...f5, ...c6, ... 0-0, etc., and it is not clear that White's plan of development has yielded him very much. My own preference would be for 7 1Wc2!?, keeping a watch on e4, and with the possible intention of playing

e4.

D) S•••dxc4!? (D)

An important but unexplored try. Practical experience, at least as far as the database is concerned, is limited to the following two examples. Cohn-Showalter, London 1899 saw Black getting a good game by unconvincing means: 6 e4 O-O?! 7 .txc4 a6 8 0-0 (8 a4!?) 8 ...b5 9 .td3 .tb7 10 1Wc2? (10 "'e2!?) 10... ~c6 11 a3 ~xd4 12 lbxd4 1Wxd4 13 .txc7 lIac8 ;. The more recent game, de Firmian-Gulko, Manila IZ 1990, 6 e3 lbd5 7 .txc4 lbxf4 8 exf4 0-0 9 0-0 ~7 10 ~e5 '12-'12 raises the interesting question of how deeply Gulko had analysed the system. I was hoping to add a third example when I was paired as Black against

252

The Queen' s Gambit Declined: 5 J.f4!

Mikhail Gurevich in the 1997 European Club Cup, but he varied with j.g5; the game was later drawn. Therefore what follows is mainly original analysis. We investigate: 01: 6e3 252 02: 6 e4!? 252 03: 6 "a4+!? 254

advantage of playing the Rubinstein Bind in traditional fashion, with kingside castling, is that White is free to probe for weaknesses on the queenside as well, without having to worry about the safety of his king, and this increases Black's defensive problems. 9j.bS j.d7=

01)

02)

6e3~!

6e4(D)

An important resource, and the rea-

son why Black delays castling. 6...0-0 7 j.xc4 is, as we have seen in Chapter 8.3, Line E, good for White. 7 j.xc4 The pawn can be a little troublesome to regain after 7 j.g3 t'hb6. 7•••lhxf48 exf4lhc6! The most accurate. Black avoids compromising his queenside pawn structure, and aims to create pressure against the d-pawn with pieces; ... j.f6 may soon follow. After 8 ...c5 9 d5 exd5 10 IOxd5 White is comfortably ahead in development, while if 8...0-0 90-0 iOd7, as in the de Firmian-Gulko game, White is probably better after

1OWe2. Unknown to me at the time of writing the main draft of this book, there is also the game Youngworth-Rai(!evic, Lone Pine 19S0, which went S...O-O 9 j.d3 t'hd7 10 "e2 t'hf6 11 O-O-O!? t'hd5 12t'hxd5 exd5 13t'he5 j.d6 14 h4 "f6 15"f3 j.f5 16 g3 c6 17 ~bl l:tfeS 18 h5 l:te7 19 g4 j.xd3+ 20 l:txd3 l:tae8. White is probably still better, but later lost. The idea of castling queenside and rushing the kingside pawns forward is tempting, but it can be difficult to find targets for the pawns if Black defends steadily. The

B

Presumably this is the reason why 5 ... dxc4 has been ignored. At first glance it would appear that White regains the pawn with ease, and keeps an enormous space advantage. However, it turns out that Black has tactical resources, based on the exposure of White's king. 6 •••bS! 7iOxbS j.b4+! A loss of tempo maybe, but still an awkward check to meet. IT now S1Oc3, then S... j.xc3+ 9 bxc3 t'hxe4 and Black stands well. White therefore has only two real choices: 021: 8lOd2 252 022: 8 j.d2 253 021) 8lOd2 O-O! (D)

Miscellaneous Systems on Moves 5 and 6

9"'c:2 1) The attempt at refutation with 9 liJxc7? is itself refuted by 9 ... e5! (the bishop is overworked as Black targets both the white knights; not, however, 9 ...liJxe4? 10 liJxaS .xd4 11 i.e3 i.xd2+ 12 .xd2 +-) 10 i.xe5 (after 10 liJxaS exf4 Black can, at the very worst, round up the as-knight, restoring material parity while retaining the initiative; 10 dxe5?? drops a piece to 1O...•xc7) 1O... liJxe4 llliJxaS i.xd2+ 12 ~e2, and now there must be several ways to win, but 12...liJc6 seems simplest. If one wants a pretty finish, 13 f3 'ifxd4 14 i.xd4liJxd4# is a new setting of Legall's mate. 2) 9liJc3!? i.b7 10 i.xc4liJxe4 11 liJdxe4 i.xe4 12 0-0 i.xc3 13 bxc3 liJd7 =. 9•••a6 9 ... i.b7 10 0-0-0 i.xd2+ II "'xd2 liJxe4 12 'ii'e3 liJa6 13 liJc3 liJd6 14 i.xd6 cxd6 IS i.xc4 is untested and probably favourable to White, who is better developed and has central control. Black can, however, argue-that the b- and c-files count for something, so not everything is clear-cut. This is just one line in a complicated position, with alternatives for both sides. The

253

text is clearer, and satisfactory for Black. 10 liJc3 10 liJxc7?! l:a7 11 tDbS axb5 12 i.xbS and Black has several strong continuations, e.g. 12...l:d713 i.eS tDg4. 10•••liJc6 1O...•xd4?! lli.e3 returns the initiative to White. The text maintains the flow of play. 11 i.xc4liJxd4 12 Wd3 cS With complicated play in which Black should be no worse, though practical tests are awaited. D22) 8 i.d2 i.xd2+ I had hoped to make the more ambitious S... c6 work, but after 9 i.xb4 cxb5 10 a4 liJc6 11 i.cS 'ii'a5+, the surprising 12 liJd2! (Burgess) is extremely difficult to meet. If then 12...liJb4 (12 ...liJxe4 13 axbS!) 13lla3! (Nunn) 13 ... liJxe4 14 axbS, Black's position is falling apart. So back to my original idea... 9tDxd2 9 'ii'xd2liJxe4 followed by ...liJd6 is fine for Black. 9•••a6 10 liJc3 "'xd4 11 tDxc4 II i.xc4 0-0 is less testing; the attack on the bishop leaves the knight temporarily stuck on d2. 11......xdl + 12 lbdlliJc6 We have reached the critical position, which needs further testing. Black's queenside pawns are uncomfortably split, but White is potentially weak on the central dark squares, and he still has to decide how to secure his e-pawn and make his bishop effective. My opinion of the position has varied from "White has a slight but

254

The Queen's Gambit Declined: 5 1./41

workable edge" to "Black is fine". For the sake of provoking further debate and testing, I suggest that the position is at least as satisfactory as Black can expect in a main-line opening, and that the most White has a very small edge. The position, although queenless, is not necessarily drawish, and Black will complete his development with ...J.b7, ... rJ;e7 and challenge the dfile. So the delicate question arises: if Black is doing OK after both 6 e3 and 6 e4, doesn't this just kill off the 5 J.f4 system? If this were so, the author would have mixed feelings; a sense of pride at having busted a whole opening system, combined with horror at the thought of what effects this would have on the sales of this book. All the reader would have to know is that 5 ... dxc4 is the recommended move, the rest of the book being consigned to the remainder bin of the chess bookstall. Mercifully though, there is a way in which White can play for an edge. As so often, the correct move is the one that has not been tried. D3)

6.a4+!? c6 The best. 6 ...1Oc6 7 e4 gives White control of the centre.

7.xc4.a5 Probably the best, although it is difficult to be certain in an uncharted

position so early in the opening. Black's potential freeing manoeuvres consist of, singly or in combination, ...lOe4/d5, ...c5, and ...b5-b4. Although we are only on move 7, we left theory two full moves ago. The usual arithmetic applies; in the absence of forcing lines, each player might have three or four or more reasonable moves, each with the same number of reasonable replies. It would be unreasonable to attempt any sort of definitive assessment of this particular position, which is critical to the 5...dxc4 system, and hence to the J.f4 system as a whole. It would be better to wait for this line to be played a few times first, so at least some main lines may become established. One line that might be worth further study is 8 J.d2 9a6 9 9xa6 .!Oxa6 10 e4 O-O!? Black is certainly not better after 7.....a5, while any advantage White has is far from crushing. As is usual in the opening we are in the zone of uncertainty between ;t and =. This is a pleasing place at which to draw the book to a close, with the indication that however much is already known about the J.f4 system, there is still much completely unexplored territory to be mapped out. It is hoped that the author of this book has fulfilled his duties to the readers: to cite, examine and assess what is already known, and to indicate where future discoveries are likely to be made.

Index of Variations 1 2 3

d4 c4

dS e6

lDc3

lOr6

4 lOO i.e7 4•••lObel7 214 5 adS exdS 6.c2 00 7 .tr4 215 and now: 7...1Ob6 215; 7...lDh5 216; 7... j.e7 8 e3 0-0 9 i.d3 219 0-0 247 5 i.f4 250 5•••cS 168 5••• b6194 5•••a6229 5•••lObS 251 5•••dxc4251 5•••00 6 e3lObd7 (6 ...1Oh5 227; 6 ...0-0 - see 5...0-0 6 e3 c6) and now: 7 h3 227; 7 i.d3 227; 7 'ifc2 212 6 e3 244 (DJ 6 cS 247 6 ad5 247 6l:[c1248 6.c2249

Others:

6•••lObel7238 6•••lDc6 245 6••• lDhS 245

6•••dxc4 246 6•••a6231: 7c5231; 7 l:[c1234; 7Wc2 234; 7 cxd5 237 6••• b6 172: 7 Wc2 173; 7 j.e2 174; 7 cxdS 175; 7 j.d3 187; 7 l:[cl 190 A)

6 ••• c6226 7 .c2213 7l:[c1224 7h3225 7 cxdS 217: 7... lOxd5 217; 7... cxdS 218; 7... exdS 218 7 j.d3220: 7...dxc4221; 7...lObd7 222 7... lObel7196 Now: 8 h3 196; 8 cxd5 208; 8 i.d3 208; 8 j.e2 209; 80-0-0210; 8 a3 211; 8l:[d1205 B)

After 6 e3: A: 6...00 B: 6 •••cS

6 ••• cS144 7 dxcS 165 7 i.d3145 7 j.e2146 7 adS 147 i.xcS 113 7 ••• 7•••lDc6165 7•••lOa6165 7•••dxc4165 7••••&5 165 8 .c2163 8 j.e2113 8 cxdS 122: 8... lOxd5 123; 8 ... exd5 135

The Queen's Gambit Declined: 5 .£141

256

8 a3 159 iLlc6 (8 ...b6 159; 8... lDe4 160; 8... dxc4 160; 8... a6 162; 8....te7 162) and now: 9 cxdS 102; 9 b4 103; 9 :CI 104; 9.te2 105 8 ••• lLlc6 8.•.dxc4 163 8••••aS 164 9 a3 149 9l:[dl.aS: IO.te2100; lolLld2101 9

•••

9••..td6151 9••••e7152 9•••l:[e8153 9•...te7154 9.•..td7155

.as

(D)

10•••l:[d821 10•••iLle4 21 11 lMe··2B 11.te223 11 l:[d224 11 ••• es 29 12 iLlb344 47 12 .tg3 47 d4 13iLlb3 12 .tgS d4 13 iLlb3: I3 .....b6 41; 13.....d830 12 ••• 1Wb6 13 .tgS 13 .tg3 d4 - see 12 .tg3 d4 13iLlb3 'ifb6

"b6

Now:

13.••.tg444 13.••d4 - see 12 .tg5 d4 13iLlb3 'it'b6

w

B2) 10 0·0-051 .te773 10•••dxc4 11 .txc4 76 .te7 - see 10... .te7 followed by ll...dxc4 10•••8685 10...:d887 10.••iLle4 87: 1liLlxe4 88; II iLlb5 89 10....td777: 11 g4 78; 11 ~bI81; 11 cxdS 83; II.tg5 84; IIiLld2 85; II h4 85 11 h452 ulLldl 74 11 ~b175 11 g4 63: 11 ...l:[d8 72; Il ... dxc4 64 11 ••• dxc4 11...l:[d860 11•••8661 11.••.td763 12 .txc453

After 9 .....a5:

Bl: 10l:[dl B2: 10 0-0-0 Or: 10 iLld2 94 10l:[c193

Bl) 10 l:[dl 20 10..•l:[e8 15 10•••dxc4 20 10•••.td721

.te7

Now:

12.••b653 12.•.8654 12..':d858