beating the griinfeld - actualités du club - 3eme open de...

Beating the Griinfeld

Beating the Griinfeld

ANATOLY KARPOV

Translated by John Sugden

B.T. Batsford Ltd, London

First published 1992 © Anatoly Karpov 1992

ISBN 0 7134 6468 2

British Library Cataloguing-in-Publication Data. A catalogue record for this book is available from the British Library

All rights reserved. No part of this book may be reproduced, by any means, without prior permission of the publisher

Typeset by Lasertext Ltd, Manchester and printed in Great Britain by Dotesios Ltd, Trowbridge, Wilts for the publishers, B. T. Batsford Ltd, 4 Fitzhardinge Street, London WIH OAH

A BATSFORD CHESS BOOK Adviser: R. D. Keene GM, OBE Technical Editor: Andrew Kinsman

Contents

Preface 7

1 Seville Variation 1 1

2 Exchange Variation with 7 �c4 - other systems 4 1

3 Modern Exchange Variation 52

4 Russian System 108

5 4 �f4 System 1 3 5

6 Fianchetto System 1 52

Index of Variations 1 87

Preface

I have to admit that the title of this book has an air of sales-talk about it, as similar titles generally do. Anyone trying to think up a recipe for beating this or that opening will scarcely succeed, and the Griinfeld Defence is no exception. But seriously: to chessplayers whose repertoire includes the Griinfeld, the author imagines that the present work will be of considerable interest and use. Hence it is indeed quite possible that the book will help many readers to score wins - with Black if they are adherents of the Griinfeld, and with White if they are looking for a way to combat this defence.

The book consists of a collection of games (or fragments of games) which reflect the contemporary state of Griinfeld theory. It is constructed around twenty-five paradigms of play by connoisseurs of this opening (like many authors I have a weakness for round figures), but the total number of examples is about ten times higher! The notes to each of the twenty-five principal games constitute a thorough discussion of the currently popular variations. As a result, all the most fashionable systems occurring in grandmaster practice in the last few years have found their way into the book. The selected games are not arranged chronologically but grouped according to themes. For this reason, the freshest examples - those from the recently concluded duel for the world chess crown, which incidentally are the most fully annotated - do not form the culmination of the book, but are placed in the middle of it . . . .

This book is written in the same format as the four-volume work probably already familiar to the reader : The Open Game (Semi-Open Game I Closed Openings I Semi-Closed Openings) in Action. That is to say, the scores of all the principal games are given in full, and in analysing the opening the reader will mostly be able to study its relation to the middlegame or even the

8 Preface

endgame. In some games, the later stages are of independent interest and consequently receive fairly detailed notes. The informal layout of material has permitted the author to select contemporary games according to his own discretion (in contrast, say, to an opening monograph, where you also have to include systems that have not been used in practice for a long time).

Whereas the four-volume work just mentioned covers the development of theory in the period 1984-7 (the last-but-one World Championship cycle), all the principal Griinfeld games in the present book are from the later period 1988-90, including the most recent World Championship match.

In my encounters with Gary Kasparov, the Griinfeld Defence has figured prominently. In our last three matches the most varied systems were tried out, and in our preparations we utilised all the most important theoretical material. As a result of these matches, the theory was in turn substantially enriched. Many ideas employed for the first time by Kasparov and myself have seen an onrush of further developments. Suitable examples of this are included in the book, while practical sources prior to 1 988 are incorporated in the notes to the principal games.

I should mention that I usually play the white side of the Griinfeld while Kasparov plays Black, and that both of us strive for victory (not only when playing each other). Such is the nature of this sharp and uncompromising opening, in which playing for the draw is inimical to both sides - although, to be sure, our fierce contests have often ended peacefully. For those who like statistics, let me give my overall score against the Griinfeld in each of these matches. London/Leningrad 1986 : +3 =6. Seville 1987: +2 -1 =7. New York/Lyon 1990 : +1 =3. As you can see, we have played almost an entire World Championship match of Griinfelds - with a definite plus score in my favour. If you like, you may conclude from this that I have unearthed the secret of Beating the Grunfeld . . . .

Obviously it is not only the World Champion and ex-champion who repeatedly contest this popular opening. It belongs to the repertoire of many prominent players. Suffice it to mention such names as Vassily Ivanchuk, Boris Gelfand and Jan Timman. Some of the valuable discoveries by these super-grandmasters will be found in this book.

In the course of working on the text, I have utilised a great

Preface 9

many sources (Informator, books and magazines), and this is reflected in the sheer quantity of references to games by masters and grandmasters. In the majority of cases where a variation or individual move is of major significance, its originator is mentioned.

In conclusion, I must thank Soviet Master Evgeny Gik, my coauthor in many books, for his help in preparing the manuscript.

Anatoly Karpov

1 Seville Variation

Game No. 1 Kuzmin-Henkin

Moscow 1989 I d4 .!Llf6 2 c4 g6 3 .!Llc3 d5 4 cd .!Llxd5 5 e4 .!Llxc3 6 be �g7 7 �c4 c5 8 .!Llel .!Llc6 9 �e3 0-0

10 0-0 �g4 The game has followed the main system of the Griinfeld. I

would just remind the reader that in the event of 10 . . . cd 1 1 cd .!Lla5 1 2 �d3 �e6 1 3 d5, theory states that White obtains a dangerous attack for the exchange.

1 w

l l f3 .!Lla5 (1)

12 Seville Variation

But now Black can answer 1 2 .td3 with 1 2 . . . cd 1 3 cd .te6, and the exchange sacrifice (14 dS) is less dangerous to him, since in several lines he has an important queen check on b6 (the details can be found in any reference work on the Griinfeld). However, in this situation White's exchange sacrifice is by no means forced. We shall later acquaint ourselves with a recent game with 14 .l:l. c l instead (Game No. 6 , Yusupov-Kasparov).

12 .txf7+ So International Master Kuzmin, one of my seconds, has

decided to participate in the theoretical debate launched by Kasparov and myself in Seville. After that 1 987 match, Kasparov stated that the plan chosen by White with 12 .txf7+ was unpromising. But to judge from the way the play went in the five games in question, it cannot be said that Black easily solved his opening problems. And in the post-match duel to which we shall presently turn (Game No. 2), he was thoroughly routed.

The pawn structure that now arises gives White every reason to count on the initiative, besides which he has an extra pawn. But then again, the position is highly dynamic and may very well suit the taste of the player of the black pieces.

Anyway, interest in this variation has not died down in three years, and numerous games enriching its theory have been played. Most of them will be mentioned in this book, over the first five main games. Of course, to some extent it will also be necessary to refer back to the Seville games.

12 13 fg 14 ..txfl

.l:l.xf7 .l:l.xfl +

Let me emphasise that the main feature of the position is not the extra pawn; the freedom of Black's game compensates for this minor deficit. White's basic plan is to block up the enemy bishop on g7, by means of the pawn chain c3/d4/e5/g5/h4. Black will rely on tactical devices to enable his bishop to escape onto the h6-cl diagonal.

14 "it'd6 (2) This queen sortie occurred in three games in the Seville match

and two of our later encounters. It has also been played in many other games in recent years. Let us nonetheless mention some alternatives.

In the ninth game in Seville, Kasparov preferred 14 .. . cd 1 5 cd

2 w

Seville Variation 13

11Vb6, and after 16 �g1 11Ve6 17 11Vd3! I returned the pawn while keeping all my positional trumps. The offshoots arising from 14 . . . cd 1 S cd will be examined in detail in Game No. S (apart from 1 S . . . "it"b6, the moves 1 S . . . "it"d7 and 1 S . . . eS have been seen).

The Seville match was not yet over when the new move 14 o o o

"it"d7 occurred in a game Chernin-Gavrikov, Lvov 1 987. That game proceeded : 15 de J:f8+ ( 1 S . .. "it"xg4 16 �f4!) 16 �g1 "it"xg4 1 7 �f4 ( 1 7 "it"d3 �c6 1 8 h3 is not bad either) 1 7 . . . "it"xd1 + (but not 1 7 . . . J: xf4? 1 8 .txf4 "it"xf4, on account of 19 "it"d8+) 1 8 : xd1 .txc3 1 9 �dS (White gains nothing from 19 �e6 : c8 20 J: d7 �f7 2 1 �gS+ �e8 22 J: d3 .tb4, or in this line 2 1 .tgS? .tf6! 22 .txf6 �xe6) 1 9 . . . .tf6 20 .th6 (the advantage is with White, but Black manages to hold on) 20 . . . J: e8 (but not 20 . . . : f7 2 1 �xf6+ : xf6 22 e S :rs 2 3 g4, and White wins) 2 1 �xf6+ ef 22 J: d7 : xe4 23 J: g7+ �h8 24 : c7 �g8 2S J: g7+ (2S .td2 �c6 26 : xb7 J:e7) 2S . . . �h8 26 : c7 �g8 27 J: g7 + �h8 28 l: c7, draw. Gutman suggests 20 g4! :e8 21 J: b 1 a6 22 �f2, followed by �e2 and .td2, with the initiative.

In answer to 14 . . . "it"d7, a more logical move seems to be 15 g5 (3), as played in Karpov-Gavrikov, European Speed Chess Championship, Spain 1988.

Despite the 'non-serious' nature of the contest, the game is of considerable interest : 15 o· · "it"e6 16 eS! "it"c4 ! 17 �g1 J: d8 18 "it"e1 �c6 19 .tf2 a6 (wouldn't 1 9 . . . bS !? have been better?) 20 a4 �aS 2 1 h4 .tf8 22 de! �b3 23 .l:!.bl ! (more precise than 23 J: d 1 : xd 1 24 "it"xd1 lbxcS) 2 3 . . . e6 (but now 23 . . . �xeS loses t o 24 : b4 "it"dS 2S J: d4) 24 c6! be 2S �d4 lbxd4 26 cd "it" xa4 27 "it"c3 (White has an obvious endgame advantage) 27 . . . J: c8 28 hS! gh 29 "it"h3 J: e8 30 "it"xhS J: e7 (30 . . . 'lfc2 3 1 J: b7) 3 1 g6 11Vc2 32


3 B

gh+ .l:l. xh7 33 11fg4+ �h8 34 .l:l.fl 11ff5 (34 . . . 11rd3 35 -*.h4!) 35 11'xf5 ef 36 -*.e3 .l:l. h5 37 g4 ! .l:l. h3 38 .l:l. xf5 .l:l. xe3 39 .l:l. xf8+ �g7 40 .l:l.f4! (the rook ending is won for White) 40 . . . a5 4 1 �f2 .1:1. b3 42 .l:l.f6 .l:l. d3 (42 . . . .l:l. c3 43 �e2 a4 44 �d2 etc.) 43 .l:l. d6 a4 44 �e2 .l:l. g3 45 .l:l. xc6 .l:l. xg4 46 �d3 47 .l:l. a6 �f7 48 �c4 .l:l. g3 49 d5 1-0.

From diagram 3, a game Gligoric-Popovic, Yugoslavia 1 988, continued differently : 15 . . . .l:l.d8 16 �g1 e6 (an evident improvement is 1 6 . . . cd 1 7 cd �h8 1 8 "ifd3 ll:lc6 1 9 .l:l.d1 ll:lxd4 20 ll:lxd4 e5) 17 .1:1. b1 ll:lc4 ( 1 7 . . . cd 18 cd ll:lc6 19 d5 ed 20 ed 11fxd5 2 1 "iWxd5 .l:l. xd5 22 .l:l. xb7 is good for White) 1 8 j.f2 b5 1 9 11rd3 a6 (White can now obtain the better game with 20 h4 ll:le5 2 1 11fh3, but Gligoric prefers to move his other rook's pawn) 20 a4 ll:le5 2 1 11fc2 (now 2 1 11fh3 ll:lc6 2 2 ab a b would lead t o unclear play - but not 21 . . . ba 22 de a3 23 .i.e 1 ! etc.) 21 . . . ll:lg4! 22 e5 b� (Gligoric gives the variation 22 . . . ll:lxf2! 23 �xf2 b4! 24 de .i.xe5!, and Black has no problems) 23 .i.g3 11fc6 24 "iWa2 ll:le3 25 ll:lf4 (White has a slight edge in the endgame, but it isn't enough for victory) 25 . . . ll:ld5 26 ll:lxd5 11fxd5 27 "iWxa4 cd 28 cd "iWxd4+ 29 11fxd4 .l:l. xd4 30 .l:l. b6 �f7 3 1 .l:l. xa6 .l:l. d5 32 .l:l. a7 + �g8 33 h4 ..txe5 34 -*.xeS .l:l. xe5 35 .l:l.e7 .l:l.e4 36 g3 .l:l.e 1 + 37 �f2 .l:l.e5 38 �f3 .l:l. e 1 39 �f4 .l:l. e2 t-t .

Perhaps an even sounder answer to 14 . . . 11rd7 i s 15 h3. This occurred in Yusupov-Popovic, Belgrade 1 989. There followed 1 5 . . . ll:lc4 1 6 .i.f2 cd 1 7 cd e5 1 8 de ll:ld2 + 1 9 � e 1 .txe5? (according to Yusupov, unclear play results from 19 . . . .i.h6 20 11fc2 .l:l. c8 2 1 11fb2 ll:lxe4 22 "iWb3 +) 20 .l:l. c 1 .l:l.d8 21 "ifc2!, and in the endgame White made no mistake in exploiting his extra pawn.

In addition to 14 . .. cd, 1 4 . . . 11rd6 and 14 . . . 11rd7, Black has


one other option : 14 • • . 1Wc8. This move was tried out in PortischKorchnoi, Reykjavik 1 988. After 15 1Wa4?! Black obtained an active game with 1 5 . . . cd 16 cd lbc4 17 .tf4 a6 1 8 g5 b5 19 1Wb3 e5 20 de 1Wc6. At this point, instead of 21 ltd1 , White should have preferred 21 1Wd3!? lbxe5 22 1Wd5+ 1Wxd5 23 ed b4, with about equal chances (but not 23 . . . lbd3 24 lt d 1 lbxf4 25 lbxf4 ltf8 26 g3 .te5 27 d6 .txf4 28 gf : xf4+ 29 "'e2, with initiative for White). Korchnoi answered 2 1 ltd1 with 2 1 . . . lt f8!, whereupon Portisch played 22 ltd5, overlooking the blow 22 . . . 1Wxd5! . After 23 ed lbd2+ 24 "'el lbxb3 25 ab lt d8, Black gained a decisive endgame advantage - though it took him fifty more moves to achieve the win!

Evidently White should react to 14 . . . 1Wc8 in the same way as if the queen had gone to d7, with 15 h3 or 15 g5. SeirawanKudrin, USA Ch. 1989, went 14 . . . cd 1 5 cd 1Wc8 16 g5, when the black queen utilised its possession of the c-file with 1 6 . . . 1Wc4. However, after 17 "'g1 lt d8 18 1We1 lbc6 19 e5 e6 20 1Wf2 lbb4 21 lDf4 ltc8 22 ltfl , it became clear that the raid with the queen had achieved nothing and the white centre was invulnerable. In the endgame after 22 . . . lDd3 23 lbxd3 1Wxd3 24 1Wf7+ "'h8 25 .tf2, White had a material and positional plus, which he duly converted to a win.

15 e5 In the 1 1 th match game in Seville, I chose 15 "'gl . For the

current state of theory on that move, see the notes to Game No. 4, where the completely new 1 5 1Wa4!? will also be discussed.

15 1Wd5 (4)

15 ••• -.e6 has also been seen. After 16 g5 lbc4 ( 1 6 . . . 1Wc4 1 7 "'g1 transposes into Karpov-Gavrikov, where play resolved itself clearly in White's favour), White has to avoid the trap 1 7 1Wd3? 1Wf5 + ! which costs him a piece; but even with 17 .tf2 lt f8 1 8 "'g1 1Wf7 19 1W e 1 lba3 20 ltcl 1Wxa2, h e achieves nothing.

A game Makarov-Hodko, USSR 1 988, went 16 h3 lbc4 17 1Wd3 ltd8 1 8 1We4! ( 1 8 lbf4 1Wa6!) 18 . . . 1Wc6 19 1Wxc6 lbxe3 + 20 "'f2 lbxg4+ 21 hg be 22 lt b 1 cd 23 cd c5 24 "'e3! cd+ (24 . . . lt d5 25 "'e4 e6 26 : b5 .tf8 27 : a5 is bad for Black) 25 lbxd4 : d5 26 lbc6. At this point, in Makarov's view, Black should have played 26 . . . lt c5 ; after 27 lt b8+ .tf8 28 lt c8 "'g7, White has no more than a slight edge.

Quite a good reply to 1 5 . . . 1We6 seems to be 16 lDf4 1Wc4+ 17


'ife2! :f8 1 8 'ifxc4+ 'Oxc4 19 �e2, with the better ending for White. One other possibility, 16 �g1 , was tried in HansenFereec, Aosta 1 989. Black restored the material balance, but after 16 . . . 'ifxg4 17 'iVd3 'ife6 1 8 -*.g5! : f8 19 h3 cd 20 cd -*.xeS 2 1 de 'ifxe5 22 -*.cl ! 'ifxa1 2 3 "ird5+ e6 24 'ifxa5 : c8 2 5 'iVd2 b5 (doubtless a more accurate line was 25 . . . 'iVb1 26 �h2 'ifc2 27 1We3 "irxa2 28 'Oc3 "irb3 29 .i.d2, with a minimal plus for White) 26 a3, his position was fairly difficult.

4 w

16 .i.f2 In this case 16 'Of4 is weak, since the queen gets to e4. A game

Lichak-Asrian, Leningrad 1 990, went 16 . . . : f8 1 7 �g1 "ire4 1 8 "irf3 "irxf3 19 gf .i.h6 2 0 10d5 -*.xe3 + 2 1 'Oxe3 cd 22 cd 'Oc6 23 : b 1 b6 24 �g2 'Oxd4, and Black had an endgame advantage.

16 :rs So far, the play coincides with game 5 in Seville. In the 7th

match game, the black rook preferred to go to d8, forcing the white queen off the central file. In later games, 16 . . . : d8 completely replaced the transfer of the rook to f8; we shall go into details in the notes to the next game in this book.

17 gS (5) A valuable novelty. The game in which this position first arose

(number 5 in Seville) continued 17 �g1 .*.h6 1 8 h4 "irf7 19 -*.g3 .i.e3+ 20 �h2 "irc4! 21 : b 1 (2 1 de is met by 21 . . . "irxg4, with a good game) 21 . . . b6 22 : b2 (but here 22 de is sounder) 22 . . . "ird5! 23 "ird3 'Oc4 24 : b1 b5, and the sharp duel should have ended in a draw (although a blunder by Kasparov eventually enabled me to win). However, 24 . . . g5 would have given Black good winning chances.


With 17 g5, White achieves his principal aim of shutting off the opposing bishop. Admittedly his king is dangerously placed opposite the black rook, but this is just a temporary problem.

5 B

17 Wf7 Gutman gives the variation 1 7 . . . "ife4 (or 1 7 . . . tt:Jc4 1 8 �g1

"ife4 1 9 tt:Jg3 "iff4 20 "ife2 b5 2 1 tt:Je4, with the better game for White) 1 8 tt:Jg1 ! tt:Jc4 1 9 ltlf3 ltle3 + 20 .i.xe3 'ifxe3 2 1 "iib3+ �h8 22 J:te 1 ! , and Black's position is not to be envied. 22 . . . J:txf3+ may look inviting, but after 23 gf "ifxf3 + 24 �g1 "ifg4+ 25 �f2 "ifh4+ 26 �e2 "ifg4+ 27 �d2 "ifxg5+ 28 �c2, there is no perpetual.

18 "ife1 h6 In reply to 18 • • • 'iff5, Henkin gives these variations :

(a) 19 h4, and now : (a1 ) 19 • • • "ifg4 20 �g1 h6 (or 20 . . . ltlc4 2 1 tt:Jg3! "ifxh4 22 tt:Je4, with a distinct plus) 2 1 gh .i.xh6 22 ltlg3! cd 23 cd "ifxh4 24 tt:Jf5 ! "ifg5 25 tt:Jxh6+ "if xh6 26 "ifxa5, and wins. (a2) 19· • • . h6 20 gh .i.xh6 21 ltlg3, with a clear advantage. (a3) 19 ... "ife4 20 �g1 cd 2 1 cd ltlc4 22 J:t c l ! (but not 22 g3? .i.xe5 23 de tt:Jxe5) 22 . . . ltle3 23 ltlf4! , and if 23 . . . l:t xf4 24 .i.xe3 ,J xh4 25 J:tc8+, White has a considerable initiative. (b) Another playable line is 19 ltlg3 "ifxg5 20 tt:Je4 'iff4 21 �g1 cd 22 cd tt:Jc6 23 .1: b 1 ! with the advantage.

In a game Kuzmin-Malishauskas, USSR 1989, Black played a new move, 18 • • • ltlc4 ( 1 8 . . . ltlc6 is also possible). There followed 19 tt:Jg3 tt:Jb2 20 tt:Je4 tt:Jd3 2 1 "ife3 (2 1 "ife2 tt:Jf4 !) 2 1 . . . tt:Jxf2 22


lt:lxf2 (Kuzmin recommends 22 . . . "it'c4 + , without the exchange in the centre; after 23 �g1 .z:td8, Black has the initiative) 23 cd "it'c4+ 24 �g1 .z:td8 25 "it'b3 "it'xb3 26 ab llxd4 27 llxa7 J.xe5

1-t (Black would lose after 27 . . . llb4 28 lt:ld3 ll x b3 29 lla8+ �f7 30 e6+ !). A much stronger answer to 1 8 . . . lt:lc4 is 19 lt:lgl! followed by 20 lt:lf3.

19 gh Stronger than 19 lt:lg3 hg 20 lt:le4 cd 21 cd "it'f5.

19 J.xh6 20 �g1 lt:lc4

An inadequate alternative is 20 . . . J.d2 2 1 "it'xd2 "it'xf2+ 22 �h1 �g7 23 "it'd3 lt:lc6 24 lldl .

21 lt:lg3! J.dl Otherwise White is a clear pawn up. The bishop move to d2

would also have been a good reply to 21 ..th4. 22 "it'el ..txc3 23 llfi!

23 e6? "it'xf2+ 24 "it'xf2 llxf2 would lose for White. 23 "it'e6

If 23 • • • cd, then 24 e6! is immediately decisive. On 23 • •• lt:ld2?, White has 24 .z:td1 cd (24 . . . "it'xa2 25 lt:le4!, or 24 . . . "it'f4 25 de!) 25 llxd2 J.xd2 26 "it'xd2 d3 27 -*.xa7.

24 de! ..txeS 25 lle1 "it'dS

25 . . . "it'f7 or 25 . . . llf4 would be safer. 26 lld1

If 26 "it'g4 (26 lt:le4 ..td4!), Black has 26 . . . llxf2! when Henkin's analysis goes : 27 "it'xg6+ �f8 28 �xf2 (28 "it'h6+ J.g7) 28 . . . J.d4+! (28 . . . "it'd4+ 29 �fl lt:le3+ 30 llxe3) 29 �fl lt:le3 + 30 : xe3 (but not 30 �e2 "it'xg2+ 3 1 �d3 "it'c2+) 30 . . . ..txe3 3 1 "it'f5 + "it'xf5 3 2 lt:lxf5 ..txc5, and Black seizes the initiative.

26 ll xf2 After 26 . . . "it'e6 or 26 . . . "it'f7, Black could still offer resistance,

but now the game ends at once. 27 llxd5 28 lt:lxe2 29 c6! 30 : xeS!

1-0

llxel �f7 b6

Game No. 2 Karpov-Kasparov

Belfort 1988


1 d4 ll:lf6 2 c4 g6 3 ll:lc3 d5 4 cd ll:lxd5 5 e4 ll:lxc3 6 be .tg7 7 .tc4 c5 8 ll:le2 ll:lc6 9 .te3 0-0 10 0-0 i.g4 1 1 f3 ll:la5 12 .txf7+ .:t xf7 13 fg .:t xfl+ 14 �xfl 'it'd6 15 eS 'it'dS

16 i.f2 .:t d8 Mter the 7th game in Seville, in which Kasparov played this

move for the first time, the rook manoeuvre to d8 became standard practice in this position. On d1 the queen was quite conveniently placed, but now the threat of 1 7 . .f .txeS compels it to leave its post. White gains nothing from 17 ltlf4 'ifc4+ ( 1 7 . . . 'iff7 1 8 ll:lh3!) 18 11Fd3 'ifxd3+ 19 ll:lxd3 cd 20 cd ll:lc6 21 ltlcS ll:lxd4 22 ll:lxb7 .:tdS 23 .:t e l : xeS 24 : xeS .txeS 2S ll:ld8 a6, with equality.

There are three ways for the queen to vacate the d-file : along the dl-a4 diagonal (to a4 or c2), or with 17 'ifel. We examine 17 'ifa4 in the present game, and 17 'ifcl in Game No. 3. But first, let us recall how the Seville game proceeded: 17 'ife1 'ife4 1 8 gS 'iffS! 19 h4 ltlc4 (he should have put pressure on the centre with 19 . . . ll:lc6 20 �gl 'ife4, leading to sharp play) 20 �g1 'ifg4 (20 . . . bS was more precise, as after White's next move, the black knight will feel uncomfortable) 21 a4 h6 22 : a2! hg 23 'it'bl ! (White thus succeeds in exploiting the weakness of the a2-g8 diagonal) 23 . . . gh 24 'it'b3 'ife6 2S ll:lf4 'iff7 26 ll:lxg6 'ifxg6 (I shall refrain from further comment until we reach one particular critical position; this book is not the place for exhaustive notes on the Seville games, they are no doubt well known to the reader already) 27 'ifxc4+ �h8 28 .:t b2! cd 29 cd 'ifg4 30 'iff7! .J:I. xd4 3 1 .txd4 'ifxd4+ 32 .:t f2 'ifxe5 33 : fs 'ife l + 34 :n 'ifeS 3S �hl? (although we are quite a long way out of the opening, it is appropriate to take stock of the situation. With 3S 'iff4! , White should come out on top : 35 . . . 'ifxf4 36 .:txf4 i.f6 37 : c4 �g7 38 .:t c7 b6 39 �fl . with a won ending) 3S . . . b6 36 'iff4 'ifhS (a withdrawal that was impossible a move earlier. If now 37 'it'b8+ �h7 38 'ifxa7, the white queen is far from the scene of action, and after 38 . . . 'ife2 Black has enough initiative to draw. But White has another way to exploit the awkward position of the black queen on the rook's file) 37 'iffS? (at this point, 37 .:t f3 ! i.f6 38 1Wb8 + �g7 39 'ifxa7 1Wc5 40 1Wa6 was decisive) 37 . . . 1We2 38 .:te l ? (here was White's third successive opportunity to win this


game. He should have checked first - 38 "it'c8+ �h7 - and only then played 39 :te l , threatening 'iltc2+. If 39 . . . 'iith5, then 40 'iltc2+ �h8 41 'iitd l ! etc.). The game lasted another 40 moves, but all my efforts to overcome my opponent were in vain. Although Black was on the brink of defeat more than once in this game, we have seen that the opening was not to blame. After 1 7 'iltel , Black had the means to obtain a perfectly reasonable position.

6 B

17 'ilta4 (6)

Now after the reply 1 7 . . . b6, my queen would withdraw to c2. You might ask what point there is in provoking a useful move of Black's b-pawn instead of placing the queen on c2 at once. The explanation is that in some variations (after Black's . . . "it'c4, for example) White can play 'ilte4, gaining a valuable tempo by attacking the knight which would by now have returned to c6. So 1 7 'ilta4 has its plus side. But I will not offer a categorical verdict as to where the queen feels more comfortable, although I have tried both possibilities in my games.

17 b6 In the event of 17 . . . lLlc6 1 8 "it'b3 c4 19 'iltxb7, White has the

better game. After 17 .. . lLlc4 18 lLlf4 'iltf7 19 g3 lLld2+ 20 �g2 lLle4 2 1 'iltc2

lLlxf2 22 'iltxf2 White would again have a plus, but at move 20 Black has the powerful 20 . . . g5 ! , pointed out by Gutman. He suggested that White should play 18 g5! himself, with the continuation 18 . . . l:t f8 19 �g1 'iltf7 20 ..tg3 lLle3 21 ..tf4 lLld5 22 l:t fl . Black's initiative has been neutralised and his bishop imprisoned, which is to say that the opening contest is settled.

After 17 • • • l:f8 1 8 g1 'llf7 1 9 .*.h4! ll:lc4 20 'ilb3, the bishop is again unable to break out and free itself: 20 . . . .*.h6 21 g5! .i.xg5 22 .*.xg5 'ilf2+ 23 '>Ph1 'ilxe2 24 h3, with a substantial advantage to White.

18 'ilcl There is nothing more for the queen to do on a4.

18 l:f8 After this, the black pieces come under pressure. Evidently 18

• • • l: c8 (7) was more exact, trying to obtain counterplay along the c-file.

7 w

From diagram 7, here are some important recent examples : Tisdall-Thorsteins, Reykjavik 1989, went 19 de be ( 19 . . . l: f8?

20 cb ll:lc4 21 g1 ) 20 l: d 1 'ilxe5 21 'ila4 l: f8 22 l: d3 c4 23 l: f3 'lidS 24 l: xf8+ xf8 25 .i.xa7 ll:lc6 26 h3 'ile4 27 'ilb5! and White had a considerable plus. However, 20 . . . 'llc4! is more accurate : 21 .i.g3 'ilxg4 22 h3, with chances for both sides.

Lputian-Dzhandzhgava, Simferopol 1988, went 19 'ildl l:d8 20 'ilcl (20 'ilc2 repeats moves) 20 . . . l: f8 (the position that has now arisen is the same as in my game with Kasparov, except that the white queen is on c l . The difference is not too important, but . . . ) 21 h3 (this is a serious inaccuracy. Correct moves are 2 1 g 1 and 2 1 g5) 2 1 . . . W"f7 2 2 W" e l .th6 2 3 ll:lg3 ll:lc4 24 e 6 W"g7 25 ll:le4 .te3 26 W"e2 b5 27 '>Pg1 .txf2+ 28 ll:lxf2 cd (Black has won his pawn back and seized the initiative; the game doesn't last much longer) 29 cd W"xd4 30 :Z. c l g5 31 l:c2 ll:le3 32 l: d2 'ilal + 33 ll:ld l l:fl + 34 h2 W"e5+ 35 g3 'ile4 36 l: d8 + g7 37 'ilb2+ r;Ph6 0- 1 .


In Douven-Ilincic, Alma-Ata 1989, after 1 9 1Wd1 l:t d8, White played 20 ll'el , an improvement on the previous example. There followed 20 . . . .!tlc4 (a more precise method was 20 . . . cd 2 1 cd -th6 22 �g1, and only then 22 . . . .!t!c4) 2 1 g5 1We4 22 �g1 1Wg4 23 .ltlg3 cd 24 cd 1Wxg5 25 1Wb4 l:t c8 26 .ltle4 1Wf4 27 1Wxe7 �h8 28 We6 l:t f8 29 1Wxc4 1Wxe4 30 :e1 , with a clear advantage.

19 .ltlf4 is well answered by 19 . . . 11'f7 or 19 . . . 1Wc4+ 20 1Wd3 -th6. Another inadequate line is 19 1Wd2 :f8 20 1Wg5 1Wf7 2. 1 1Wh4 -th6! (but not 2 1 . . . .!tlc4? because of 22 g5!, as in SchneiderLjubojevic, Pernik 1 988) 22 �g1 .!tlc4!, and the initiative is with Black. At move 20, the unfortunate 20 �g1? led to a quick loss in Schulz-Filipovic, Korimofen 1 989 : 20 . . . -*.xeS 21 l:t d 1 -txh2 + !

It only remains to add that 18 . •• 1Wc4 would be met by 19 1We4 with favourable consolidation, for instance : 1 9 . . . l:t f8 20 �g1 ! -th6 2 1 -th4! . Notice one peculiarity of the position resulting from 1 7 . . . b6: Black now constantly has to reckon with the exchange d4xc5.

19 �g1 1Wc4 The bishop cannot now break out onto the open board: 19 . . .

-th6 20 h4 1Wf7 2 1 .ltlg3, or 19 . . . .!t!c4 20 h4. 20 1Wd2!

White continues with his plan of restricting the bishop's mobility. After 20 1We4, he would have to reckon with 20 . . . -th6 and especially with 20 . . . .!tlc6!?, threatening to capture on e5; naturally, 2 1 1Wxc6 1Wxe2 cannot arouse White's enthusiasm.

20 1We6 20 . . . 1Wf7 is not good, since after 21 .ltlg3 everything fits together

for White : his knight is transferred to e4, and his queen to e2. Nor does 20 . . . -th6 21 1Wxh6 1Wxe2 rid black of his worries, on account of 22 1We3 1Wxg4 23 de be (23 . . . .!tlc4 24 1Wd4) 24 1Wxc5.

21 h3 .!tlc4 22 ll'gS! (8)

A critical moment. In addition to .!tle2-f4, White will now be threatening -tf2-h4 in some variations.

22 b6 22 . . . -tf6 is no good, if only because of 23 ef (23 .ltlf4 is also

strong) 23 . . . ef 24 .!tlf4. 23 ll'cl 11'£7

To obtain counterplay Black had to opt for 23 . . . b5, intending

8 B


(for example) 24 ll:lf4 'flf7 25 ll:ld3 b4 !? - even though 25 .tg3 maintains a plus for White. 23 .. . 11d5 would not work in view of 24 'flc2!, while after 23 . . . h5 the queen would return to g5.

24 .tg3 g5 This move was roundly condemned by the commentators. But

I do not recall that any serious alternative was suggested to give Black counterplay. 24 • • • 11d5 (for example) has been recommended, but then 25 ll:lf4 'fle4 26 ll:le6 would put Black in a very dangerous position. There are three variations - let us look at them : (a) 26 . . • .l:l.c8 27 'ilb1 ! 'fle3 + 28 .tf2 'flxc3 + 29 'flxg6 'flxa1 + 30 �h2, with unavoidable mate. (b) 26 . . . ll:le3 27 11d2 cd 28 cd .l:l. c8 29 .l:l.e 1 .l:l. c2 30 .l:l. xe3 'flc6 3 1 d5, and Black is lost. He could have tried driving the queen to e2, with 27 . . . ll:lc4 28 'fle1 ll:le3 29 'fle2 cd. Then 30 cd .l:l. c8 would give Black strong counterplay, but 30 ll:lxf8 .txf8 (30 . . . d3 3 1 11f2) 3 1 'flf3! 'ild3 32 cd 'ilxd4 33 .l:l. e 1 clears up the situation completely - since 33 . . . ll:lc2+ 34 .tf2 ll:lxe1 is refuted by the intermediate 35 'ilb3 + . (c) 26 • • • cd 27 ll:lxf8 ltle3 28 11d2 de 29 'fle2 .txf8 30 'flf3, with a big advantage.

25 'flc2 11d5 26 .tf2 b5 27 ll:lg3 .1:1. £7

Forced. If 27 . . . b4, then 28 ll:lf5 is unpleasant, while 28 . . . .l:l. f7 is met by 29 e6 'flxe6 .l:l. e 1 11d7 3 1 cb.

28 .l:l.el If White had wanted, he could have put a stop to Black's

queenside counterplay with 28 : c t .

28 b4 29 1Wg6 �f8

29 . . . be loses immediately to 30 o!Llf5 f8 3 1 e6 .l:l. xf5 32 gf o!Lld6 33 de.

30 o!Lle4 30 o!Llf5 is also strong: 30 . . . e6 3 1 o!Llxh6 .l:l. f4 32 1Wxg5 , or 32

1Wh7. 30 .l:l. xfl

The exchange sacrifice is incapable of improving matters. All it demands from White is a modicum of accuracy.

31 �xf2 be 32 'tiffS+ �g8 33 1Wc8+ h7 34 1Wxc5 1Wf7+ 35 g1 c2 36 o!Llg3 .i.f8 37 o!Llf5 �g8 39 .l:l.c1

1-0 Black's pieces on the kingside never succeeded in breaking free.

As this game was played in the World Cup in the French town of Belfort, someone jokingly called 1 7 1Wa4 the Belfort variation . . . .

I might add that this game was declared the best game played in the first half of 1988 (Informator, vol. 45). Also, it proved to be our last decisive encounter until the following (already our fifth) World Championship match.

Game No. 3 Karpov-Timman Rotterdam 1989

1 d4 o!Llf6 2 c4 g6 3 o!Llc3 d5 4 cd o!Llxd5 5 e4 o!Llxc3 6 be J..g7 7 J..c4 c5 8 o!Lle2 o!Llc6 9 J..e3 0-0 10 o-o J..g4 1 1 f3 o!Lla5 12 J..xf7+ .l:l.xf7 13 fg l:Xfl+ 14 �xfl 1Wd6 15 e5 1Wd5 16 J..f2 .l:l. d8

In the fifth game in Seville, Kasparov chose 1 6 . . . .l:l. f8 (see Game No. 1 ). As we know, the rook move to d8 was first employed in the 7th game of the match.

17 1Wc2 In the first game with this line, I chose 17 'tlfe l (see Game No.


2); in Belfort 1988, I preferred 1 7 'ilt'a4 (Game No. 2). 17 'ilt'c4

In Griinberg-Ilincic, Prague 1 989, an equal game resulted from 1 7 . . . .l:t c8 1 8 illf4 'ilt'f7 1 9 'ilt'e4 J.h6 20 g3 .l:tf8 2 1 ¢>g2 J.xf4 22 gf 'ilt'xf4 23 'ilt'xf4 .l:t xf4 24 ¢>g3 .l:t f8.

18 'it'b2 But not 1 8 g5? on account of 1 8 . . . cd. If 1 8 'ilt'e4, then 1 8 . . .

.l:t d5! , threatening 1 9 . . . .l:t xe5, is quite good for Black. 18 J.h6

In Ljubojevic-Timman, Linares 1989, Black moved his rook again : 18 ... .l:t f8. There followed 19 ¢>g1 J.h6 20 .l:t dl ! 'ilt'a4 2 1 .!:te l cd (better was 2 1 . . . illc4 2 2 'ilt'b3 'ilt'xb3 2 3 ab illb2) 2 2 illxd4 'ilt'c4 23 h3 b6 24 illf3 .l:t d8 25 .td4 (25 g5! is also playable) 25 . . . J.f4 26 ¢>f2 1Fd5 27 'it'bl illc4 28 'ilt'e4 illb2 29 'ilt'c2 'it'b5 30 .l:t b l illd3 + 3 1 ¢>fl 'ilt'c4 32 'ilt'e2 b5 33 ille l ! illxe5 34 .l:t xb5 'ilt'f7 ! 35 ¢>g1 illc6 36 illf3 illxd4 37 cd .l:t c8 (but not 37 . . . J.d6 38 illg5 'ilt'f4 39 'ilt'e6+ ¢>g7 40 'ilt'xe7 + with advantage, or 39 . . . ¢>f8? 40 .l:t f5 + and wins) 38 .l:t c5 .l:t b8, and the players decided to draw by repetition. It remains to be noted that at move 28, White could have kept the initiative with 28 .l:t e2!

After 18 . . . .l:t f8 19 ¢>gl J.h6 20 .l:t d 1 , an innovation, 20 • • • 'ilt'e6, was played in L. Hansen-Jasnikowski, Warsaw 1990. (Instead, 20 . . . 'ilt'f7 2 1 .!Llg3! illc4 22 'ilt'e2 'ilt'xf2+ 23 'ilt'xf2 J.e3 24 ille4! .l:t f4 25 .l:t d3 is hardly good for Black. ) However, after 2 1 h3 .!Llc4 22 'ilt'xb7 .te3 (22 . . . ille3 23 .!:ta l 'ilt'f7 24 J.xe3 J.xe3 + 25 ¢>hl 'ilt'f2 26 'it'b5 cd 27 illxd4!) 23 J.xe3 illxe3 24 .l:t d3 'ilt'f7 (or 24 . . . .l:tfl + 25 ¢>h2 illxg4+ 26 hg 'ilt'xg4 27 .l:t f3 'ilt'h5+ 28 .l:t h3 'ilt'xe2 29 'ilt'xe7 .l:t f7 30 'ilt'xc5 and wins - Hansen) 25 'ilt'f3! 'ilt'xf3 26 gf .l:t xf3 27 de! White had a clear endgame advantage.

9 B

19 h4 (9)


This position first occurred in Karpov-Kasparov, Amsterdam 1 988, which proceeded as follows : 19 • • . 11f7 20 �gl .J:I. f8 21 .!Dg3 (2 1 .ig3 is met by 21 . . . .ie3+ 22 �h2 11e6 23 g5 11g4) 21 . . . .!Dc4 22 'ife2 11xf2+ ! (other continuations lead to advantage for White : 22 .id2 23 .!De4, or 22 . . . cd 23 cd 11xf2+ 24 11xf2 .ie3 25 11xe3 .!Dxe3 26 .J:I.c l ) 23 11xf2 .ie3 24 11xe3 (24 .J:I. fl? .J:I. xf2 25 .J:I. xf2 cd) 24 . . . .!Dxe3 25 de .J:I. c8 (but not 25 . . . .!Dxg4 26 .J:I. b l , when White i s already better) 26 .J:I.b1 .J:I.xc5 27 .J:I.xb7 .!Dxg4 28 .J:I. b4 (White may also play 28 : xe7 : xc3 29 .!Dfl : a3 30 e6 .J:I. xa2, which quickly draws) 28 . . . h5 29 .!De4 .J:I. xe5 30 g3 (30 �fl .!De3+ 3 1 �e2 .!Dd5 32 .J:I.b8+ is simpler) 30 . . . �f7 3 1 �g2 �e6 32 �f3 (a more accurate choice is 32 .!Dg5+ �f6 33 .J:I.f4+ .J:I.f5 34 .J:I.e4 .J:I.f2+ 35 �g1 .J:I.xa2 36 .:t f4+ �e5 37 .J:I.e4+ �d6 38 .J:I.e6+ �d7 39 .J:I. xg6, and a draw can be agreed) 32 . . . �f5 ! . Black now has slightly the better endgame, not that there is much danger to White. After a further two dozen moves, we agreed a draw.

The question arises whether the strange retreat with the queen to f7 can really solve all Black's problems. It turns out that this is not quite the case. In Vyzhmanavin-Ernst, Stockholm 1990, Black was dealt an annihilating blow, radically altering the assessment of the variation. The game went 19 . . . 11f7 20 g5! (White isn't afraid of the pin on his bishop. The main thing is to shut the opposing bishop out of play) 20 . . . .!Dc4 2 1 e6! (this is just the intermediate move that constitutes the refutation of Black's defence. White deflects the queen, now allowing his opponent to double on the f-file, and only afterwards picks up the b-pawn) 2 1 . . . 11f5 2 2 .!Dg3 11xe6 2 3 11xb7 .J:I. f8 24 .J:I.e1 ! (now Black can't avoid loss of material) 24 . . . .!De3+ 25 �gl .J:I.xf2 26 gh cd 27 cd .:t f8 28 11b2 lljc4 29 .J:I.xe6 1-0.

Perhaps Timman had a foreboding of such misfortunes, and therefore avoided the move that justified itself in Amsterdam ( 1 9 . . . 11f7), preferring to occupy the f-file with his rook.

19 .1:1.£8 20 g5!

The stock manoeuvre; here too it guarantees White a plus.

20 11d3 21 'ilbl!

Gaining a couple of tempi for bringing the queen to the defence.


A mistake would be 2 1 gh? lbc4 22 1i'c l 1Wf5 23 it'el lbe3 + 24 �gl lbc2.

21 1i'e3 22 it'el R.g7 23 �gl

But not 23 lbcl? cd and everything is fine for Black, as after 24 1t'xe3de, he wins.

10 w

23

24 lLig3!

1t'e4 (10)

White returns the pawn but seizes all the key squares with his pieces. A familiar precept!

24 1t'xh4 25 lbe4

25 lLif5 is also strong : 25 .. . 1t'xg5 26 lLixg7 irg4 (26 . . . �xg7 27 R.h4 1i'g4 28 R.xe7, etc.) 27 d5 �xg7 28 R.xc5 1t'd7 29 1i'e4, and Black is in a bad way.

25 l:txf2 An exchange sacrifice (again not entirely voluntary!) analogous

to the one that occurred in Karpov-Kasparov at Belfort. 25 . . . 1t'g4 does not help either : 26 lLixc5 b6 27 1i'e4 1t'xg5 28 1i'd5+ l:t f7 29 lbe6, and White should win.

26 lLixf2 cd 27 l:tdl ! d3

On 27 . . . lLic6, I had in mind 28 cd lbxd4 29 1i'e3 lbc6 30 1t'b3+ �f8 3 1 1i'xb7 1t'c4 32 1i'c8+ �f7 33 lbg4! . Nor is 27 . . . de 28 1i'xc3 lbc6 29 1t'b3 + any better for Black.

28 1We3! lLic6 28 . . . lLic4 is well answered by 29 1Wd4.

28 §eville Variation

29 tl:lxd3 'lra4 30 'lrf3! 'lraS 31 e6 tl:ld8 32 tl:lf4 .i.eS 33 tl:ld5! 'IreS+

If 33 . . . tl:lxe6, then 34 l:t fl ! is decisive. 34 �h1

1-0 There is no defence against the threats of 35 tl:lf6+ and 35 :fl .

Game No. 4 Naumkin-Neverov

Moscow 1989 1 d4 tl:lf6 2 c4 g6 3 tl:lc3 d5 4 cd tl:lxdS 5 e4 tl:lxc3 6 be .i.g7 7

.i.c4 c5 8 tl:le2 tl:lc6 9 .i.e3 0-0 10 0-0 .i.g4 11 f3 tl:laS 12 .t.xf7+ .:.x£7 13 fg J: xfl + 14 �xfl 'ird6

15 �g1 In the 5th and 7th match games in Seville I continued 1 5 e5,

and the current state of theory on that line is given in the notes to the first three games of this book. In the 1 1 th match game, however, I played the king to g1 here. White is prepared to return the pawn in order to obtain a sturdy centre. But before going on to a detailed discussion of the prospects for either side, it is worth mentioning one other possibility : 15 'lra4!?

This occurred in Zakharov-Henkin, Voronezh 1 989, where Black replied 1 5 . . . 'lrxh2! (11) .

ll w

A bold piece sacrifice. Instead, the timid 1 5 . . . b6 1 6 �g1 cd 1 7 cd 'lre6 1 8 d5! 'lrxg4 1 9 :tel , followed by 20 h3, is in White's favour.


The game proceeded : 1 6 'ifxa5 l:l. f8+ 1 7 �e1 (but not 1 7 J..f2? on account of 1 7 . . . 'ifh4! 1 8 g3 'ii'h 1 + 1 9 �g1 'ifh2, and wins) 1 7 . . . 'ii'h 1 + 1 8 �g 1 ! (on 1 8 .!.g1 ?! , Black has the very strong 1 8 . . . ..th6! 19 'ifxc5 'ifxg2, threatening 20 . . . 'iffl mate) 1 8 . . . 'ifxg2 (but here, 1 8 . . . J..h6 comes too late : 1 9 �d2 l:l. f2+ 20 �d3 c4+ 21 .;;.xc4 J..xe3 22 'ii'd5+ .;;.g7 23 l:l.e 1 , with the better endgame chances for White; if he prefers, he can give perpetual check by 23 �f3 'ifxa 1 24 'ife4+) 19 'ii'b5 ..th6 (after 19 . . . cd 20 cd 'ifxe4 2 1 'ife2 and 22 l:l. d 1 , White consolidates; but now, a highly unusual distribution of forces comes into being) 20 'ife2 'ifg3+ 21 �d2 l:l. f2 22 J..xh6 l:l. xe2+ 23 �xe2 'ifh3! 24 J..e3 ! 'ifxg4 25 l:l.fl ! h6! (a very strong move according to Henkin, who gives the following variations : 25 . . . cd 26 J..h6! de+ 27 �c1 'ifc8 28 �d4, threatening �e6 and l:l.f8+; 25 . . . h5 26 �f4, followed by l:l. g 1 ; or 25 . . . �g7 26 �f4 g5 27 l:l. g 1 'ifh4 28 l:l. xg5 + ! . Now the game heads towards a draw) 26 J..xh6 'ifxe4 27 l:l. f8+ �h7 28 .!.g5 'ife6 29 a4 cd 30 cd 'ifa2+ 3 1 .;;.el 'ifxa4 32 lH7 + .;;.g8 33 .J:I. xe7 'ii'b4+ 34 �f2 a5 35 J..f6 'ii'b6! 36 J..e5 a4 37 �f4 'ii'b2+ (Black loses after 37 . . . �f8 38 J..g7+ !, or 37 . . . a3 38 �e6!) 38 �g1 ! 'ifc1 + 39 �h2 'ii'b2+ 40 �g1 'ii'b 1 + 41 .;;.h2 'ii'b2+ t-t.

So the extravagant 1 5 'ifa4 is not dangerous to Black. It must be acknowledged that 1 5 e5, which we have already examined, is the most precise; White can move his king if the need arises. Nonetheless, 1 5 .;;.g 1 should also be studied.

15 'ife6 In Polajzer-Anka, Dortmund 1 988, Black played 15 . • • l:l.d8.

It isn't clear that the rook on d8 is useful to Black after 16 'ii'd3 'ife6 1 7 g5! . However, White replied 16 'ifa4 'ifa6 1 7 .J:I.e1 'ii'd3 1 8 J..f2 �c4 1 9 'ii'b5 l:l.f8 20 'ii'b l 'ii'd2 2 1 'ii'b3 b 5 2 2 'ii'xb5 �e3 23 h3, and now, in Gutman's view, 23 . . . J..h6! would have given Black a powerful initiative.

16 'ii'd3 'ifc4 So far, the play has followed the 1 1 th game in Seville, and the

text move is better than the capture on g4, which occurred (with a transposition of moves and the insertion of the pawn exchange on d4) in the 9th match game. In Chernin-Malishauskas, Lvov 1987 (played while the Seville match was still in progress), Black didn't hurry to exchange queens, preferring 16 • • • cd 17 cd l:l. d8. After 18 g5 �c4 1 9 J..f2 b5 20 a4 (20 h3 was worth considering)


20 . . . ba 2 1 lllf4 'flf7 22 llld5 .l: f8 23 i.g3 lllb6 (but not 23 . . . lllb2 24 'flc3 a3 25 h3, followed by i.e5, with advantage) 24 h4 (24 lllxb6 'ilb3!) 24 . . . lllxd5 25 ed 'ilxd5, the chances were equal. Instead of 1 8 g5 White can play 18 h3, but after the queen exchange Black's chances, once again, are no worse : 1 8 . . . 'ilc4 19 'ilxc4 lllxc4 20 i.f2 e5 2 1 d5 i.h6 22 h4 (22 a4 llld6!) 22 . . . i.d2 23 a4 llld6!. Note that variations involving the pawn exchange in the centre ( . . . c5xd4, c3xd4) will be examined more closely in Game No. 5.

17 'ilxc4 It hardly pays White to avoid the queen exchange : 1 7 'ild2 cd

1 8 cd 'fla6! or 1 7 . . . 'fle6. 17 lllxc4 18 i.f2

Vasser Seirawan, who has also incorporated this variation in his repertoire, twice played 18 i.gS (12) here. Both games are worth examining.

12 B

Seirawan-Lputian, St John 1988, continued 18 . • . h6 1 9 i.xe7 cd (Black can win a piece with 1 9 . . . .l: e8 20 .i.xc5 b6, but after 2 1 i.xb6 ab 22 e5, White obtains more than enough pawns for it. Furthermore, once his king is centralised and the knight jumps to e4 via g3, the white pawn mass can become highly mobile) 20 cd .l:e8 21 .l: c 1 llla5 (2 1 . . . .l: xe7 22 ::txc4 .l: xe4 23 �f2 .l: xg4 24 J:t c8 + �h7 25 .l: c7 is no better for Black, but 2 1 . . . b5 was worth considering) 22 .l: c7 lllc6 23 i.c5 J:t xe4 24 ci>f2 J.xd4+ (in the event of 24 . . . lllxd4, Lputian gives 25 .1: xg7 +! �xg7 26 i.xd4+ �f7 27 �f3 J:t e6 28 i.xa7 J:t a6 29 J.e3 J:t xa2 30 i.xh6 b5 3 1 lllc1 , with advantage) 25 lllxd4 lllxd4 26 .l: xb7 lllc6

27 h3 :la4 28 a3. White has an extra pawn, which he eventually exploited to win (though it took him 30 moves!).

Seirawan-Hort, Lugano 1 988, went 18 • • . cd 19 cd eS 20 :lc1 (20 d5 h6 2 1 .i.cl �d6 22 �g3 : c8 23 .i.e3 :lc3 24 f2 .i.f6! 2S a4 .i.gS led to equality in Dlugy-Nikoloff, Toronto 1 989; Black can also play the immediate 20 . . . tLld6 21 �g3 : c8 22 .i.e3, with adequate counterplay) 20 . . . bS 2 1 de .i.xeS 22 :ld1 : c8 (22 . . . :e8 is more accurate and gives equality} 23 .i.f4 ! .i.g7 24 :ldS a6 2S �f2 :le8 26 f3 f7 27 h4 �eS+ 28 .i.xeS :xeS 29 : d3 b4 30 �f4. Here again, White has an extra pawn, but this time Black managed to hold out.

Interestingly, his loss to Seirawan made such a strong impression on Lputian that he took the first opportunity to play this variation with White. In Lputian-Hansen, Dortmund 1988, Black played (from diagram 1 2} the immediate 18 • . • e5. There followed 19 d5 b5 ( 1 9 . . . h6 20 .i.cl is sounder for Black} 20 : b1 :b8 21 f2 aS 22 thc1 h6 (22 . . . �a3 23 :l b3 b4 24 cb cb 25 d6) 23 .i.e3 �xe3 (in Lputian's view, 23 . . . tLld6 24 f3 :lf8+ 25 e2 c4 26 e1 �xe4 27 :lxbS �xc3 28 :lxaS :ld8 would have given equality; but not 23 . . . :lf8 + 24 e2 �xe3 2S �xe3 :lfl 26 :la 1 ! and 27 �b3!} 24 xe3 c4 2S �e2 .i.f8 26 tZ'lg1 .i.cS+ 27 e2 .i.xg 1 28 :lxg1 f7 29 a3 rj;e7 (29 . . . b4 would have retained some saving chances} 30 :l b 1 , and White won the rook ending. After the game, Lputian explained that at move 2 1 , the correct course was 21 a4! b4 22 cb cb (or 22 . . . :lxb4 23 tZ'lc3 �d6 24 :le 1 , preparing 2S .i.e7) 23 �c1 .i.f8 24 �b3 �d6 2S �d2 b3 26 .i.e3 a6 27 fl! :l b4 (28 . . . b2 29 e2) 28 .i.c5 :lxa4 29 .i.xd6 .i.xd6 30 :lxb3, and White has the better chances.

18 cd Black can play 1 8 . . . eS at once, without opening the c-file.

After 19 dS ( 19 de :ld8 20 gS .i.f8} 1 9 . . . b6 20 gS :lf8, he has a perfectly secure fortress.

19 cd e5 After 19 . . . b6 or 19 . . . bS, White plays 20 :lbl! with the better

chances. 20 d5 (13)

20 :lcl �d6 2 1 de .i.xeS is not dangerous for Black. 20 �d6

A refinement on the 1 1 th game in Seville. But before going any further, let us recall how that game (which played an important

13 B

part in the match) proceeded. It might, incidentally, have been worth including it (accompanied as it is by some new annotations) among the 'basic' games in this volume. But we will not break our rule: only games played in 1988 or later are to be specially numbered.

20 . .. .t..h6 21 h4 .t..d2 22 'fJ.d1 .t..a5 If 22 . . . b5, White has 23 lt!c1 !, heading for b3. Gutman quotes

the following analysis by H. Wirthensohn : 23 . . . aS (23 . . . .t..xc1 24 'fJ.xc1 lt!d6 25 'fJ.e1 a6 26 .t..g3 'fJ.e8 27 <;Pf2) 24 lt!b3 .t..b4 (24 . . . .t..c3 25 d6 a4 26 d7 rJ.d8 27 lt!c5 lt!b6 28 rJ.d6) 25 .t..c5 .t..xc5 26 lt!xc5 'fJ.c8 (26 . . . lt!d6 27 lt!d7 lt!xe4 28 lt!xe5 lt!d6 29 'fJ.c1 ) 27 lt!b7 'fJ.c7 28 d6 'fJ.d7 29 'fJ.d5 lt!b6 30 lt!c5, and White i s on top.

23 'fJ.c1 b5 Here 23 . . . lt!d6 is no good : 24 lt!g3 .t..b6 25 J.xb6 ab 26 'fJ.c7

'fJ.a4 27 'fJ.e7 lt!xe4 28 lt!xe4 'fJ.xe4 29 d6 �f8 30 'fJ.xh7 'iPe8 3 1 h5, and Black is doomed.

24 'fJ.c2 ltld6 25 ltlg3 ltlc4 26 ltlf1 ltld6 27 ltlg3 ltlc4 28 g5 �/7 White has consolidated his advantage. If 28 . . . a6 (28 . . . .t..b6

29 a4 lt!a3 30 'fJ.b2 ba 3 1 'fJ.a2), there follows 29 lt!fl lt!d6 30 'fJ.c6 lt!xe4 3 1 lt!g3! lt!c3 32 d6! �f7 33 d7 'fJ.d8 34 .t..c5 'fJ.xd7 35 'fJ.f6+ �g8 36 h5 gh 37 lt!xh5 'fJ.d8 38 'fJ.xa6 .t..c7 39 lt!f6+, and Black is in a bad way.

29 ltl/1 ltld6 30 ltlg3 ltlc4 31 �/1 �e7 Korchnoi suggested 31 . . . .t..b6 32 .t..xb6 ab 33 'iPe2 'fJ.a4, but

a stratagem we have seen before - 32 a4 ! lt!a3 33 'fJ. b2 ba 34 'fJ.a2 - gives White the advantage.

32 .t..c5+ <;Pf7 (14)

33 J:.f2+

14 w


A critical moment in the game. After the obvious 33 �e2, Black would scarcely be able to save himself: 33 . . . i..b6 34 a4 i..xc5 35 ab J:.b8 36 �d3! lhb5 (36 . . . a6 37 �xc4 ab+ 38 �b3 i..d6 39 J:. c6 �e7 40 �e2 J:. a8 41 �c3 J:. a3+ 42 �b2, or 40 . . . b4 4 1 �c1 , with a won position) 37 �xc4 J:.b7 38 lL!fl i..d6 (38 . . . a5 39 �d2, or 38 . . . i..b4 39 �e3) 39 �d2, etc.

33 . . . �gl 34 J:.f6 i..b6 35 J:.c6?? By retreating his bishop to f2, White would preserve the better

chances. Instead of this, I commit a fatal blunder. The sad fact is that the possibility of the black knight jumping to the edge of the board has escaped my attention.

35 . . . �a5 36 i..xb6 �xc6 37 i..cl J:.f8+ 38 �e2 J:.fl 39 i..d6 J:.dl 40 i..c5 �a5 41 �!1 .:tel 42 i..d6 J:.c2+ 43 �d3 J:.xa2 44 �e3 �fl 45 �g4 �c4 46 �xe5+ �xe5 47 i..xe5 b4 48 i..f6 b3 49 e5 J:.xg2 50 e6+ �f8! 0-1.

Now at last we return to the game Naumkin-Neverov, which, as it happens, was decided very quickly.

21 �g3 i..h6 22 i..cS

White proceeds ineffectively and is soon in a difficult position. The immediate 22 h4 was better.

22 23 h4 24 i..f2 25 �e2 26 J:.dl

�c4 J:.c8! i..f4 i..d2 aS!

White's loss of time in moving his bishop about begins to tell: Black creates a passed pawn, which decides the issue.

27 gS bS 28 ltlg3 b4 29 lllfl J.c3 30 lllh2 a4 31 d6 b3 32 ab ab 33 ltlg4 b2 34 d7 .IUS 35 :n <t;g7 36 ltlf6 llld2 37 ltle8+ <t;f7 o-1

So 1 2 J.xf7+ doesn't guarantee victory. One or two inaccuracies, and Black may assume a decisive initiative.

Game No. 5 Seirawan-Popovic

Manila 1990 1 d4 ltlf6 2 c4 g6 3 ltlc3 dS 4 cd lllxdS 5 e4 lllxc3 6 be J.g7 7

J.c4 cS 8 llle2 lllc6 9 J.e3 o-o 10 0-0 J.g4 1 1 f3 ltlaS 12

J.xf7 + ll xf7 13 fg ll xfl+

15 B

14 �xfl cd 15 cd (15)

Curiously enough, this precise position arose as long ago as 1 955, in the game Spassky-Korchnoi from the 22nd USSR Championship. Play continued : 15 ... 1i'd7 16 h3 1i'e6 17 1i'd3 1i'c4 18 1i'd2 (or 1 8 1i'xc4 lll xc4 19 J.g5 e6 20 ll d 1 b5, and despite White's extra pawn and the queen exchange, the initiative is with Black; Panteleyev-Prakhov, Bulgaria 1 970. An improvement was 20 .l:l bl !) 1 8 . . . 1i'a6! 19 1i'c2 ltlc4 20 1rb3 <t;h8 21 <t;g 1 ltld2 (21 . . . .l:l f8 is stronger) 22 J.xd2 1i'xe2 23 .te3, with equal chances. If Black wants, he can force a draw with 23 . . . .l:l c8 24 .l:lfl .l:lc2 25 ll f2 1i'el+ 26 ll fl 1i'e2.

The immediate pawn exchange on d4 at move 14 occurred only once in the Seville match - in the ninth game. Let us take quite a close look at it (recently some new thoughts about it have come to mind!) :


15 . . . flb6 16 �g1 With the queen on b6, 1 6 h3 can be strongly answered by 16

. . . �c4 1 7 ..t.f2 flb2! 1 8 �gl l:t f8 1 9 l:t b l flxa2 20 l:t xb7 .th6! with the initiative.

16 ... fle6 With the white king on g 1 , the line just mentioned no longer

gives Black anything : 16 . . . �c4 1 7 .tf2 flb2 1 8 �f4! l:t f8 1 9 �d3.

17 fld3! White returns the pawn while retaining all the positional trumps.

The position had first been seen (it arose by transposition) in Alfeyevsky-Werner, corr. 1 984. That game went 17 �g3 l:t d8 1 8 l:tc l ( 1 8 l:t b l .txd4 1 9 .txd4 �c6 20 �e2 �xd4 2 1 �xd4 flc4) 1 8 . . . flb6 19 �e2 �c6 20 l:t b l �xd4! 2 1 l:t xb6 �f3+ 22 �f2 l:t xdl 23 l:t xb7 �xh2 24 g4 �g4+ 25 �f3 �h2+, with a peaceful conclusion by perpetual check.

1 7 .. . flxg4 After this, I succeed in obtaining a substantial plus. After 1 7 . . .

l:t d8 1 8 g5 �c4 1 9 .tf2 b5, the play would transpose into Chernin-Malishauskas which we have examined before (see page 3 1 ). (I should point out that the ideas and variations in Games 1 -5 have much in common, while sometimes being distinguished by small nuances; I hope the reader will not get tangled up in the 'undergrowth'.) Instead of the capture on g4, a better idea is 1 7 . . . flc4, aiming to exchange o ff the strongest pieces. This idea was employed in game 1 1 of the Seville match, with the difference that the exchange on d4 took place later (see notes to Game No. 4). White may avoid the queen exchange by retreating with 1 8 fld2 (exploiting the absence of the pawn from c3 and thus gaining a tempo). After 1 8 . . . fla6 19 flc2 �c4 ( 1 9 . . . flc4 is well answered by 20 :t e l ! - emphasising once again that the removal of the cpawns is favourable to White) 20 .tf2 .tf8 (20 . . . .th6 21 h4) 21 g5, and Black's position is devoid of counterplay.

18 l:t/1 l:tc8 The alternative 1 8 . . . l:t f8 is worse : 1 9 :t xf8+ .txf8 20 d5 b6

2 1 �d4! 19 h3! fld7 20 d5 �c4 21 ..t.d4 More precise than 2 1 �d4 �xe3 22 flxe3 l:t c4 23 �e6 flxe6

24 de .td4 25 flxd4 l:t xd4 26 l:t f7 l:t xe4 27 l:t xe7 b5, with equality; but 21 .txa7 b6 22 ..t.b8!? was worth considering.


21 . . . e5 If 2 1 . . . lbe5, then 22 ..txe5 .txe5 23 lbd4 .txd4+ 24 -.xd4

b6 25 e5 -.c7 26 -.g4! is not bad for White. 22 de -.xe6 23 .txg7 r;J;xg7 24 lb/4 -.d6 25 -.c3+ (16)

25 ... �h6

16 B

The only move. There would be a striking finish after 25 . . . -.es 26 lbe6+ ! �g8 (26 . . . r;J;h6 27 'ifc1 + g5 28 l:l.f5) 27 -.xc4 ! l:l. xc4 28 l:l. f8 mate. Another line that fails is 25 . . . r;J;g8 26 lbd5 lbb6 27 lbf6+ �h8 28 -.b2 -.c5+ 29 �h2 -.c3 30 lbe8! lbd7 (30 . . . �g8 31 -.bs l:l.c5 32 lbf6+! ) 31 -.xc3 l:l.xc3 32 l:l.f7 l:l. d3 .(32 . . . l:l. c8 33 lbd6) 33 e5 r;J;g8 34 e6, and White wins.

26 lbd5? Black's king is in quite a dangerous position, and White should

have tried to approach it from the other direction with 26 lbd3! The threat of lbd3-f2-g4+ is then fairly serious (26 . . . lbb6 27 'ifd2+ �g7 28 'ifb2+ r;J;g8 29 -.b3 + �h8 30 l:l.f7, with a winning position).

26 . . . -.e5 27 Vd3 White could have maintained the tension with 27 'ife l . Many

annotators recommended 27 -.b4 instead, but this is not dangerous for Black : 27 . . . 'ifxe4! 28 lbf6 aS! 29 -.bs (29 'ifc3 -.e3 +, or 29 'ifa4 'ifd4+ 30 r;J;h 1 l:l.f8 3 1 lbg4+ �g7, with advantage to Black) 29 . . . 'ifd4+ 30 .l:tf2 (in the event of 30 �h1 lbe3 3 1 'ifxb7 lbxfl 32 lbg4+, the black king fearlessly marches forward : 32 . . . r;J;g5 33 -.e7 + r;J;f4 34 -.f7+ r;J;g3 35 -.xn h5 36 -.f3 + �h4 37 g3+ �xh3 38 lbf2+ -.xf2! 39 -.xf2 l:l. c l + , and wins) 30 . . . -.at+ 3 1 l:l. fl -.d4+ 32 l:l.f2 -.a l +, and the contest could end in perpetual check.

27 . . . �g7 28 !i:Jf6 11d6 29 11c3 11e5 30 11d3 11d6 31 11c3 11e5 32 1ib3 l:l. c7 33 11d3 l:l./7 34 11xc4 l:l. x/6 35 l:l. dl b5

The position is drawn, but Black is playing somewhat recklessly. A simpler method is 35 . . . l:l.e6 36 l:t d7+ l:l.e7 37 l:l. xe7+ 11xe7 38 11d4+ �f7 39 11xa7 'lrxe4.

36 l:l. d7+ �h6 37 11e2 11c5+ 38 �h2 11e5+ 39 g3 11c3 40 �g2 11c4 41 11e3+ g5 42 l:l. d2 11/1 + 43 �h2 11/3

The sealed move. Once again, just as happened in game 7 of the match, I persistently sought winning chances after resumption, and reached an endgame a pawn up - but alas, a drawn result was unavoidable. We will follow this interesting game to the end.

44 11d4 l:l. e6 45 e5 11/5 46 l:l. e2 a5 47 11d5 b4 48 11xa5 11d3 49 l:l.g2 11d4 50 11a8 11xe5 51 11/8+ �g6 52 11xb4 h5 53 h4 gh 54 11xh4 l:l. d6 55 11c4 l:l. d4 56 11c6+ �g7 57 11b7+ �h6 58 11c6+ �g7 59 l:l. c2 l:l. h4+ 60 �g2 11e4+ 61 11xe4 l:l. xe4 62 l:l. c7+ �g6 63 l:l. a7 l:l. e3 64 �h3 l:tc3 65 l:ta8 l:l. c4 66 a4 �g5 67 a5 l:l. a4 68 a6 <i;h6 69 �g2 l:l. a3 70 �/2 �g7 t-t .

15 eS!? This move had first occurred in the game Kir. Georgiev

Ivanchuk, Reggio Emilia 1 989/90. 16 dS

The natural reaction - White creates a passed pawn. Despite the pawn extra, the endgame arising from 16 de i.xe5 1 7 11xd8+ l:l. xd8 holds no danger for Black. Seirawan-H. Olafsson, Reykjavik 1 990, continued 1 8 l:l. c l !i:Jc6 19 g3 l:l. d3 20 i.f4 i.d4 21 !i:Jxd4 l:l. xd4 22 l:l. b l l:l. b4 23 l:l. xb4 !i:Jxb4 24 a4 !i:Jc6 25 �e2 a6 26 i.d2 b5 27 ab ab 28 i.c3 rj;f7 29 <i;e3 �e6. Subsequently Black tried to make something of his passed b-pawn, but without success, and a draw was agreed after another 25 moves.

Nor does White do any better with 16 l:l. cl 11d7 17 de 11xd 1 + 1 8 : xd 1 !i:Jc4 1 9 i.f2 i.xe5 20 l:l. d7 b 6 2 1 i.d4 i.xd4 22 !i:Jxd4 !i:Jd2+ 23 <i;e2 !i:Jxe4 24 !i:Jc6 a5 25 l:t b7 !i:Jc3 + 26 <i;d2 !i:Jxa2 27 : xb6 !i:Jb4, and the game is level; Schiissler-Kudrin, SaintMartin 1 990.

16 !i:Jc4 17 i.f2 11f6 1s �g1 : rs 19 11e1

The position is unclear after 19 i.xa7 b6, or 19 i.c5 l:t c8 20 i.xa7 b6.

19 R.h6 Threatening . . . R.d2.

20 lllg3 W'a6 20 . . . W'xf2+ would be over-hasty : 2 1 W'xf2 R.e3 22 W'xe3,

followed by r:t cl. 21 <lo>h1

The threat was . . . r:t xf2, for example : 21 W'e2 r:t xf2! 22 <lo>xf2 R.e3+; while if 2 1 lllfl , then 2 1 . . . lllb2! is decisive.

21 W'a4 2 1 . . . W'a3 is also worth trying.

1 7 B

22 R.gl! (1 7)

The white queen's problems can be solved later. In Georgiev-lvanchuk, White played 22 "ife2, and there followed : 22 . . . b6 23 h4 R.f4 24 lllfl (24 g5 llld6 25 r:t e 1 was more precise) 24 . . . llld6 25 l:le 1 l:l c8 26 g3 r:t c2 27 W'f3 W'xa2 28 <lo>g1 R.h6 29 g5 R.g7 30 llle3 r:t c7 3 1 lllg4 r:t f7 32 W'e3 W'c2 33 h5 lllc4 34 W'c1 W'xc1 35 r:t xcl gh 36 r:t xc4 hg 37 r:t c8+ R.f8 38 R.e l ! (lvanchuk gives 38 <lo>fl ! b5 39 <lo>e2 b4 40 R.c5 a5 41 r:t a8 b3 42 <lo>d2, with a great deal of play left; or 38 . . . <lo>g7 39 <lo>e2 b5 40 r:t a8) 38 . . . <lo>g7 39 R.c3 R.d6! 40 r:t c6 R.c5+ 41 <lo>g2 r:t f2+ 42 <lo>h1 R.d4 43 R.b4 r:t f7! 44 r:t e6 r:t b7! 45 r:t c6 a5 46 d6? (lvanchuk considers this the decisive mistake, although even after the more stubborn 46 R.a3 b5 47 d6 r:tf7! 48 r:t c7 R.b6 49 .D. c6 R.d8 50 R.b2 r:t d7 5 1 R.xe5+ <lo>f7, Black has a clear plus; not however 47 . . . R.b6?? 48 r:t xb6!) 46 . . . ab! 47 r:t c7+ <lo>f8 48 r:t xb7 b3 49 r:t b8+ <lo>f7 50 d7 b2 5 1 r:t f8+ <lo>e6 52 d8lll+ <lo>e7 53 r:t fl �xd8, and White soon resigned.

One other game is worth mentioning. In Naumkin-Miralles, Voskresensk 1 990, Black answered 22 11re2 with 22 • • • .i.g5, and only after 23 lOft did he play 23 . . . b6. (He could have brought about simplifications with the bold 23 . . . l0d6! 24 .i.c5 l0xe4 ! 25 .i.xf8 11rd4 26 .i.c5 ! 11rxa 1 27 11rxe4 11rxfl + 30 .i.g1 .i.d8, threatening . . . .i.b6.) Naumkin now seized the initiative: 24 h4 .i.e7 25 g5 .i.c5 (25 . . . l0d6 26 l0d2) 26 .i.g3 l0d6 27 l0d2 .i.d4 28 J: c 1 ! 1ixa2 29 J: c6! J: d8? (it was better to take the central pawn : 29 . . . l0xe4 30 11rxe4 11rxd2 3 1 J: xg6+ hg 32 11rxg6+ h8 33 11rh6+, with a draw; sharp play would result from 3 1 h2 11re3 32 11rg4!?) 30 h2! aS (Naumkin gives these variations : 30 . . . 11ra1 3 1 l0f3, or 30 . . . l0xe4 3 1 11rxe4 11rxd2 32 J: xg6+ hg 33 11rxg6+ ci>f8 34 11rf6+ ci>e8 35 11re6+ ci>f8 36 g6 .i.g1 + 37 �xg1 11rxd5 38 11rf6+, with a substantial advantage to White) 3 1 h 5 11ra1 32 l0f3 gh (nor can he save himself with 32 . . . 11rb2 33 l0xd4 11rxd4 34 11rg4 11rxe4 35 11re6+ .!On 36 J:c8!) 33 l0xd4 11rxd4 34 11rxh5 l0e8 35 g6 1-0.

22 b6 Before playing the knight on d6, it is essential to cover the c5-

square. 23 11rc3 J: f7

23 . . . l0d6 is premature because of 24 11rxe5 .i.g7 25 11re6+. 24 J: bl .i.d2

24 . . . 11rxa2 is bad on account of 25 .1: b4, and the pawn on e5 falls. Seirawan recommends 24 . . . l0d2 25 J: b4 11rd1 26 11rxe5 .i.e3 27 11re8 + with a draw, but instead White can maintain the pressure with 25 .1: e 1 .

25 11rd3 After 25 11rb3 11rxb3 26 ab l0d6, White's extra pawn is of no

significance. 25 11rxa2

25 . . . .i.h6 (25 . . . .i.f4 26 l0e2 and l0c3) is answered by 26 a3, and the a-pawn is invulnerable : 26 . . . l0xa3 27 : a t .

26 d6! l0xd6 The only move. On 26 . . . J:t d7, Seirawan gives 27 11rd5+ ci>g7

28 J:t fl ! (28 11re6 J:t xd6 29 11re7+ ci>g8 30 J:t fl .i.f4) 28 . . . .i.f4 29 J:t xf4! ef 30 .i.d4+ h6 3 1 g5 mate. (Editor 's note - it appears that Black can escape with 30 . . . �f8. Therefore White should prefer Stohl's 30 lt:lf5 + ! ! gf (30 . . . ci>f8 3 1 .!Oh6 �g7 32 g5!) 3 1 i.d4+ ci>f8 3 2 11rxf5 + ci>g8 3 3 11rg5+ , forcing mate).


27 'irxd6 'irxb1 28 'irxd2 aS

28 . . . 'irb5 29 h3 'ird7 30 'irc3 'ire? 3 1 'irb3 is more tenacious, although White's chances are still better.

29 'ird8+ �g7 30 'irg5 h6

White threatened the deadly 3 1 lDf5+ . 31 'irxe5+ �h7 32 h3! J:d7

White was intending to bring his bishop into play after 33 �h2. But now the denouement comes instantly.

33 lDh5! 1-0

The debate about the 12 .*.xf7+ variation still continues.

2 Exchange Variation with 7 Jtc4 - other systems

Game No. 6 Yusupov-Kasparov

USSR Ch 1988 1 d4 �£6 2 c4 g6 3 �c3 d5 4 cd �xd5 5 e4 �xc3 6 be .i.g7 7

.i.c4 c5 8 �e2 �c6 9 .i.e3 0-0 10 0-0 .i.g4 11 £3 �a5 12 .i.d3

12 .i.xf7+ attained wide popularity after the World Championship match in Seville, but capturing on f7 is, of course, not obligatory - the bishop may retreat instead. It is most securely placed on d3, although occasionally 1 2 .i.d5 is also seen.

12 cd 13 cd .i.e6 14 .:tel (18)

The once fashionable Sokolsky attack, 14 d5 .i.xa 1 1 5 "tl'xa1 f6, has now practically fallen into disuse. Black is the exchange up and can extricate himself without too much difficulty.

18 B

14 .i.xa2 15 "tlra4

This is more precise than 1 5 d5 at once.

42 Exchange Variation with 7 J.c4 - other systems

15 J.e6 16 d5 J.d7 17 'lfb4 (19)

An alternative is 17 'lfa3, whereupon, apart from the more conventional 1 7 . . . e6, Black has 17 • • • bS!?. After 18 :. fd1, the reply 18 . . . J:.b8 is rather slow, for example : 1 9 'lfb4 a6 20 l0d4 J:. e8 2 1 J.e2 J.e5 22 J:. a l ! J.c7 23 l0e6! J.d6 24 'lfxd6! fe 25 'lfe5 1 -0; Piskov-Lputian, Belgrade 1988. In Ilic-Fercec, Kladovo 1989, Black introduced the important innovation 18 . . . b4!, giving the pawn back but obtaining adequate counterplay : 19 'lfxb4 :. b8 20 'lfel l0b3 21 J:. b l a5 22 .tc2, and now 22 . . . a4 is logical, with a complicated game.

19 B

In this diagram we have a standard position in contemporary practice. Black can now choose between 1 7 . . . e6 (the game continuation) and 1 7 . . . b6. Before going any further, let us examine the latter.

1 7 . . . b6

Safer than 1 7 . . . b5, for example : 18 :fd1 J.e5 19 J.c5 �b7 20 J.xe7 'lfb6+ 21 �h1 aS 22 'lfd2 J:. fc8 23 f4 J.g7? (the bishop should have been placed on d6) 24 e5 b4 25 J.c4 b3 26 J.a3 ! with a won position for White; Balashov-Sibarevic, Lugano 1988.

18 f4 Or 1 8 .ta6 J.c8 1 9 J.b5 a6 20 J.d3 b5 2 1 f4 e5! 22 J.c5

J:. e8 23 J.d6 'lfb6+ 24 J.c5 (24 �h1 J.g4! 25 J.c7 'ife3) 24 . . . 'lfd8 25 J.d6, with a draw; Savchenko-Lputian, Tallinn 1988.

18 . . . e5 Stronger than 18 . . . e6 ( 1 8 . . . J:. c8 19 l0c3 l0b7 20 J.a6 l:.c7

2 1 e5 l0c5 22 J.c4 'lfb8 23 'lfa3! J:. cc8 24 J:. fd 1 J:. fd8 25 .ixc5 !

Exchange Variation with 7 J.c4 - other systems 43

and Black is in serious trouble; Vyzhmanavin-Ivanchuk, Tashkent 1 987) 19 d6 eS 20 f5 : c8 2 1 .!Llc3 J.c6 22 .!Llb5 1Wd7 23 f6! with a clear plus; Balashov-Hansen, Malmo 1 987/8. However, after 19 . . . .!Llc6 20 1Wb3 e5 ! 21 f5 : c8 22 .!Llc3 .!Lld4 23 J.xd4 ed 24 .!Lld5 : xc1 25 .!Lle7+ �h8 26 : xc1 gf 27 ef J.f6, Black has everything in order; Dolmatov-Gavrikov, Kiev 1 986.

19 :j2 In Vaiser-Gavrikov, Tallinn 1 988, Black seized the initiative

after 19 fS?! :e8 20 J.a6? J.f8 2 1 1fc3 b5. At move 20, White should have played 20 d6 J.f8 2 1 J.c4 .!Llxc4 22 1fxc4 1Wf6 23 fg (23 .!Llc3? : ec8 24 gf : xc4! 25 gh+ �xh7 26 : xf6 J.e6 27 .!Llb5 : xe4, with advantage to Black; Heinig-Gauglitz, Berlin 1 988) 23 . . . 1fxg6 24 :f3 1fe6, with complex play (Gauglitz).

19 . . . : c8 20 : cf1 1Wc7 It is hard to give a preference to either side.

17 e6 This position occurrred as far back as 1954, in the game Geller

Liliental from the 2 1 st USSR Championship. Play continued 18 de J.xe6 1 9 : fd l b6 20 J.a6 1Wh4 21 .!Lld4, and now by means of 2 1 . . . J.c8 22 J.b5 J.e5 23 g3 1Wf6 24 f4 J.d6 25 1fa4 J.g4 Black could have kept his extra pawn with a sound position.

Nor does 18 d6 give White anything, for example : 1 8 . . . .!Llc6 1 9 1fxb7 : b8 20 1fc7 : b3! 21 : fd 1 .!Lle5 with the initiative; Razuvayev-Lputian, Sochi 1 987.

In the same Soviet Championship as the game we are annotating ( 1988), Belyavsky-Kasparov continued as follows : 18 .!Llc3 ed ( 1 8 . . . b6 is also interesting : 1 9 f4 ed 20 .!Llxd5 J.e6 2 1 : fd 1 J.xd5 22 J.b5 1ff6 23 : xd5 : ac8 24 : xc8 : xc8 25 e5 1fe6 26 1We4 J.f8 27 J.d7 : c4 28 1Wd3 1fe7 29 e6, with approximate equality; Naumkin-Krasenkov, Vilnius 1988) 19 ed (Black similarly has a good game after 1 9 .!Llxd5 J.e6 20 : fd l J.xd5 etc.) 19 • • • : e8 ( 1 9 . . . b6 is also playable : 20 .!Lle4 : e8 2 1 J.d4 .!Llb3 22 J.xg7 .!Llxcl 23 1Wd4 : xe4! 24 fe .!Llxd3 25 J.h6 f6, with a good game; Utemov-Obodchuk, USSR 1 988) 20 J.f2 J.f8 2 1 1Wb2 (after 2 1 1Wf4 g5! 2 2 Wg3 .!Llb3 2 3 : b1 .!Llc5 24 J.c2 f5! Black would seize the initiative, so White is not justified in playing for a win here) 2 1 . . . J.g7 22 1Wb4 J.f8 23 1Wb2 t-t .

The next two examples are identical with Belyavsky-Kasparov up to move 20.

Yusupov-Timman, Rotterdam 1 988, continued sharply with 20

44 Exchange Variation with 7 J..c4 - other systems

. • . J..e5 2 1 .!tJe4 .tfS 22 .tbS .C. f8 23 J.c5 b6! (not 23 . . . .txe4? 24 fe 11Vh4 2S g3 .txg3 26 .C. c2! and White has an obvious plus) 24 .txf8 11Vxf8 2S 11Vxf8 + �xf8 26 .C.fe 1 ! (after 26 .C. fd1? the exchange on e4 is good for Black : 26 . . . .txe4 27 fe �e7) 26 . . . .!tJb3 27 .C. cd l .!tJd4 28 .ta6 .!tJc2 29 .C.e2 .!tJd4 30 .C. ee l .!tJc2 3 1 .C.e2 .!tJd4 t-t.

One other game by Yusupov likewise ended in a quick draw : 20 . . . b5! 2 1 .C.fd l (or 2 1 .!tJe4 .!tJb7 22 .txbS aS 23 11Va4 .txbS 24 11VxbS .!tJd6, with equality; Naumkin-Mokry, Nemestovo 1 987) 21 . . . .!tJc4 22 J..xc4 aS 23 11Vb3 be 24 11Vxc4 .C. c8 2S 11Vd3 !-t; Yusupov-Smejkal, Munich 1 988.

18 .C. fd1 It seems to me that a very strong move here is 18 .!tJd4, but it

doesn't appear to have occurred in practice yet. 18 ed 19 ed .C. e8 20 .tf2 b5!

This looks like a serious weakening, but Black secures for his knight (via c4) the shortest route to the centre, where the main action is going to take place.

21 .!tJd4 .!tJc4 22 .!tJc6

By exchanging the troublesome knight on c4, White would restore the material balance but could hardly count on an advantage : 22 J.xc4 aS 23 11t'cS be 24 11Vxc4 a4. Yusupov endeavours to extract the maximum from the position.

20 w

22 .txc6 23 de .!tJb2! (20)

Exchange Variation with 7 J.c4 - other systems 45

24 J.xbS! White would lose after the hasty 24 c7? 'irxd3 25 l: xd3 etlxd3

26 cS('ir) : axeS 27 : xeS : xeS and 2S . . . l: c l + . 24 etlxdl 25 c7 'irdS!

It looks as if both sides are playing to win. As Kasparov has pointed out, a draw would result from 25 . . . 'ireS 26 J.xeS etlxf2! (26 . . . etlc3? 27 l: e l ! 'irxc7 2S l: e7) 27 'ire7 etlh3 + 2S gh 'irxeS 29 'irxeS+ : xeS 30 cS('ir) J.d4+ 3 1 �g2 : xeS 32 : xeS+ �g7.

26 J.xe8 The ingenious 26 J.c6 is answered by the calm 26 . . . 'ire6 27

J.xeS etlxf2, and now 2S 'irbS would lose to 2S . . . 'ire3! . On the other hand, 26 'irc4 'irxc4 27 : xc4 etlxf2 2S J.xeS etlh3+ 29 �fl : xe8 30 c8('ir) : xeS 31 : xeS + J.fS 32 gh �g7 leads to a draw.

26 etlxf2 27 c8('ir)

27 J.c6 would be met by 27 . . . etlh3 + . Two other variations also lead to drawn endgames : 27 J.xf7+ �xf7 2S 'irf4+ �e7 29 cS('ir) : xeS 30 : xeS etlh3+ 3 1 gh 'ird1 +, or 27 'irc4 etlh3+ 2S �fl 'irxc4+ 29 : xc4 : xeS etc., as in the note to White's 26th move.

21 w

27 28 : xc8

: xc8 et:Jh3+! (21)

The knight has sped right the way across the board to get its king out of trouble.

29 gh 29 �fl?? loses to 29 . . . 'ird3 + 30 �e 1 'ire3+ 3 1 �dl etlf2+

32 �c2 'ird3+ 33 �c1 'irdl mate.

46 Exchange Variation with 7 .tc4 - other systems

29 1i'dl + 30 �g2!

And now, 30 �f2?? is answered by 30 . . . .td4+ 3 1 �g3 1i'g1 + 32 �f4 1i'e3 + 33 �g3 1i'gS mate.

30 1i'e2+ 31 �gl

i-t There is no escaping the perpetual check: 31 �g3 .teS + 32

f4 (32 �g4? fS+) 32 . . . 1i'e3+ 33 �g2 1We2+ 34 �g1 1i'd 1 + 3S �g2 1i'e2+.

Game No. 7 Polugayevsky-Kudrin

New York 1989 1 d4 �f6 2 c4 g6 3 �c3 d5 4 cd �xd5 5 e4 �xc3 6 be .tg7 7

.tc4 c5 8 �e2 �c6 9 .te3 0-0 10 :t el (22)

In the present game we are still dealing with the main system (Exchange Variation) of the Griinfeld, but turn our attention to a line in which White postpones castling by one move, resulting in play of a wholly different character.

In passing, we should also mention the innovation 10 .:t b 1 !?, employed in Korchnoi-Kasparov, Reykjavik 1988. There followed : 10 . . . �aS 1 1 .td3 cd 1 2 cd b6 1 3 0-0 (immediate kingside activity with 1 3 h4 could come up against counterplay with obscure consequences : 1 3 . . . eS 14 dS fS 1 S hS f4 etc.) 1 3 . . . e6 14 'it'a4 ( 14 'it'd2 .tb7 1 S h4 .:t c8 1 6 .:tfc1 1i'd7 17 hS : xc1 + 1 8 .J:l xc 1 .J:l c8 leads to equality) 14 . . . .tb7 1 S .:t fd 1 .l:t c8 1 6 .td2 �c6 17 .tc3 1i'h4! 18 .te l .D.fd8 19 f3 1i'e7 20 .tbS?! (in Kasparov's view 20 .tf2 is better, leading to equality after 20 . . . �aS, but then Black also has 20 . . . .th6!?) 20 . . . a6! 2 1 .txa6 .txa6 22 1i'xa6 �xd4 (the initiative is with Black, but after 23 �xd4 .txd4+ 24 �h1 , or 24 .:t xd4 1i'cS 2S .tf2, White has adequate defensive resources; instead, he commits the decisive error) 23 .tf2? : a8! 24 1i'd3 (24 1i'c4 bS) 24 . . . .:t a3, winning the queen and with it the game (2S �xd4 .:t xd3 etc.).

10 cd The more placid 10 • . . .td7 is also payable. In Vyzhmanavin

Mikhalchishin, Moscow 1 989, there followed: 11 o-o J:tc8 12 'it'd2 1Wa5 13 d5 �e5 14 .tb3 c4?! l S .tc2 e6 1 6 .:t b l ! b6 1 7 f4 �g4

Exchange Variation with 7 J..c4 - other systems 47

22 B

1 8 de fe 19 e5! and White gained a clear plus. At move 14, Black introduced a valuable novelty in Shirov-Epishin, Daugavpils 1989: 14 . • . e6! 15 f4 .!Dg4 16 c4 ( 16 de J.xe6 17 J.xe6 fe 18 J.f2 : cd8 19 1rc2 .!Dxf2 20 : xf2 1ra6, with the better prospects owing to the threat of . . . 1rc4) 1 6 . . . 1rxd2 1 7 J.xd2 ed 1 8 cd J.b5 19 ..tc4 J.xc4 20 : xc4 b5 2 1 : c2 .!Df6 22 .!Dg3 (in Shirov's opinion White could have maintained equality with 22 .!Dc3 b4 23 e5 be 24 ef cd 25 fg �xg7 26 : xd2, or 24 . . . J.xf6 25 J.xc3) 22 . . . : fe8! 23 d6 .!Dd7! and the initiative passed to Black.

1 1 cd 1ra5+ 12 �n J.d7

An inferior choice is 12 . . . : d8 13 h4 h5 14 1rb3 e6 1 5 d5 .!DeS 16 de .!Dxc4 17 ef+ �h7 18 1rxc4 J.g4 19 f3 : ac8 20 1rb3 : xc1 + 2 1 J.xc l J.e6 22 1rxe6 :t d 1 + 23 �f2 : xh 1 24 1rd6! 'iltxa2 25 1re7!. However, 12 . . . 1ra3!? deserves attention. This novelty occurred in Savchenko-Dimov, Varna 1989. After 1 3 1rb3 1rxb3 14 J.xb3 :t d8 15 d5 .!Da5 16 J.a4 e6 1 7 . J.g5 f6 18 J.f4 e5 19 J.d2 b6, the game is about level.

23 B

13 h4 (23)


13 l:l. ac8 In the diagram position Black has various options. Let us look

at them. The blocking move 13 • • • h5 is inadequate; Polugayevsky gives

14 �f4 �h7 1 5 e5!? l:l. ac8 16 J.d3 �g8 17 e6! fe 1 8 l:l.c5! with powerful threats.

Polugayevsky-Korchnoi, Haninge 1 988, went 13 . . . e5 14 d5, and Black gave up a pawn to no avail : 14 . . . �d4 ( 14 . . . �e7 is safer) 15 �xd4 ed 16 J.xd4 J.xd4 17 1hd4 l:l.ac8 18 �e2! (but not 1 8 �gl? on account of 18 . . . b5 !). Black has no compensation for the pawn, and the game ended quickly: 18 . . . l:l. fe8 19 f3 l:l. xc4 20 l:l. xc4 11fxa2+ (20 . . . J.b5 doesn't help : 2 1 �e3 'irxa2 22 : ee l 'irxg2 23 1Wd2 1Wg3 24 'irh2, etc.) 2 1 �e3 'irxg2 22 l:l. hcl f5 2 3 e5! g5 24 hg 'irh2 25 e6 b5 26 l:l.c7 1-0.

Polugayevsky won in confident style. The main game we are examining will also end quickly in his favour. Does this mean that the outlook as a whole is grim for Black? No - it turns out that 13 . . . l:l. fc8! is a good deal more accurate. After the end of the main game, we shall consider what a difference it makes to occupy the c-file with the other rook.

14 h5 e5 Without this central advance Black can hardly hope for counter

play. Let us look at the alternatives : (a) 14 . . . e6 1 5 hg hg 1 6 e5! (but not immediately 16 'ird3 b5 1 7 J.b3 �b4 1 8 1Wd2 l:l. xc l + 19 �xc 1 �c6 20 'ird3 b4 2 1 �e2 l:l. d8, and Black has very strong counterplay in the centre. LputianDvoiris, Simferopol 1 988, continued 22 J.h6? J.xd4! 23 'irh3 J.f6 24 J.e3 J.c8 25 tt:lf4 !i:Je7, and White got nowhere. 22 �g1 was better) 16 . . . �e7 17 'ird3 l:l. fe8 18 J.d2 'ira4 19 J.b3 l:l. xcl + 20 J.xcl 'irb4 2 1 'irh3 �f8 22 J.h6 �g8 23 J.xg7+ �xg7 24 'irh8+ �f8 25 l:l. h7 �e7 26 'lrg7 �d8 27 'irxf7, and it is all over; Griinberg-Gauglitz, E. Germany 1989. (b) 14 . . . b5 15 J.b3 e5 (in the present circumstances, this counterstroke in the centre is not good) 16 hg hg 1 7 de! (Black was vainly hoping for 17 d5? �d4 1 8 l:l. xc8 l:l. xc8 19 �xd4 ed 20 J.xd4 J.xd4 2 1 1Wxd4 l:l.c 1 + 22 J.d1 f6 23 g3 'ira4! 24 'irxa4 ba 25 �g2 l:l.a l 26 J.f3 l:l. xa2 27 l:l.c1 a6, when the black bishop settles on b5 and the pawn on a4 acquires formidable strength) 17 . . . �xeS 1 8 .l: xc8 J.xc8 ( 1 8 . . . l:l. xc8 1 9 'f4 �g4 20 J.xf7+ !) 1 9 �f4 J.b7? (loses at once, but then Black also has a hard time in other

Exchange Variation with 7 J..c4 - other systems 49

vanat10ns : 1 9 . . . l:l. e8 20 �xg6 �xg6 2 1 "ifh5 J..e6 22 "ifh7+ �f8 23 J..c5+ �e7 24 J..xe6 fe 25 l:l. h3, and wins; or 1 9 . . . b4 20 �d5 l:l. e8 2 1 J..g5 J..b7 22 �g1 J..xd5 23 J..xd5, etc.) 20 �xg6! �xg6 2 1 "ifh5 1 -0; Dautov-Huzman, Kecskemet 1989. Black resigned because he cannot simultaneously defend the knight on g6 and the h7-square.

15 hg hg 16 d5!

More energetic than 1 6 J..d2 1Wb6 1 7 l:l. b 1 "ifc7 1 8 d5 �aS! 19 J..d3 �c4, with a good game for Black; Guseinov-Huzman, Baku 1988.

16 �d4 After 1 7 . . . �e7 1 8 J..g5, White has a solid advantage.

17 �xd4 l:l. xc4 17 • • • eel 18 ..txd4 l:l. xc4 1 9 l:l. xc4 11fa6 ( 19 . . . J..b5 20 J..xg7,

and wins) 20 1Wd3 transposes into the game continuation. At move 1 8, it might seem that Black can improve with 18 . . . J..b5 19 J..xg7 l:l. xc4 ( 19 . . . J..xc4+ 20 l:l. xc4 !) 20 l:l. xc4 ..txc4+ 21 �g1 �xg7 22 11fd4+ , and after 22 . . . f6 23 "ifxc4 "ife1 + 24 "iffl 11fxe4 the endgame is more pleasant for Black (Huzman and Vaikerman). However, in place of the check on d4, Polugayevsky unearthed the quiet move 22 "ifcl! ! (24) .

24 B

In spite of his extra piece, Black is helpless : (a) 22 . . . l:l. c8 23 "ifh6+ �f6 24 "iff4+ �e7 25 1We5+ �d7 26 l:l. h7 l:l.f8 27 "ife6+ �d8 28 1Wd6+ �e8 29 "ilb8+ �e7 30 d6+ ! (b) 2 2 • • • 11fc5 (22 . . . "ilb4 23 a3!) 2 3 1i'b2+ ! f6 24 11fxb7+ l:l. f7 25 l:l. h7+ �xh7 26 "ifxf7+ �h6 27 "ifxf6, and Black has no defence.


(c) 22 • . • f6 l3 .l:l h3! (another quiet move, looking for all the world like a study; instead, the straightforward 23 'lrh6+ �f7 24 'lrh7+ �e8 2S 'lrxg6+ �d8 26 l:l h7 'lre1 + 27 �h2 'lrxf2 28 'lrg7 'lrf4+ leads only to a draw). Now Polugayevsky gives two variations : (c1 ) l3 . . . 'lrxa2 24 'lrh6+ �f7 2S 'lrh7+ �e8 26 'lrxb7 'lra1 + 27 �h2 'IreS+ 28 g3 'lrd4 29 'lrc8+ �f7 30 'lrd7+ �g8 3 1 'lrh7 mate. (c2) 23 . . . J.a6 24 'lrh6+ �f7 2S 'lrh7+ �e8 26 d6! 'IreS 27 'lrc7 J.bS 28 .l:l h7, and it is all over.

18 .l:l xc4 'lra6 As we already know, 18 . . . ed 1 9 J.xd4 J.bS loses to 20 J.xg7.

25 w

19 'lrd3! ed 20 J.xd4 J.b5 (25)

One's first impression is that Black has defended himself while keeping the extra piece. But the unexpected continuation was :

21 'lrh3! J.xc4+ 22 �g1 f6 23 'lrh7+ �f7 l4 .l:l h6!

1-0 A fine game. But let us go back to move 1 3, where Black could

have played a little differently with 13 . . . .l:l fc8!. This innovation was used in Nogueiras-Ljubojevic, Barcelona 1 989. Now, after 14 hS e5 1 S hg hg 16 dS �d4 1 7 �xd4 ed 18 J.xd4 .l:l xc4 ! 19 .l:l xc4 ( 19 J.xg7 'lra6!) 1 9 . . . 'lra6 20 'lrd3 J.b5, we reach the same position as in diagram 2S, with the difference that Black's rook is on a8, not f8. This little nuance reverses the verdict on the position. Black has an escape square on f8 for his king, and already holds

Exchange Variation with 7 i.c4 - other systems 51

victory in his hands, seeing that 2 1 1Wh3 (which settled matters in Polugayevsky-Kudrin) is unplayable; there is no threat of 1Wh7 mate, and Black simply captures the bishop on d4.

I should add that in answer to 14 h5, Ljubojevic refrained from 14 . . . e5, preferring to retreat to the back rank with his knight : 14 • • • ll:ld8. Incidentally, this retreat too was made possible by . . . l:l. f8-c8 - the rooks are not now disconnected. There followed : 1 5 hg hg 16 i.d2? (a most unfortunate move) 16 . . . 1Wa4! (Black seizes the initiative. Mere equality results from 1 6 . . . 1Wb6 1 7 l:l. b 1 i.a4 1 8 1W e 1 1Wf6 1 9 l:l. c 1 ll:le6, while in this line 1 8 . . . 1Wc7? actually loses to 19 l:l. c 1 1Wd7 20 f3 i.xd4 21 1Wh4 1Wd6 22 i.b4!; Dragomai:etsky-Krasenkov, Moscow 1989) 17 i.b3 1Wa6 1 8 �g1 1Wd3 19 l:l. xc8? (it isn't so clear after 19 i.h6) 1 9 . . . l:l. xc8 20 e5 i.e6! 2 1 �fl i.xb3 22 ab l:l. c2! and Black won.

One month later, the position after 1 3 . . . l:l. fc8 arose in a game Polugayevsky-Ftacnik, Haninge 1 989. Polugayevsky now innovated with 14 e5!, and only after 14 . . . ll:ld8 continued with 1 5 h5. Alas, a third beautiful win was not to be: 1 5 . . . i.b5 16 i.xb5 1Wxb5 1 7 hg hg 1 8 �g1 l:l. xc l 19 ll:lxcl ll:le6 20 1Wg4 l:l. c8 21 1Wh4? (in Polugayevsky's view, 2 1 1We4 would have kept some initiative for White) 2 1 . . . 1Wb1 22 1Wh7+ �f8 23 �h2 1Wf5 24 ll:le2 l:l. c2 25 l:l.c1? (this allows Black to pick up a pawn and go into a won ending; it was essential to play 25 l:l. e1 , with drawing chances) 25 . . . l:l. xc l 26 ll:lxc l ll:lxd4! 27 f4 1We4 28 1Wh3, and now instead of 28 . . . ll:lf5 (after which Black took 35 moves to win), Ftacnik demonstrated a quicker method : 28 . . . ll:le6 29 i.d2 g5 30 1Wc3 �g8 3 1 g3 gf 32 gf ll:lxf4 33 1Wc8+ �h7 34 1Wg4 i.xe5 etc.

So the keen debate about the fascinating variation 10 l:l.c1 is far from being resolved.

3 Modern Exchange Variation

Game No. 8 Shirov-Akopian

Tbilisi 1989 1 d4 lDf6 l c4 g6 3 lDc3 dS 4 lDf3 �g7 5 cd lDxdS 6 e4 lDxc3 7 be c5 s : bt o-o 9 �el

It is interesting that the 1976 edition of ECO gave this set-up no more than a cursory mention. Yet in the last few years, the system with 8 : b 1 has virtually become the most popular choice against the Griinfeld Defence, perhaps even surpassing the old variation in which the bishop is brought out to c4.

Another method of developing the white pieces, behind that strong pawn centre which characterises the Exchange Variation, involves an early �c1-e3. In this case, the light-squared bishop is brought to e2 (later than usual), and the queen's rook may occupy either c1 or d1 (after 1Fd1-d2). This variation was an object of dispute in my last match with Kasparov - see Games No. 14 and 1 5.

I would add that the popularity of the variation with : b 1 and .te2 is above all associated with the names of the young grandmasters "Boris Gelfand and Alexander Halifman.

9 1FaS The main line for Black is 9 . . . cd 10 cd, and only then 10 . . .

1Fa5+. We shall deal with it in Games 1 1 - 1 3. The fairly popular moves 9 . . . b6 and 9 . . . l0c6 are examined in Games 9 and 10.

Modern Exchange Variation 53

As for 9 . . . .i.g4, it has gone out of use. White gains the advantage, either with the immediate 10 J:l. xb7, or after sacrificing the dpawn : 10 0-0 cd 1 1 cd .i.xf3 1 2 ..txf3 .i.xd4 ( 12 . . . 11fxd4 1 3 11fxd4 ..txd4 1 4 J:l. xb7) 1 3 J:l. xb7.

10 0-0 Of course, sacrificing the pawn on a2 is part of White's plan.

Whether he will be able to work up a dangerous initiative in return is another question. Incidentally, at the beginning of the 1 980s White used to prefer 10 J:l. b5 11fxc3 + ( 10 . . . 11fxa2 1 1 J:l. xc5 is worse for Black) 1 1 .i.d2 11fa3 12 J:l. a5 11fb2 13 J:l. xc5, and at last Black would take the pawn with 13 . . . 11fxa2. White would then continue 14 J:l. a5, 14 11fc1 , or 14 0-0, leading in all cases to a complex game with chances for both sides.

10 11fxa2 (26) At this point, capturing the c-pawn is extremely risky. White's

best reply to 10 . . . 11Vxc3 is 1 1 d5. Let us see how play might develop: 1 1 . . . 11fa5 (if 1 1 . . . l'hd7, then 1 2 .i.g5 J:l. e8 1 3 11fa4 is unpleasant, while 12 . . . .i.f6 loses virtually by force : 1 3 .i.d2 11fa3 14 11fc2 l'he5 1 5 J:l. b3 l'hxf3 + 16 gf 11fa4 1 7 .i.b5 11fd4 1 8 ..tc3) 12 .i.g5 11fc7 1 3 Wf c l (another good line is 1 3 11fd2 .i.g4 14 J:l. fc1 b6 1 5 h3 ..txf3 16 ..txf3 e5 1 7 d6 11fc6 1 8 -*.e7 J:l. e8 1 9 .i.e2 l'hd7 20 .i.b5 11fb7 2 1 11fc2; Legky-Veingold, Lvov 1 984) 1 3 . . . ..tg4 14 .i.f4 11fc8 1 5 e 5 t2Jd7 1 6 J:l. e 1 (or 16 11fe3, connecting the rooks and increasing the pressure in the centre) 1 6 . . . l'hb6 1 7 d6 ed 1 8 ed, and in return for the pawn White has a highly promising position; Chiburdanidze-Malaniuk, Odessa 1 982.

26 w

11 .i.gS 11fe6 An interesting alternative is 11 • • • t2Jd7 12 .i.xe7 J:l. e8 13 .i.d6

cd 14 l'hxd4 l'hf6 15 f3 l'hd5!? 1 6 J:l. a 1 11fb2 1 7 11fa4 .i.f5 1 8 ed


.i.xd4+ 19 lrxd4 .D. xe2 20 g4 .D.d2 2 1 lrb4 .i.d3 22 lrxb2 .D. xb2 23 .D. fd 1 , with equality; Vaiser-Dvoiris, Bamaul 1984. Vaiser suggests 15 .J:la1 lrb2 16 .lObS!? .ll:lxe4 17 .i.a3 ird2 1 8 .ll:lc7.

12 ird3!? A valuable idea, first employed in Epishin-Henkin, Bamaul

1 988. Formerly, White used to advance one of his central pawns here. For example : 12 eS .D. d8 1 3 lra4 lrc6 ( 1 3 . . . .i.d7 14 dS .i.xa4 1 5 de f6 1 6 ef ef 1 7 .i.c4! .D.e8 1 8 .i.f4 b6 19 .i.dS .i.c6 20 c4! gave White a clear advantage in Yusupov-Tukmakov, Moscow 1 983) 14 irb3 .i.e6 ( 14 . . . lrc7 1 5 .i.c4 .D. f8 1 6 e6 f6 1 7 .i.h4 .ll:lc6 1 8 .i.g3 ird8 19 lra2 cd 20 cd b 6 2 1 d S .ll:leS 2 2 .D.fd 1 �h8 2 3 .ll:lxeS fe 2 4 d6! gave White a large plus in another game Yusupov-Tukmakov, Erevan 1 982) 1 5 c4 cd 1 6 .i.xe7 .D. c8 1 7 .ll:lxd4 .i.xc4! 1 8 .ll:lxc6 .i.xb3 1 9 .ll:lxb8 .i.e6 20 .i.d6 .i.f8 2 1 .i.f3 .i.xd6 2 2 .i.xb7 .i.xb8 2 3 .i.xa8 .i.xeS 24 .i.b7 .J:lc2 2 5 .D. fc 1 .J:l a2 2 6 .l:t d 1 .i.c7 2 7 .J:l a 1 .D. b2 2 8 .l:t db 1 .l:t d2 2 9 .l:t d 1 .l:t b2 1-1; Damjanovic-Schmidt, Athens 1984. This one game serves to illustrate the harmlessness of 1 2 eS.

12 b6 After 12 • . . lrd6 1 3 lre3, or 12 • • • .ll:ld7 1 3 lre3 .ll:lf6 14 .ll:leS!,

the initiative is with White. On the other hand, 12 . . • .D.d8!? deserves to be tried.

13 d5 In this case, 1 3 lre3 is no good : 1 3 . . . .i.a6! 14 dS 1rd6 1 5 c4

e5!, and White has nothing for the pawn.

27 w

13 ird6 (27)

14 e5! The point of White's play. He now obtains the two bishops and


a big advantage in space and development. Still, Black does have two extra pawns.

Again 14 1We3 would be well answered by 14 . . . eS! . 14 -*.xeS IS et:lxeS 'itxeS 16 1Wd2

But not 1 6 f4? 1Wd6 17 1We3 :es, or 1 6 1We3? 1Wxe3 1 7 fe : eS! 1 S d6 eS! 1 9 -*.e7 -*.e6 20 .i.f3 ll!d7 21 -*.xaS : xaS.

16 1Wd6 Perhaps 1 6 . . . et:ld7 does more to limit White's options. The

virtually forced continuation is 1 7 -*.f3 1Wd6 1 S : fe1 f6 (worse alternatives are 1 S . . . et:leS? 19 : xeS! 'itxeS 20 :e 1 W'fS 21 d6 -*.d7 22 de : res 23 -*.xaS : xaS 24 h3, and 19 . . . f6 20 l:t e6! -*.xe6 2 1 -*.f4 1Wd7 22 de 1Wxd2 23 -*.xd2) 19 : e6 fg 20 : xd6 ed 21 l:t e l aS, and the position is unclear.

17 1We3! Here this move is entirely in place, and better than 17 .i.f3

ll!d7 (see the previous note), or 17 -*.f4 eS 1S de 1Wxd2 19 -*.xd2 fe 20 .tf3 : xf3 ! 21 gf .tb7.

17 : e8 But not 1 7 . . . f6 1 S .tf4.

18 .tf3 llld7 The play hinges on some subtle points. In many lines the

counter-stroke . . . e7-eS is good, but not here : 1 S . . . eS? 19 de : xe6 20 : bd 1 ! and wins.

19 .tf4 It is only now that White deviates from the game Epishin

Henkin already referred to. That game went 19 : ret et:lf6 ( 19 . . . f6 20 1We6+ !) 20 c4 .tfS 21 : at, and after 21 . . . a6?! 22 -*.f4 1Wd7 23 h3 hS? 24 .teS �h7 2S 1Wf4 l:t gS 26 'itgS, White quickly won. But instead of 21 . . . a6 Black has 21 • • • eS!, after which White doesn't have adequate compensation for the two pawns, for example: 22 de : xe6 23 .txaS (43 1Wc3 et:le4!) 23 . . . : xe3 24 fe. In Epishin's view, White can improve with 21 l:tbdl -*.g4! (otherwise 22 h3) 22 .txf6 .txf3 23 .teS .txd l 24 .txd6 ed 2S 1Wc3, but Henkin has indicated a line in which Black again has nothing to fear: 25 . . . .te2 26 f3 bS! 27 cb .txbS 2S : xeS+ :xeS, 29 1Wf6 : e 1 + 30 �f2 : e2+ 3 1 �g3 c4 32 1Wxd6 a6. But then, the text move should not have brought White any concrete gains either.


19 Wf6 20 d6 : b8 21 : bd1 (28)

2 1 de is bad : 2 1 . . . :Z. xe7 22 Wd2 .!LieS 23 ..tgS .!Llxf3 + 24 gf :Z. d7! 25 Wxd7 WxgS+, and Black wins.

28 B

21 eS? A serious mistake. He could not have solved his problems with

21 . • . ..tb7 22 de :Z. xe7 (22 . . . ..txf3 23 :Z.xd7 ..tg4 24 : xa7 :Z. bc8 25 :Z.e l ) 23 :Z. xd7! : xe3 24 fe ..txf3 25 ..txb8 Wxc3 26 :Z. xf3 Wel + 27 : fl Wxe3 + 28 ..thl We8 29 :Z. fd l ! and wins. But the essential move was 21 . . • e6!, and already after 22 ..tc6 Wd8 23 :Z.fe l bS it is White who has to think about equalising.

22 ..tg5 Wg7 If 22 . . . Wf5, then 23 ..te4 We6 25 ..tdS WfS 25 f4!, while 22

. . . Wh8 is well answered by 23 ..tc6. 23 ..th6 Wf6 24 ..tc6! :Z. e6?

24 . . . Wd8 25 f4 ..tb7 26 ..txd7 Wxd7 27 fe f5! was more tenacious, although after 28 ..tgS ..te4 29 ..tf6, White has an advantage due to the subsequent h4-h5.

25 ..tg5 Wg7 26 f4 h6

If 26 . . . f6, then 27 fS ! gf 28 ..th6 Wg6 29 :Z. f3 f4 30 :Z. xf4! 27 ..te7 ef 28 Wxf4 g5 29 Wa4 .!Oe5 30 ..td5 ..tb7 31 Wxa7 .!Lld7 32 ..txb7

1-0


Game No. 9 Halifman-Pribyl Leningrad 1989

1 d4 ltlf6 2 c4 g6 3 ltlc3 d5 4 ltlf3 J.g7 5 cd llJxd5 6 e4 ltlxc3 7 be c5 8 l:l. b1 0-0

9 J.e2 b6 10 0-0

Another familiar plan is 10 J.e3 J.b7 1 1 e5 llJc6 12 h4, with a kingside attack. However, in a game Nemet-Korchnoi, Switzerland 1985, Black seized the initiative with ll . . . cd! 12 cd ltla6 1 3 1rd2 ltlc7 14 h4 ltld5 1 5 h5 llJxe3 1 6 fe e6 1 7 .i.d3 l:l. c8. There is also another possibility (after 1 1 . . . cd 12 cd): 12 • • • J.e4! 13 l:l.c l llJc6 14 h4 l:l.c8 1 5 llJg5 J.xg2 1 6 l:l. h2 J.d5 1 7 h5 llJb4 18 hg hg 19 a3 llJa2, and Black is clearly better; Vaiser-1. Sokolov, San Bernadino 1989.

In Cebalo-1. Sokolov from the same tournament, White launched his attack at once with 10 h4!?. His opponent's ineffective response unexpectedly led to a quick and crushing defeat : 10 . . . J.g4 ( 10 . . . J.b7 is more logical) 1 1 '/o>fl !? cd 12 cd llJd7 (and here, the right continuation is 1 2 . . . llJc6 1 3 d5 J.xf3 14 gf ltld4) 1 3 e5! l:l. c8 14 J.e3 llJb8 1 5 h5 ! 1rd5 1 6 hg hg 1 7 1rd2 J:Hd8 18 l:l. h4 ! .i.xf3 19 gf J.xe5? 20 l:l. b5! 1re6 2 1 l:l. xe5, and i t i s all over.

29 B

10 J.b7 1 1 1rd3 (29)

1 1 J.a6 In earlier games, Black played 1 1 . . . cd 12 cd e6 13 .i.g5 1rd6,

but as a rule failed to equalise. Here is one example : 14 'iWe3 l:l. c8 1 5 l:l. fd 1 llJd7 1 6 h4 l:l. c2 1 7 e5 1rc7 1 8 J.d3 l:l.c3 ( 1 8 . . . l:.Xa2 19


l:tbc1) 19 'lre2 .*.d5 20 : be l h6 2 1 .*.e3 l:t dS 22 h5! with a big advantage; Lputian-Lalic, Sarajevo 19S5.

12 'lre3 1 2 'lrc2 promises White less. Chernin-Dorfman, Moscow 19S4,

continued 1 2 . . . cd 1 3 cd 'IreS 14 'lrdl .*.xe2 1 5 'lrxe2 'lra6 16 'lrxa6 lbxa6 1 7 .*.a3 l:t fdS 1 S d5 f5 19 .*.xe7 l:t d7 20 d6, and the players agreed a draw.

12 cd 12 . • . e6 is too timid; after 13 de .*.xe2 14 'lrxe2 be 1 5 .*.f4

'IreS 1 6 .*.d6 :t dS 1 7 e5 'lrc6 l S h4 lbd7 19 c4 ! a6 20 : fe l lbb6 21 'lre4, Black has no prospects whatsoever; Lerner-Gorelov, Ivano-Frankovsk 19S2.

In the last few years, Black has tried moving his queen to cS or d7 without exchanging pawns in the centre. These experiments have been unsuccessful, for example : 12 • • • 'IreS 1 3 d5 .*.xe2 14 'lrxe2 .*.xc3 1 5 e5 'lrf5 1 6 : b3 .*.a5 1 7 lbh4 'lrd7 1 S .*.h6, with advantage; Halifman-Lau, Rotterdam 1 9SS. Or 12 . . . 'lrd7 1 3 de be ( 1 3 . . . .*.xe2 14 'lrxe2 be is safer) 14 l:t xbS!? l:t axbS 1 5 .*.xa6 'lra4 16 'lre2, with a won position; Gelfand-I. Sokolov, Yugoslavia 1 9SS.

13 cd 'lrd7 A recommendation of Grandmaster Gavrikov. A much earlier

game Browne-Martz, USA 19S2, went 1 3 . . . .*.xe2 14 'lrxe2 llJc6 1 5 d5 lbd4 16 llJxd4 .*.xd4 1 7 l:td l .*.g7 1S .*.a3 'lrd7 19 e5, with a clear advantage to White.

14 .*.a3 14 .*.xa6 lbxa6 1 5 'lra3 lbc7! gives White nothing; Halifman

Epishin, Leningrad 19SS, continued 1 6 :t d l :t fcS 1 7 .*.b2 lbb5 lS 'lrd3, and now Halifman recommends lS . . . e6!

Black also obtains a good game after 14 l:t dl?! 'lra4 1 5 .*.d2 .*.xe2 16 'lrxe2 :t c8 17 e5 lba6 18 e6 f6 ( 1 8 . . . f5!) 19 d5 'ifc4 20 'lrxc4 : xc4; Lputian-Zilberstein, Blagoveshchensk 1988. However, Salov's innovation 14 d5!? deserves attention. After 14 . . . .*.xe2 1 5 'lrxe2 'lra4 1 6 .*.g5 l:t e8 1 7 l:t fc l , White obtained a slight edge in Salov-1. Sokolov, Haifa 1 989.

14 .*.xel Another possibility is 14 • • • l:t e8 IS d5 .*.xe2 16 'lrxe2 'lra4 1 7

'lre3 lbd7 l S l:t fc 1 : ac8 1 9 h3 l:t xcl + 20 l:t xc l llJf6 2 1 lbd2 .*.h6!, as in Gelfand-Malishauskas, Vilnius 1 988. However, Pribyl gives IS : fd! as more accurate.


15 'irxe2 .l:l. e8 Black is not in a hurry to bring his knight out. In Gelfand

Groszpeter, Palma de Mallorca 1989, Black played 1 S . . . c!Llc6, and after 16 dS c!LleS 17 c!LlxeS -*.xeS 1 8 f4 �d6 19 .i. b2 e6 20 .1:1. bd1 White gained the initiative and soon won.

16 .l:l. fcl e6 17 .l:l. c3 .l:l. c8

1 7 . . . c!Llc6? loses to 1 8 'irbS .l:l. ac8 1 9 .I:!. bel a6 20 'irxb6. 18 .l:l. xc8+

1 8 .l:l. bc1 .l:l. xc3 1 9 .l:l. xc3 c!Llc6 20 "irbS .l:l.c8 2 1 .l:l. c4 'irb7 leads to equality.

18 'irxc8 19 .l:l.cl 'irb7

An alternative is 19 . . . 'ira6 20 "ire3 c!Lld7 2 1 .l:l. c7 c!Llf6 (2 1 . . . jfa4 22 h3) 22 c!LleS .l:l. c8! 23 .l:l. xf7 "irbS 24 g3 jfbl + 2S �g2 jfxe4+ 26 jfxe4 c!Llxe4 27 .l:l. xa7 �xeS 28 de .l:l.c2, with full equality; Dzhandzhgava-Malishauskas, Uzhgorod 1988. A better line for White, according to Halifman, is 20 jfxa6 c!Llxa6 2 1 �fl. with some advantage.

30 w

20 'irc4 c!Lla6 (30)

This position had first occurred only a few days earlier, in the game S. Ivanov-Pribyl from the same international tournament (Leningrad 1989). After 2 1 dS ed 22 ed .l:l. d8 23 d6 c!LlcS! 24 �xeS be 2S 'irxcS 'ira6 26 h4 .l:l. xd6 27 c!LlgS h6, the players agreed a draw. The main game we are examining was to end in victory for Black, but that was hardly the rightful outcome. For the moment Halifman doesn't want to settle for a draw, and adds fuel to the flames.


21 h4 .l:l. d8 22 e5!? h6 23 h5 gh 24 .l:l. c3 b5! 25 'irc6 'irxc6 26 .l:l. xc6 lt!b8 27 .l:l.c7 .U8 28 ..txf8 �xf8 (31)

31 w

29 d5! White's last chance - he hopes for success by sacrificing another

pawn. As Pribyl has pointed out, a forced draw results from 29 .l:l. xa7 lt!c6 30 .l:l. b7 lt!xd4 3 1 lt!xd4 .l:l. xd4 32 .l:l.xb5 .l:l. d l + 33 �h2 .l:l. d2.

29 ed 30 lt!d4 .l:l.e8! 31 f4 .l:l. e7 32 .l:l.c5 lt!a6 33 .l:l. xd5

In the event of 33 .l:l. c6 lt!c7 34 .l:l. xh6, Black can save himself with 34 . . . �g7! (but not 34 . . . lt!e6? 35 lt!f5 .l:l. c7 36 .l:l. h8 mate) 35 .l:l. xh5 lt!e6 36 lt!f5 + �g6 37 lt!xe7+ �xh5 38 lt!xd5 �g4.

33 lt!c7 34 .l:l. d6 lt!e6 35 lt!c6 .l:l.c7 36 f5 lt!g5 37 f6?

Now all White's previous work comes to nothing. After 37 e6! Black would still have to struggle for the draw.


37 . . . �g8 38 �fl 'it>h7 39 lllb4 a5 40 llld5 .:t el + 41 'it>e2 .:tel+ 42 'it>e3 .:t xg2 43 .:t d7 .:t g3+ 44 'it>e2 .:t g4 45 'it>d3 h4 46 'it>e3 .: a4 0-1

Game No. 1 0 Halifman-Henkin

Leningrad 1989 1 d4 lllf6 2 c4 g6 3 lllf3 J.g7 4 lllc3 d5 5 cd lllxd5 6 e4 lllxc3

7 be c5 8 .:t b1 0-0 9 J.e2 lllc6

10 d5 White is unsuccessful with 10 J.e3 cd 1 1 cd 'ifa5+ 1 2 J.d2

'ifxa2; Paunovic-Ristic, Vrnjacka Banja 1 983. In comparison with the variation 9 . . . cd 10 cd 'ifa5+ , which we examine in Games 1 1 - 1 3, Black has the useful extra move . . . lllc6.

10 llle5 Safer than taking the pawn. Let us look at that possibility : 10

. . . J.xc3 + 1 1 J.d2 J.xd2+ 1 2 'ifxd2 llla5 (moving that knight to b8 or d4 is worse. An example of the latter is 12 . . . ll:ld4 1 3 lllxd4 cd 14 'ii'xd4 'ifa5 + 1 5 'ifd2 'ifxd2+ 1 6 'it>xd2 .:t d8 1 7 'it>e3 b6 18 .:t bc l e6 19 J.c4 e5 20 J.b3 J.d7 21 .:t c7 a5 22 d6, and Black has numerous problems; Kasparov-Natsis, Malta 01 1 980) 1 3 h4 (White may also castle, with e4-e5 to follow) 1 3 . . . J.g4 14 h5 J.xf3 1 5 gf ! (not allowing the knight onto c4). Despite the simplification, White's attacking potential is very noticeable, for example : 1 5 . . . e6 1 6 'ifh4 'iff6 1 7 hg 'ifxg6 1 8 'ifh2 'it>h8 1 9 'it>d2! f6 20 .:t bg l 'iff7 21 de 'ife7 22 .:t g6, and the attack continues; Bjarnason-Griinberg, Biel 1985.

11 lllxe5 J.xe5 12 11Vd2 (32)

White can also play 12 c4 at once, for example : 12 • • • 'ifd6!? ( 1 2 . . . 'ifa5 + occurs more often) 12 'ifd2 J.d4 14 J.b2 J.xb2 1 5 : xb2 e 5 16 0-0 b6 1 7 : b3, with a complicated position; GelfandGhinda, Halle 1 987. Alternatively Black has an interesting countersacrifice of a pawn : 12 • • • f5 1 3 ef J.xf5 14 .:t xb7 'ifd6 1 5 .:t b3. Perhaps he should now go in for immediate exchanges with 15 . . . J.e4 1 6 J.f3 J.xf3 1 7 : xf3 : xf3 1 8 'ifxf3 : f8, trying to exploit his advantage in development. Danner-Shvidler, Biel 1 982, saw instead 1 5 . . . J.d7 16 'ifd3 .i.f5 17 'ifd2 J.e4 1 8 J.g4 J.xg2 19 J.e6+ 'ifxe6 ( 1 9 . . . �h8 20 .:t g 1 J.e4 2 1 .:t h3 with an attack!) 20 de

.i.xh1 2 1 11fe2 l:tf5 22 f3 .J:I. af8 23 .i.h6 .J:1. 8f6 24 11fe4, and Black has worries. A move that looks fairly logical is 12 • • • 11fc7, covering the pawn on b7 and preventing White from castling. A correspondence game Danner-Nesis ( 1 983-5) continued 1 3 h3 f5 14 ef .i.xf5 1 5 .J:I. b3 ..td4 1 6 0-0 e5, and Black achieved a sound position.

32 B

In the first half of the 1980s, Black almost invariably played 1 2 . . . e 6 here, and after 1 3 f4 h e retreated his bishop t o g 7 o r h8. I do not plan to give a thorough study of the positions which then arise, but it is worth reminding the reader of the difference that the choice of retreat square makes - or more exactly, of why it is essential to withdraw the bishop into the corner.

Consider the variation 12 . . . e6 13 f4 .i.g7 14 c4 .J:I.e8 1 5 e5 f6 1 4 d6 fe 1 7 .i.b2 ef 1 8 .i.xg7 <t;xg7 19 0-0 (33) .

33 B

Practice has abundantly shown that 19 . . . .J:I. f8 20 .J:I. xf4 .J:I.xf4 2 1 11fxf4 secures White a distinct plus. But at move 19 Black


cannot play 19 . . . eS, on account of 20 d7!, and the pin on the 7th rank will be lethal.

However, if earlier Black plays 13 . . . .th8, and accordingly 1 8 .txh8 �xh8 occurs, then 19 0-0 can be answered effectively by 19 . . . eS! , giving Black a solid position after 20 :t fdl 'it'd7 2 1 'it'dS :t b8 22 'it' xcS b6; Bikhovsky-Polovodin, Irkutsk 1983 .

All the same, I should point out that after 13 . . . .th8 14 c4 :t e8 15 e5 j6, White doesn't have to play 1 6 d6; he has the stronger 16 f5! (34) .

16 . . . ed

34 B

Other pawn captures are no better: 1 6 . . . ef 1 7 e6! followed by 1 8 0-0 and 19 g4; 1 6 . . . fe 1 7 fg ( 1 7 fe 'it'h4+ 1 8 �d1 doesn't look bad either) 17 . . . hg 18 0-0 ed 19 cd 'it'd6 20 :t b3 .tg7 2 1 :t g3 etc.; 1 6 . . . gf 1 7 :t b3 :t e7, and now 1 8 .tb2, 1 8 d6 and 1 8 'it'h6 are all highly unpleasant for Black.

1 7 fg Practice has also seen 17 e6 d4 1 8 g4 b6 19 .tf3 : b8 ( 1 9 . . .

.ta6!?) 20 'it'g2 'flc7 2 1 0-0 gS 22 .tdS 'it'e7 23 h4, with the initiative; Vaiser-Pribyl, Sochi 1984.

17 . . . hg, and now White has two lines : (a) 18 0--0 : xeS 1 9 .tf3 d4 20 :t e l 'it'e8 2 1 .ta3 .tfS 22 : xeS 'it'xeS 23 .tdS + 'it'xdS 24 cd J..xb l 2S J..xcS, and Black is in a bad way; Cebalo-Raicevic, Citorle 1984. (b) 18 cd : xeS 19 0-0 b6 20 d6 J..e6 21 .tf3 :t c8 22 J..b2 fS 23 'it'h6! and White has a won position; Agzamov-Pribyl, Sochi 1984.

ll b6 A comparatively new idea. Black limits the effect of his

opponent's rook on the b-file, and makes b7 available to his own bishop.

13 f4 i.g7 14 c4 e6

14 . . . e5 looks logical, but the following two examples from recent practice show that it is not so simple for Black to achieve equality.

Epishin-Ftacni�, Belgrade 1988, went 15 0-0 ef 1 6 11Vxf4 i.d4+ 17 i.e3 i.xe3+ 18 11Vxe3 11Ve7 19 11Vf4 (Epishin gives 19 e5! l%e8 20 e6 fe 2 1 i.f3 as even stronger) 19 . . . f6 20 a4! with the better game for White.

Gelfand-Fta<::nik, Debrecen 1989, deviated with 15 . . . f5!? 16 i.b2 1rd6 1 7 1rc3 l% e8 1 8 i.d3 l%e7? (in Ftacnik's view, 1 8 . . . fe 1 9 i.xe4 l%e7 20 f5 gf 2 1 i.xf5 e4 leads to unclear play) 19 ef gf ( 19 . . . e4 20 f6 ed 2 1 fe! and wins) 20 fe! i.xe5 2 1 11Vd2 i.xh2+ 22 �hl i.e5 23 •s5+ 1rg6 24 11Vxe7 1rh6+ 25 �gl 11Ve3+ 26 �hl 11Vh6+ 27 �gl 1re3 + 28 l% f2! i.h2+ 29 �xh2 11Vxe7 30 l% f3! 11Vd6+ 31 l% g3 + �f7 32 : fl h5 33 i.xf5 h4 34 i.g6+ + �g8 35 i.h7 + + ! �xh7 36 l%f7+ �h6 37 i.cl + 1-0.

15 i.b2 Considerably stronger than 15 o-o, which Black used to meet

with 15 . . . l% e8. But 15 . . . i.d4+! is better; after 16 �h1 l% e8 1 7 i.d3 ed 1 8 ed i.f5! 19 i.xf5 gf, Black obtained good play in Novikov-Krasenkov, Odessa 1989. The game continued very sharply and ended in a draw; here is how : 20 i.b2 1rf6 2 1 i.xd4 cd 22 : be l l% e4 23 g4 l% ae8 24 gf 1rxf5 25 l%g1 + �f8 26 1rd3! 11Vxf4 27 11Va3

·+ l% 8e7 28 : eft 11Vh4 ! 29 1rd6 l% g4 30 11Vd8+ l%e8

1 3 1rd6+ : e7 32 11Vd8+ l% e8 33 : xf7+ �xf7 34 l% fl + l% f4! 35 1rc7+ �g8 36 11Vxf4 11Vxf4 37 : xf4 d3! 38 l%d4 �f8 39 l% xd3 l%e 1 + t-t .

15 16 l% xb2 17 e5

i.xb2 l% e8

On 17 0-0, Black would equalise with 1 7 . . . ed. l7 i.b7 18 0-0 ed 19 i.f3 f6 20 l% e1 fe (35)

This position arose for the first time in Kudzhma-Henkin, Kramatorsk 1989. In that game Black obtained a fine position


after 2 1 .I:!. xeS 11Ff6!. There followed 22 11Fel (22 .*.xd5 + was more to the point, giving equality after 22 . . . .*.xd5 23 11Fxd5+ �h8) 22 . . . .l:l. xe5! 23 11Fxe5 (23 fe 11Ff4) 23 . . . 11rf7 24 .l:l. e2 (better is 24 cd .l:l. e8, when Black has a minimal edge) 24 . . . de 25 .*.g4 11Fg7! and White's position is not easy.

35 w

21 fe! A significant refinement.

21 11Fc7 22 .*.xd5+ .*.xd5 23 'it'xd5+ �g7 24 e6 .l:l. ad8 25 11Ff3 .l:l. e7

He should have blockaded the pawn with his queen, although after 25 . . . 11Fe7 26 .1:1. be2 White still has the more pleasant position.

26 .1:1. £2 11Fd6 27 11Ff6+ �g8 28 11Fg5!

This accurate move casts doubt on Black's set-up. After the immediate 28 h4 .l:l. de8, White would have to acquiesce in a draw with 29 11Ff7+ �h8 30 11Ff6+. But now 28 . . . .l:l. de8 is answered by 29 11Fd5! .l:l. d8 30 .l:l. d2! with a won position.

18 �g7 29 'it'f6+ �g8 30 11Fg5 �g7 31 h4! .l:l. de8 32 11Fd5 .l:l. d8 33 11Fg5

Of course not now 33 .l:l. d2? 11Fg3!, and it is Black who wins.


33 :tde8 34 :t dl 11fc6?

As Henkin has shown, the only defence was 34 0 0 0 11fc7. Then 35 h5?! :txe6 36 h6+ �f8 37 :t xe6 :t xe6 38 :t d8+ �f7 39 :t h8 11fd6 40 :t xh7+ �g8 41 :t g7+ �h8 42 :t xa7 :t e l + ! 43 �f2 :t e5 44 :t a8+ would lead to a draw. Still, after the quiet 35 :t del White would retain some positional advantage.

36 w

35 11fe5+ �g8 (36)

36 :t d7! Black's pieces are stalemated.

36 b5 36 0 0 0 h5 doesn't help : 37 :te3, followed by 38 :tg3. If 36 0 0 0

11fa4, White's simplest course would be 3 7 h5 11fxc4 38 :t xe7 :t xe7 39 11fb8+ �g7 40 h6+ ! �xh6 42 11ff8+ :tg7 43 e7.

37 cb 11fxb5 38 hS 11fb6 39 :t xe7 :t xe7 40 h6 c4+ 41 �h1 11fd8 42 :t b1 ! :tc7 43 11ff4!

After 43 11fd5 :t c8! , or 43 11fd4 'tiffS!, Black could still hold on. 43 g5

The threat was 44 e7! 44 11fxg5+!

1-0


Game No. 1 1 Vaiser-Pein

Budapest 1989 l d4 lt!f6 2 c4 g6 3 lt!c3 dS 4 lt!f3 ,j_g7 S cd lt!xdS 6 e4 lt!xc3

7 be cS 8 l:t bl 0-0 9 ,j_el cd

10 cd 'lfaS+ At the present time, the exchange on d4 followed by the check

with the queen is Black's most popular choice in this variation. l l ,j_dl

Formerly, 1 1 'lfd2 was seen more often, but in the resulting endgame White can hardly count on a plus. For that reason, the search for the initiative now proceeds on different lines, and involves a pawn sacrifice. This variation of the Griinfeld occurred only once in my encounters with Kasparov - in the 1 3th game in Seville, where 1 1 'lt'd2 'lfxd2+ 12 ,j_xd2 was played (37) . Let us briefly examine that game.

3 7 B

1 2 . . . e6 (in answer to 1 2 . . . b6 at once, White has various choices : 13 ,j_d3, 1 3 d5 or 1 3 l:t c 1 , but best of all is 13 0-0. Then after 1 3 . . . ib7 14 d5 ,j_a6 1 5 l:t fe l ,j_xe2 16 l:t xe2 lt!a6 1 7 .i.e3, White has the better chances. If 17 . . . f5, Gutman recommends 1 8 �d4! fe 19 l:t xe4 if6 20 lt!e5! . I f 1 7 . . . l:tfe8, he recommends 18 l:t c2!, instead of 1 8 lt!d4 �xd4 19 .i.xd4 e6 20 de, which led to a draw in Halifman-Tseshkovsky, Minsk 1 985) 1 3 0-0 ( 1 3 .i.c4, 1 3 l:t c l and even 1 3 h4 have been seen, but there i s n o point in starting active operations with incomplete development) 1 3 . . . b6 14 l:tfdl (it is hard to decide how to arrange the rooks on the first rank; 14 l:tfe l , 14 l:l. bc l and 14 l:l. fc l have also been played.


As the d-pawn is due to advance, I place a rook behind it) 14 . . . ..tb7 1 5 d5 ed 1 6 ed lDd7 ( 1 6 . . . ..txd5 loses to 17 ..tb4 .l:l. d8 1 8 ..tc4) 1 7 ..tb4 .l:l. fc8 1 8 ..te7 (neither 1 8 tDd4 nor 1 8 ..tb5 gives White anything. But now his initiative looks fairly substantial -for example 1 8 . . . ..tf8 19 d6 ..txe7 20 de tDf6 2 1 .l:l.e 1 ! . However, Black manages to find a powerful retort which solves all his opening problems) 1 8 . . . ..tf6! 1 9 d6 �g7! 20 .l:l. e 1 (20 ..txf6+ is answered by 20 . . . �xf6, but not 20 . . . lDxf6 2 1 1 d7 .l:l. d8 21 lDe5 ! ) 20 . . . .l:l. c5 2 1 ..tb5 ..tc6 22 ..txc6 .l:l. xc6 23 .l:l. bd1 ..tc3 24 .l:l. e3 f6 25 g4 g5 26 h4 h6 27 hg hg 28 lDd4 (now Black seizes the initiative; the correct line was 28 .l:l.c 1 .l:l. ac8 29 ..td8! with equality) 28 . . . ..txd4 29 .l:l. xd4 .l:l. h8 30 .l:l. e l .l:l. c2 (Black could have continued quite actively with 30 . . . .l:l. c3 3 1 �g2 �f7; but now peace ensues quickly) 3 1 a4 a5 32 f4 ! �g6 33 fg �xg5 34 .l:l.fl �g6 35 .l:l. f2 .l:l. hc8 36 .l:l. df4 .l:l. xf2 t-t.

1 1 "Wxal 12 0-0 b6 (38)

The main variations are 1 2 . . . tDd7 and 12 . . . "We6; these will feature in Games 12 and 1 3. There is no point in exposing the knight to a pawn attack : after 12 . . . lDc6 1 3 d5 lDe5 14 lDd4, White has clearly the better chances. But then, in the last few years, practice has shown that 1 2 . . . b6 is also dangerous for Black. The present game will convince us of this.

38 w

13 "Wet! "We6 The alternative is 1 3 . . . ..tb7 14 ..tc4 "Wa4 1 5 ..tb5 "Wa2 (now

of course White can force a draw by repetition, as occurred for instance in Chemin-Dvoiris, Lvov 1990; but that is not much of an achievement) 1 6 .l:l. e l ! (but not 1 6 "Wet? "Wc2! 1 7 d5 ..ta6! 1 8


We3 .txb5 19 J: xb5 �a6 20 l:l. bb1 �c5 2 1 .tb4 aS! , and Black seizes the initiative; Yusupov-Korchnoi, Reykjavik 1 988) 1 6 . . . : c8 1 7 Wd1 e6 1 8 We2 �c6 1 9 We3 l:l. d8 ( 19 . . . e5 is no better : 20 d5 �d4 2 1 �xd4 ed 22 Wf4 .tf8 23 l:l. a l Wc2 26 .i.d7! with more than enough for the pawn; Halifman-Epishin, Vilnius 1 988) 20 : a 1 Wb2 21 .txc6 .txc6 22 .tc3 Wb3 23 J:l. eb1 Wc2 24 �d2 .th6 25 Wxh6 Wxc3 26 �f3 f6 27 l:l. c 1 'lfb2 28 'Wf4! 'it;>g7 29 'Wc7 + .td7, and now instead of 30 'Wd6, as played in Epishin -Mark Tseitlin, Leningrad 1988, Epishin gives 30 : cb 1 ! 'We2 3 1 J:l.e 1 'Wb5 3 2 : xa7 : xa7 3 3 'Wxd8 .te8 3 4 e 5 J:f7 35 h3, with the advantage.

14 .tc4! 15 :e1 16 .tb4 17 J: xe6!

'Wxe4 'Wb7 .te6

A typical exchange sacrifice giving White a dangerous initiative.

39 B

17 fe 18 �g5 (39)

18 �c6 A novelty which, however, does not solve all Black's problems. In Gelfand-Dorfman, Minsk 1 986, Black played 18 • . • �h8,

which gave him a won position after 19 �xe6 �d7 20 .txe7? :fc8. In Vaiser-Andrianov, Naberezhnie Chelny 1 988, White's play was strengthened : 19 J: b3! �d7 20 l:l. h3 �f6? (loses at once. A few days later, in a game Konyashkin-Titlyanov, Black replied 20 . . . h5! , and after 2 1 'Wb1 :rs 22 �f7+ �h7 23 .txe6 �f6 24 .txf5 gf 25 'Wxf5 + 'itfg8 26 'Wg6 'We4 27 �h6+ �h8 28 �f7+, the game had to end in perpetual check. However, the right line was 2 1 'Wd1 ! �f6 22 'Wb1 'Wc6 23 .txe6, with numerous threats)


21 'irb1 ! lt!h5 22 .l:l. xh5 .l:l. f6 23 .l:l. xh7+ �g8 24 'irb3 'irc6 25 'irh3 1 -0.

19 lt!xe6 wh8 20 .tc3

But not 20 .td5? .l:l. fc8 2 1 lt!g5 h6, and Black has everything in order.

20 .t£6 After 20 . . . .l:l. f6 2 1 lt!xg7 wxg7 22 d5 lt!d8 23 d6! .l:l. c8 24 'ire3

e6 (24 . . . ed? 25 .txf6+ �xf6 26 'ird4+ �g5 27 .1:1. b3 and wins) 25 .txf6+ wxf6 26 'ird4+ �f7 27 .l:l. d 1 , White's advantage is indisputable.

21 'irh6! .l:l. g8 The decisive mistake. In Vaiser's opinion, 21 . . . .l:l. f7! 22 iDeS

'ireS 23 .txf7 be 24 .txg6 'irg8 25 .te4 leads to unclear play. 22 .l:l. el .l:l. g7

In his analysis Vaiser has shown that on 22 . . . lt!d8, White attains his goal with 23 lt!f8!

23 g4! lt!a5 23 . . . g5 is no better : 24 .td3! .l:l. f7 25 .tg6 wg8 26 .txf7+

�xf7 27 d5! 24 .td3 'irc6 25 .tal .1:1.£7

Things also go badly for Black in the endgame that results from 25 . . . 'irf3 26 g5 'irh5 27 'irxh5 gh 28 f4 .l:l. gg8 29 wf2 .tg7 30 lt!xg7 �xg7 3 1 .l:l. xe7+ �f8 32 l:t xh7.

26 g5 .tg7 27 d5! 'irxd5 28 .txg7 + �g8 29 .txg6

1-0

Game No. 1 2 Polovodin-Maslov

Leningrad 1990 1 d4 lt!f6 2 c4 g6 3 lt!c3 d5 4 lt!f3 .tg7 5 cd lt!xd5 6 e4 lt!xc3

7 be c5 8 l:t b1 0-0 9 .tel cd 10 cd 'ira5+ tt .td2 'irxa2 12 0-0 lt!d7

Over the last three years, a good many games and analyses relevant to this move have accumulated. I shall here quote the most important of them.


13 .tb4 The natural move, which is usually played almost automatically.

However, in Nemet-Kozul, Liechtenstein 1989, White decided to postpone this active bishop sortie and drive the queen away : 1 3 J:l. a 1 'ire6 ( 1 3 . . . 'irb2 i s also playable) 14 'ir b 1 'irb6 1 5 'ird3 'ird8 16 l:tfcl b6 17 .tb4 ltlf6! (but not 17 . . . .tb7 18 e5!) 18 ltle5 .tb7 19 ltlc6 .txc6 20 .J:I. xc6 aS 21 .td2 (other bishop moves are answered by 21 . . . b5, when White has inadequate compensation for the pawn) 21 . . . J:l. c8 22 .J:I. ac 1 l:l. xc6 23 l:l. xc6 'ira8 24 d5 ltld7. Black's chances are better; here is how the game ended : 25 'ifa3 .td4 26 .th6 J:l. e8 27 .tb5? ltlc5! 28 .J:I. c7 'ird8 29 .J:I. a7 'irb8 30 .txe8 'ifxa7 3 1 'irh3 'irb8 32 .tc6 ltlxe4 33 d6 ltlxd6 34 'ird7 ltlf5 35 .td2 'ire5 0-1 .

40 w

13 ltlb6 (40)

Before proceeding to the game continuation ( 14 ltle5), let us look at some alternatives.

The exchange of the pawns on e4 and e7 took place in VaiserKozul, Ptuj 1989: 14 .txe7 J:l. e8 1 5 J:l. a 1 (or 1 5 .tc5 l:l. xe4 16 .td3 J:l.e8 1 7 J:l. a 1 'irb2 with a draw, but Black also has the interesting reply 15 . . . ltld7!) 1 5 . . . 'ire6 16 .tc5 'ifxe4 1 7 ltle5 .txe5 18 .tf3 .txh2+ 19 �xh2 'iff4+ 20 �g1 ltld7 21 .txa7 ltlf6 22 'irc1 1Wf5 23 J:l. e 1 J:l. xe 1 + 24 'ifxe1 'iff4. White has the advantage of the bishop pair, which compensates for the pawn minus.

An equal game results from 14 .tb5 .td7 1 5 .txd7 ltlxd7 16 .t.xe7 l:l.fe8; Riemersma - Conquest, Dordrecht 1 988. I t is worth considering 14 d5, with the threat of 1 5 .txe7.


The game Halifman-Gavrikov, Moscow 1 988, took the following interesting course :

14 'fldJ!? r1 e8 15 liJg5! Exploiting the unfavourable placing of the enemy queen, White

sets up the threat of 16 f1 a l . The tactical justification, given by Halifman, is : 1 5 . . . .txd4 16 .td 1 ! f1d8 ( 1 6 . . . .te6 1 7 'ifxd4 'ifxb 1 18 liJxe6 fe 19 .tc3, or 16 . . . 'ifc4 17 'iff3 and 1 8 .tb3) 1 7 .tb3 .txf2+ ( 1 7 . . . 'flxf2+ 18 wh1 ) 1 8 �h1 f1 xd3 19 .txa2 and wins.

15 . . . .te6 16 d5 .i..d7 The attempt to take advantage of the queen on the d-file with

16 . . . f1 ad8 is no good in view of 17 de f6 1 8 'flh3 fg 19 .tb5. 17 'fl/3 f5! The only move. 1 7 . . . .tf6 loses at once to 1 8 .txe7!, while 1 7

. . . f6 1 8 .C. a 1 'flc2 19 .C. fc l 'ii'b2 20 ..tc3 'flb3 2 1 .C. cb1 'ifc2 22 .C. b2 is also bad for Black.

18 .td3 .i../6 19 ef Rather recklessly played. 19 liJe6 or 19 h4 was stronger, and

may have given a sharper attack. 19 . . . ..txg5 20 fg 'flxd5 21 .te4 'fle6 22 .i..d2! Not 22 'flh5 h6 23 f4 'flxe4 24 fg, which fails to the intermediate

check 24 . . . 'fle3 + . 22 . . . hg! Other moves quickly lose. 23 .txg5 .tc6 24 .txc6 'flxc6 25 'ii'h3 'fld5 26 .te3 Here White had the much stronger 26 'flh6! 'iff7 27 .te3,

winning the pawn back and retaining a positional advantage. 26 . . . liJc4 27 'flg4 �g7 28 .td4+ J-J. So 14 'fld3 gave White a dangerous attack. However, the

vulnerability of the queen on d3 was emphasised in a game Tukmakov-Gavrikov, Moscow 1 989, which went 14 . . . .i..e6! (41) .

Black threatens 1 5 . . . .tc4, which Makarichev suggested parrying with 15 liJdl .C.fd8 16 d5. But then Black would carry out much the same decisive tactical operation as in the actual game : 1 6 . . . .txd5! 1 7 ed liJxd5! and there is no defence against 1 8 . . . liJxb4 or 18 . . . liJf4. The correct line for White is 15 .C. at 'flc4 16 .txe7 'flxd3 1 7 .txd3 .C.fe8 1 8 .tc5 .tc4 19 .txc4 liJxc4, with a roughly equal ending; or 1 5 . . . 'flb3 1 6 .txe7 f1 fe8 17 .tc5 liJa4 18 'flxb3 .txb3 19 .tb5 liJxc5 20 .txe8 .C. xe8 2 1

41 w


: a3, as in S. Ivanov-Novik, USSR 1 989, and now 2 1 . . . .tc4! 22 de .txfl 23 �xfl , again with an equal position.

Tukmakov played 15 d5??, not even suspecting the danger awaiting him : 1 5 . . . lllxd5!, an attractive stroke which decides the issue at once. After 16 ed .tf5, not only the rook on b1 but also the bishop on e2 is left under attack, and White could very well cease resistance. This actually happened 12 moves later.

14 llle5 (42) 14 : a 1 "ife6 1 5 "ifc2 is interesting, but has yet to be tried out

in practice. It was Neverov who suggested jumping to e5 with the knight, and the move was first played in a game AltermanRogozhenko, USSR 1989.

42 B

14 f6 In the original game with this line, Black decided to eliminate

the knight at once, and was quickly crushed : 14 • • • .txe5 1 5 de "ife6 1 6 "ifd4 .td7 1 7 f4 f6 ( 1 7 . . . .tc6!? was more tenacious; after 18 f5 "ifd7, the play isn't so clear) 18 ef "ifxf6 19 "ife3! :ac8


20 .tc5 .tc6 21 f5 'ife5 22 l:l. xb6! ab 23 .tc4+ l:l.f7 24 .txf7+ ! and White won.

The retreat of the queen to e6 in these circumstances led to disaster in Ivanov-Maslov, Leningrad 1990 : 14 . . • 'ife6 16 fe f5 1 6 .tc5 fe 1 7 .txb6 l S de ab 19 'ifd4 'ifc6 20 l:l. fc l (or 20 .tc4+ .te6 21 .txe6+ 22 l:l. xb6 'iff5 23 l:l. xb7, with the better chances) 20 . . . l:l. a4 (he should have played 20 . . . 'ifd7 2 1 'ifxb6 'ifdS, with a complex struggle) 2 1 'ifdl ! 'ifd7 22 : xeS! and wins.

14 . . . .te6 is also dangerous for Black: 1 5 .txe7 .: feS 1 6 .th4! (stronger than 16 .tc5 .txe5 17 de .tc4 lS .txc4 'ifxc4, with equal chances; Halifman-Kindermann, West Berlin l9S9) 16 . . . g5 ( 1 6 . . . .:t eeS may be better) 1 7 .txg5 (or 1 7 .tg3 : adS l S .tb5 .: rs, as in Sakayev-Novik, USSR 1990, and now 1 9 'ifh5! f6 20 d5 fe 2 1 de 'ifxe6 22 'ifxg5 with a clear plus) 17 . . . f6 1S .:t a l 'ifb3 ( 1 S . . . "ifb2 1 9 .!Lid3 'ifxd4 20 .te3 'ifxe4 was better) 19 'ifd2 l:l. acS 20 .td1 'ifc3 2 1 'ifxc3 l:l. xc3 22 .td2, and White wins.

15 .!Lic4! In the game Shevelev-Lagunov, USSR 19S9, in which 14 . . . f6

was played for the first time, White withdrew his knight to f3, giving Black a good place for his queen : 15 .!Lif3 'iff7! 16 'ifc1 .td7 1 7 'ifa3 l:l. fcS! 1 S .txe7 (otherwise 1 S . . . .tfS) 1 S . . . l:l. eS 1 9 .tc5 l:l. xe4 20 l:l. xb6 (20 .txb6 l:l. xe2 is worse) 20 . . . l:l. xe2 2 1 .l:t xb7 a S 22 'ifd3 .:t eeS 2 3 d5 (43) .

43 B

Now after 23 . . . J:. abS 24 l:l. a7 a4, the game was eventually drawn. Lagunov points out an interesting trap that Black could have set : 23 . . . .:teeS! 24 .ta3 (also after 24 l:l a7 l:l. xa7 25 .txa7 a4, the advantage is with Black) 24 . . . .tf5! 25 'ifb5 .td3! and wins.


Apart from the retreat of the knight at move 1 5, practice has seen t5 :.at 11re6 t6 .tg4 f5 1 7 ef gf 18 .tf3 l:r. d8 19 .l: e 1 .txe5 20 .l: xe5 11rxe5 21 de .l: xd 1 + 22 .l: xd 1 �f7, with a roughly equal game; Neverov-Malishauskas, Podolsk 1 989. Or t6 �d3!? 11rxe4? ( 1 6 . . . 11rf7 is safer) 1 7 .tf3 11rf5 1 8 .l:e 1 .l:e8 1 9 .tc5 .td7 20 .txb7 .l: ab8 2 1 .l: xa7 with the advantage; Agrest-Maslov, USSR 1990.

t5 �xc4 t6 :at �bl!

t6 . . . 11rb2 loses the queen to 17 .txc4+ �h8 18 .l: b l ! . On the other hand, t6 . . . �e3 transposes into the game after 1 7 : xa2 �xd l .

44 B

t7 .l: xal �xdt t8 .tc4+ �h8 t9 : xdt (44)

How should we assess this rook and bishop ending? Black is a pawn up, and has two connected passed pawns on the queenside. White has powerful pawns in the centre, and the more active bishops. Overall, White's position looks the more attractive, and indeed in the present game his conduct of the ending brings success. (From now on, I shall make use of the winner's notes.)

t9 .td7 After t9 . . . a6 20 f3 : e8 21 .tb5 .te6 22 d5 ab 23 : xa8 : xa8

24 de .tf8 25 : d7 : a6 26 .txe7 .txe7 27 : xe7 : b6 28 : e8+ �g7 29 .l: b8 .l: xe6 30 : xb7+ (threatening .l: xb5), White would gain a clear advantage (Almeira). But a better move, undoubtedly, was t9 . . . : d8, clarifying the position in the centre at once : 20

J.xe7 l:l e8 21 J.c5 l:l xe4 22 l:l xa7 l:l xa7 23 J.xa7. Material equality is restored, and White has a minimal edge.

20 J.xe7 l:l fe8 21 J.d6

The game Petursson-Almeira, Lyon 1990, went 2 1 J.c5 b6, and only then 22 J.d6. After 22 . . . J.e6 23 d5 J.d7 24 J.d3 f5 25 ef gf, White utilised the weakening of a6 with 26 J.a6!

21 .l:ld After 21 • • • J.f8 22 J.xf8 .a. xf8 23 .td5 ..i.c6 24 J.xc6 be 25 .a. a6 .a. fc8 26 f3, the double rook endgame is difficult for Black. The ingenious tactical trick 21 . • . .a. xe4 22 .td5 .te6 doesn't tum out so well after 23 .txe4 J.xa2 24 J.xb7. A safer line for Black is 21 • • • .i.e6 22 d5 J.d7 23 J.d3 f5 24 ef gf.

22 .i.f7 a5 22 . . . J.f8 23 J.xf8 .a. xf8 24 J.d5 J.c6 25 J.xc6 be 26 .a. a6

similarly leads to a bad ending for Black. 23 f4!

The advance of the centre pawns quickly decides the game. 23 .l:l a6

Black could resist more stubbornly with 23 . . . J.f8 (23 . . . .i.h6 24 g3) 24 .i.xf8 .l:l xf8 25 J.d5 J.c6 26 ..i.xc6 be 27 .l:l c l .l:l fd8 28 .l:l a4, although even then it isn't simple to hold the ending.

24 eS ..i.f8 25 J.d5!

A tactical resource which consolidates White's advantage. After 25 J.xf8 .l:l xf8 26 .i.d5 b5! , the black pawns would already be much the more dangerous.

25 .l:l a7 If 25 . . . J.xd6, the intermediate move 26 J.xb7 is decisive.

Nor can Black save himself with 25 . . . .l:l xd6 26 ed b6 27 l:l bl J.xd6 (27 . . . .a. b8 28 l:l xa5!) 28 .a. xb6 .i.xf4 29 f2 .txh2 30 .l:l xf6.

26 J.xf8 l:l xf8 27 e6 J.c8

Other moves likewise fail to rescue Black : 27 . . . J.c6 28 J.b3 a4 29 d5 .i.b5 30 d6!, or 27 . . . i.b5 28 l:l b2 .ta4 29 l:l a l .tc6 30 J.c4.

28 l:l el bS If 28 . . . b6, then 29 .tc6 l:l e7 30 d5, with a won position.

29 .tc6 l:l e7


30 d5 b4 Or 30 . . . J:[ d8 3 1 .i.xb5 : xe6 32 : xe6 .i.xe6 33 de J:[ xd 1 +

34 �f2 J: d2+ 3 5 �f3, and the pawn queens. 31 : at

1-0

Game No. 1 3 Gelfand-Ivanchuk

Tilburg 1990 I d4 .!Zif6 2 c4 g6 3 .!Zic3 d5 4 cd ll:lxd5 5 e4 .!Zixc3 6 be .i.g7 7

.!Zif3 c5 8 : bt 0-0 9 .i.e2 cd 10 cd 1Wa5+ I I .i.d2 'lfxa2 12 0-0

The present game will feature 12 • • . .i.g4 - we shall come to it shortly.

In Games 1 1 and 1 2, we looked in detail at 12 • • • ll:ld7 and 12 . . . b6. Since the black queen must return to its own camp sooner or later, the immediate 12 • • • 1We6 is also sometimes played. Let us digress a little from our principal game and consider this manoeuvre. White protects his pawn with 13 1Wc2, giving this position (45) :

45 B

The queen doesn't feel very comfortable on e6, and it is now best to move it to d6 or c6.

Gelfand-Kindermann, Debrecen 1989, went 13 • • • 1Wd6 14 d5 b6 ( 14 . . . .!Zia6 is safer) 1 5 .i.b4 1Wd8 1 6 J:[ fd 1 .!Zia6 1 7 .i.a3 ll:lc5 1 8 .!Zid4 .i.xd4 19 : xd4 1Wd6 20 1Wc3 1Wf4 21 .i.b5 .i.d7 22 .i.xd7 ll:lxd7 23 .i.xe7 J:[ fe8. At this point, as Gelfand has shown, White could have won beautifully with 24 d6 l:t ac8 25 J:[ c4 ! 1Wxe4 (25 . . . J:[ c5 26 l:t xc5 .!Zixc5 27 1Wd4) 26 l:t xc8 1W xb 1 + 27 1Wcl 1Wxcl +


28 .lhc1 fS 29 .l:l. c7 ll:lf8 (29 . . . ll:lcS 30 .I:!. xeS !) 30 h4, etc. Instead he played 24 g3?, and after 24 . . . 11'eS 2S -*.a3 .l:l. ac8 26 11Fd2 .l:l. c2 27 11'xc2 11'xd4 28 .l:l. e 1 ll:lcS 29 -*.xeS be 30 .l:l. d 1 11'xe4 3 1 11'xcS 11Ff3!, a level ending arose and a draw soon resulted : 32 .l:l. c l (32 11Fd4? .l:l. e4!) 32 . . . .l:l. d8 33 .l:l. e 1 .l:l. xdS (33 . . . 11FxdS? 34 .l:l. e8+) 34 .l:l. e8+ �g7 3S 11'f8 + �f6 36 11'e7+, 1-t .

Interestingly enough, the position in diagram 4S occurred in another game between the same players a year later (Dortmund, 1990). This time, after 13 . . . 11Fd6 Gelfand innovated with the immediate 14 ..i.b4!. There followed 14 . . . 11'd8 1 S dS -*.g4 ( 14 . . . b6 transposes into the previous game, which should not have turned out well for Kindermann) 1 6 .l:l. fc 1 ll:la6 17 -*.a3 -*.xf3 1 8 gf! (White has obtained a definite plus) 1 8 . . . 11'd7 19 -*.xa6 ba 20 11'c7 11Fh3 21 11Fg3! 11'xg3 22 hg, and Black has a miserable position.

Gelfand-Kamsky, Til burg 1 990, went 13 . . . 11'c6 14 11Fd3 11'd6 1 S .*.b4 11'd8 16 11'a3 (of course White had no reason to exchange queens, but then his queen didn't feel too comfortable on d3 either) 16 . . . ll:lc6 1 7 .i.cS (46) .

46 B

After lengthy manoeuvres ( . . . 11Fd8-aS-a2-e6-c6-d6-d8), the black queen has returned to its 'rightful' place. Several tempi have been lost, and in addition White is threatening d4-dS. However, Black found an interesting tactical resource : 1 7 . . . .*.xd4 !? 1 8 ll:lxd4 lllxd4 1 9 ..i.xe7 ll:lxe2+ 20 � h 1 11'e8 21 .*.xf8 11'xf8 22 11'e3 aS (while White surrounds the knight, the black a-pawn speeds forward) 23 11'xe2 a4 24 .l:l. bS 11'e7 2S .l:l. a 1 .*.e6 26 eS a3 27 11Ff3 a2 28 11'xb7 11Fd8 (Black is the exchange down, but his far advanced a-pawn enables him to save himself) 29 h3 .l:l. c8 30 .l:l. b2


'ifgS 3 1 J:t bxa2 (forcing the issue) 3 1 . . . J:t c l + 32 J:t xc l 'ifxc l + 33 �h2 .txa2 34 'ifaS+ �g7 3S 'ifxa2 'iff4+ 36 g3 t-t .

And now, one more digression : the new move 1 2 . . . aS!? was played in Hebert-Kozul, Toronto 1990. No other examples are available, but in this case the advance of the rook's pawn brought Black success : 1 3 .tgS a4 14 .tbS (better is 14 'ird3) 14 . . . .td7 IS .txd7 l0xd7 16 J: xb7 'ife6 17 dS 'ifxe4 IS J: xd7 a3 19 : xe7 'irb4 20 : eel a2 21 .te7 'irbS 22 'ird2? (the right course was 22 .txf8 a l ('if) 23 'ifxal : xal 24 .txg7 : xe l 2S :xe l �xg7 26 h4! 'irxdS 27 J:t e3, with equality) 22 . . . : res 23 d6 al ('if) 24 : xal : xa l 2S : xal .txal 26 h3 'ird7, and Black won.

12 .tg4 Black leaves his pawn on b7 undefended, but takes aim at the

white d-pawn. 13 J: xb7 .txf3 14 .txf3 .txd4 I S eS l0a6 16 J:t xe7 : adS leads to sharp play, with material equality.

13 .tgS! h6 In the game Sakayev-Bukhman, USSR 19S9, Black retreated

with 1 3 . . . 'ife6, and there followed : 14 dS (at this point, taking the b-pawn doesn't pay at all : 1 4 : xb7 'ifxe4 I S : xe7 'irdS, with advantage to Black) 14 . . . 'ifxe4 (or 14 . . . 'ird7 I S 'irb3 b6 1 6 'ifa3, and White has the initiative) I S 'ird2! f6 (the preliminary I S . . . a S was better: 1 6 J:t xb7 f6 1 7 .te3, with compensation for the pawn) 16 .te3 aS 1 7 :fc l ! l:t cS 1 S h3 .td7 19 J:t xcS+ .txcS 20 .td3 'ifa4 (20 . . . 'irxdS loses to 21 J:t c l 'ird7 22 .tc4+ �hS 23 .te6! 'ifxe6 24 'irdS+ 'irgS 2S : xeS) 2 1 .tcS 'ird7 22 'ife2 .tfS 23 .tc4 �h8 24 : b6 l0a6 2S .td4 l0c7 26 lOeS! fe. White could now have won immediately with 27 'ifxeS+ �gS 2S d6+ e6 29 de.

14 .te3 l0c6 Ivanchuk twice reached this position in the Manila Interzonal,

1 990. Gurevich-Ivanchuk continued : 14 . . . b6 l S h3 .txf3 16 .txf3 eS 1 7 dS l0d7 I S :e 1 J:tfdS 19 'ird3 lOcS 20 'irbS .tfS 2 1 : ed 1 l0b7 22 J: a l t-t .

IS dS lOaS In the game between the same opponents in Manila, Black

preferred I S . . . lOeS, and after 1 6 : xb7 e6 1 7 :e l .txf3 I S gf J:t fdS 19 d6 'ira3 20 d7 'ife7 2 1 .tbS 'irh4 22 .tfl a draw was agreed. White has a strong passed d-pawn, but his king's residence is weakened, which gives Black sufficient counterchances. In the present case Ivanchuk is attempting to improve on his play, but


Gelfand succeeds in exploiting the position of the knight on the edge of the board.

16 .tc5 .tf6 17 e5!

The positional sacrifice of a second pawn. 17 .txe5 18 h3 .txf3 19 .txf3 lUeS

In a detailed examination of the game, Gelfand and his trainer Kapengut indicate the stronger 19 . . . .l:t ae8 20 d6 (20 :t e l lllc4 2 1 .l:t xb7 .*.d6 22 .l:t xa7 11Vd2 leads to equality) 20 . . . ed 2 1 .td5 lllc4 22 11Vg4 (the tempting 22 .l:t xb7 de 23 11Vf3 llld2 24 11Vd3 comes up against 24 . . . .th2 + ! 25 �xh2 lllxfl + 26 �gl 11Val ! -lvanchuk; but 22 .tb4! is worth considering) 22 . . . de 23 11Vxg6+ (if 23 .txc4, Black can defend his g-pawn with 23 . . . 11Vc2) 23 . . . .*.g7 24 .l:t xb7 llle5! 25 11Vxg7+ �xg7 26 J.xa2 .l:t b8 27 .l:t xa7 : a8, with equality.

20 d6! Now it turns out that Black cannot take the d-pawn, so he has

to part with the exchange. 20 11Vc4

The endgame which now arises is difficult for Black, but the alternatives are thoroughly bad: 20 . . . ed 21 .td5 ll\c4 22 .l:t xb7, or 20 . . . 11Ve6 21 d7 .l:t ed8 22 11Va4 lllc4 23 .txb7 .l:t ab8 24 .txa7.

21 d7 11Vxc5 22 de(11V)+ .l:t xe8

For the moment Black is quite well off from the material point of view, with three pawns - more than the equivalent - for the exchange. His pieces, however, are badly co-ordinated, and this leads to more losses.

23 11Va4 23 11Vd7 looks tempting, with the aim of exploiting the exposure

of the black king after 23 . . . 11Vc8 24 11Vb5 J.c7 25 .l:t fc l .l:t d8 26 1t'c5 lllc6 27 .txc6 .td6 28 11Ve3! . But the raid with the queen would be refuted by 23 . . . �f8! 24 .l:t b5 11Vc7.

23 lllc6 24 .l:t xb7

After 24 :t fc l 11Vd4 25 .l:t c4 11Vd7 26 .l:t xc6 be 27 .txc6 11Vf5! 28 J.e4 11Vc8, it is Black who has the better game.

24 .l:tc8


25 -*.xc6 25 .l: b5 11rc3! 26 .i.xc6 11rxc6 is insufficient to win.

25 : xc6 26 .J: xe7

Already the advantage in material is with White. 26 • • • 11rxe7 27 11rxc6 -*.d4 28 11rd5 -*.b6 29 g3 b5 30 �gl

�g7 31 .J: bl 11re2 32 11rf3 11re6 White's advantage is indisputable, but to convert it to a win he

needs to exchange queens. 33 11rc3+ �h7 34 .l: el 11rd5+ 35 11rf3 11rd7 36 .J:dl 11re7 37

.l: d3 �g7 38 11rc6 11re2 39 l:t£3 11ra2? 40 11rc3+ �g8 41 11rb3 White's main task is accomplished; the rest is simpler. 41 • • • 11rxb3 42 : xb3 �f8 43 �f3 �e7 44 : bl �e6 45 �e4

�d6 46 : dl+ we6 47 �£4 �£6 48 : c2 The rook cannot be prevented from penetrating the Black

position. 48 . . . -*.d4 49 �e4 .tb6 50 .l:l. c6+ �e7 51 f4 �d7 52 .1:1. (6

�e7 53 .l:l. c6 �d7 54 �d5 -*.c7 55 .l:l. c4 aS 56 .l:l. c6 -*.d8 57 .l:l. a6 �e7 58 �e4 .tc7 59 �dS .td8 60 .l:l. a7+ �f6 (47)

4 7 w

61 J: a8! Now if 61 . . . .tc7, White wins with 62 �c6, as the bishop is

trapped. If 61 . . . �e7, then 62 �c6, and an original zugzwang decides the game; once the possible kingside pawn moves are exhausted, Black must move his a-pawn for lack of anything else.

61 -*.e7 62 .l:l. xaS

1-0 White cannot be stopped from approaching the f-pawn with his

king and then winning it with the aid of the rook, after which further resistance is senseless.


Game No. 14 Karpov-Kasparov

World Championship (15th game) Lyon 1990

Before proceeding with this game, I should explain that nearly all the material for the present book was assembled before my last match - the fifth - with Kasparov; and that when I began putting the manuscript in order immediately the match was over, I allotted two places to the four Griinfeld games played in New York and Lyon. The point is that the ninth game of the match was spoilt by my blunder in the ending, while in the thirteenth I never managed to extract much from the opening; so I decided not to include either of them among the principal (numbered) games. The fifteenth and seventeenth match games were a different matter - they were assigned to places 14 and 1 5. But once I set about annotating the first of these games - with my impressions of it still fresh - I unexpectedly realised that it contained a far greater number of interesting subtleties and unusual situations than might at first have been supposed. Immersing myself in the analysis, I detected - strangely enough - more and more new ideas and problems which had to be looked into. The result was that whereas in other cases I had merely told the story of a game, here I produced something more like a novel.

To be sure, by the exhaustive analysis, by the sheer abundance of the variations given, this game stands clearly apart from all others in the book. What is offered here for the reader's attention is virtually the definitive theory of an important Griinfeld system that was tested out in this match.

1 d4 l0f6 2 c4 g6 3 l0c3 d5

Ever since the World Championship Match of 1986, the Griinfeld Defence has been firmly incorporated into Kasparov's repertoire, and together with the King's Indian it today constitutes his main weapon against the closed openings. Obviously his choice of the Griinfeld in the present match was no surprise to me.

4 cd Essentially, this seems to me to be more logical than 4 .tf4 (to

which I turned during the 1 986 match), or 4 l0f3 .tg7 5 •b3 de 6 •xc4 (another popular line which was thoroughly tested both


in 1 986 and in the Seville encounter of 1 987). The merit of the plan initiated by 4 cd is that White creates a strong pawn centre within the shortest number of moves!

4 lDxdS 5 e4 lDxc3 6 be j_g7

There was a time when it was thought essential to start undermining the pawn centre at once with 6 . . . c5. The argument was that otherwise (after 6 . . . j_g7), White could play 7 j_a3, as recommended by Simagin, and hold up Black's flank counterthrust for a long time. But now the bishop move to a3 is considered fairly innocuous on account of 7 . . . lDd7 (or 7 . . . 0-0 8 lDf3 lDd7) 8 lDf3 c5.

7 j_e3 The development of forces which commences with this move is

less thoroughly explored than 7 j_c4 c5 8 lDe2, or 7 lDf3 c5 8 : b1 (7 'ili'a4+ and 7 i.b5 +, which are coming into fashion, do not appeal to me). Together with my team of seconds I had prepared this variation specially for the match. It must be said that, in the main, my hopes of gaining an opening advantage with it were not unrealised.

7 c5 8 'ii'd2 (48)

White still has quite a wide choice of continuations here. Thus, 8 j_b5+ is playable, as are 8 lDf3 and 8 :Z.c 1 , two lines which tend to transpose.

48 B

This is the position that formed the starting point for the theoretical contest in the games (and analysis! ) of the 1990 match.


Kasparov endeavoured four times to equalise for the Black side, in opposition to my efforts to develop an opening initiative. The overall score ( + 1 = 3 to White) may be considered acceptable to both sides, and testifies rather to the problematic character of the variation than to the attainment of a definite verdict.

8 0-0 In the ninth game, where I tried this system for the first time,

Kasparov chose a more straightforward plan, exchanging pawns in the centre. That game is also a very important one, so we will now embark on a long digression to look at it in detail :

8 . . . cd 9 cd lllc6 10 r:J. dl ECO only considers 10 lllf3 .t.g4 1 1 r:J. d 1 (nor can White count

on anything after 1 1 .i.b5 0-0 12 .t.xc6 be 1 3 : c 1 .t.xf3 14 gf 1rd7) 1 1 . . . 0-0 1 2 .t.e2 : cs 1 3 0-0 b6 ( 1 3 . . . .t.xf3 14 .t.xf3 e5 1 5 d5 llld4 doesn't look bad either) 14 d5 .t.xf3 15 gf llle5 1 6 r:J. c l 1Wd7 1 7 f4 lllc4!. After 1 S .t.xc4 1rg4+, Black is guaranteed a draw; Haik-Granda, Dubai 01 1 9S6.

10 . . . 1Va5 (49) The game Kozul-Dorfman, Marseille 1 9S9, in which the inno

vation 10 r:J.d1 was tried out, went 10 . . • e6 1 1 lllf3 0-0 1 2 .i.b5!? (the tempting 13 .i.h6 is parried by 1 3 . . . 1Wa5 14 1rxa5 lllxa5 1 5 .t.xg7 �xg7 1 6 .t.d3 .t.d7 1 7 �d2 : res, o r by the quite good alternative 1 3 . . . f5!? 14 .i.xg7 �xg7 1 5 e5 1rd5 16 .t.e2 b6 1 7 0-0 ib7, retaining counter-attacking possibilities in the middlegame) 12 . . . id7 1 3 0-0 (after 13 ih6 ixh6 14 1Vxh6, the reply 14 . . . 1Wa5+ is rather risky on account of 1 5 :d2. Similarly, 1 3 . . . llle5 14 lllxe5 ixb5 1 5 ixg7 �xg7 is inadequate in view of 1 6 h4. But 1 3 . . . f5 14 .t.xg7 �xg7, as in the previous note, is fully viable, since 15 e5 is met by 15 . . . lllxe5 16 de ixb5 17 1Wb4 1Vb6 - the simple 17 . . . ireS is also possible - 1S lllg5 :res 1 9 l:l. d6 ic6 20 lllxe6+ �gS, with complications; or 1 S llld4 r:J. ad8, with the idea of answering 1 9 1Vxb5 with 19 . . . : xd4!, or 19 1re7+ with 19 . . . �gS! 19 a4 id7, or 19 lllxb5 with 1 9 . . . r:J. xd 1 + 20 �xdl a6 2 1 1We7+ r:J. f7) 1 3 . . . a6 ( 1 3 . . . llle5 is unconvincing; although 14 de ixb5 1 5 1Vb4 ireS 16 : fel b6 is satisfactory for Black, the unassuming 14 ie2 gives White the better prospects) 14 ie2 1ra5 1 5 d5 1rxd2 (of course not 1 5 . . . ic3? 1 6 de! ixd2 1 7 ixd2, with a won position) 1 6 : xd2 ed 1 7 ed llle7, and the ending clearly favours White. For example, he can play to restrict the knight's mobility with 1S d6 lllf5 ( 1 S


. . . �c6 19 ..tb6) 19 ..tb6 and then g2-g4; or else 1 8 ..tc5 l::tfe8 19 d6 �f5 20 g4 �h6 2 1 h3, with the advantage.

In Seirawan-Sokolov, Novi Sad 01 1 990, Black played the modest 10 . . . 0-0, but after 1 1 �f3 ..i.g4 1 2 ..te2 l:t cS 1 3 0-0 1rd7 14 h3, he sacrificed his bishop : 14 . . . ..th5 1 5 g4 ..txg4 1 6 h g 1rxg4+ 1 7 �h2 1rxe4 1 8 �g5 1rfS 19 f4 e 6 20 ..td3 1rf6 2 1 ..te4. Black's compensation for the piece i s insufficient, and White eventually prevailed.

Whatever the reason, Kasparov preferred to exchange queens . . .

49 w

1 1 1i'xa5 �xa5 12 �!3 This endgame had arisen once before, in Yusupov-Gulko,

Linares 1 989. On that occasion White preferred 12 ..td3 0-0 1 3 �e2 ..td7 1 4 l:t c 1 l:t fc8 1 S �d2 e6 1 6 : xeS + : xeS 1 7 l:t c 1 l:txcl 1 8 �xcl �fS, and the minor piece ending was quickly drawn after 19 �e2 a6 20 ..tf4 �c6 2 1 ..td6+ �e8 22 �c3 ..tf6 23 ..tc7 ..te7 24 f3.

12 . . . 0-0 Here 12 . . . ..tg4 is premature on account of 1 3 ..tb5+ . 13 ..te2 ..td7 (50)

14 ..td2! Typical of Grtinfeld endings. If the knight retreats, then after

14 . . . �c6 1 S dS �eS (in the event of 1 S . . . �d4 1 6 �xd4 ..txd4 1 7 ..tb4 ..tf6 1 8 f4 aS 19 ..ta3 ..tc3 + 20 �f2 l:t fe8 2 1 lt b1 , or 20 . . . ..tb4 2 1 ..txb4 ab 22 l% d2, Black has many problems) 1 6 �xes ..txe5 1 7 f4 ..tc7 1 8 :c t : res ( 1 8 . . . ..tb6 19 ..tb4 : res 20 �d2) 19 ..te3 e6 20 ..tc4, White goes ahead with his policy of constriction. 19 �f2 is also playable.


50 w

14 . . . b6 15 0-0 After pondering it for a long time, I refrained from doubling

Black's pawns with 1 5 .t.xa5 ba, for he would then acquire an adequate initiative. However, as Zaitsev has shown, 1 5 .t.a6!? was more promising. The analysis of this endgame points to the conclusion that the entire variation, beginning with 8 . . . cd, fails to secure equality for Black.

15 . . . 'D.fd8?! This appears to be the root cause of his subsequent troubles. A

more natural choice seems to be 1 5 . . . .l:fc8 1 6 :c 1 (if 1 6 .t.a6, then 1 6 . . . 'D. c2 1 7 .t.xa5 : xa2 1 8 .t.b7 'D. b8. Therefore White should first play 1 6 .t.xa5 ba, and then 1 7 .t.a6 'D. c2 18 'D. c 1 : xa2 19 : c7 'D. d8, but even so, Black i s risking nothing after 20 h3 'D. a3! 2 1 : xa7 'D. xf3 22 gf .t.xd4 and 23 . . . .t.xh3) 1 6 . . . .t.g4 ( 16 . . . lDc6 is problematic; Black aims to answer 1 7 .t.a6 with 1 7 . . . lDxd4 1 8 .t.xc8 lDxf3 + 1 9 gf .t.xc8 20 �g2 .t.e6, o r if 20 : c7, then 20 . . . .t.h3. In either case Black has definite compensation for the exchange, but is it sufficient for equality? Thus, in the last variation, after 21 'D. fcl e5 22 'D. b7!, the rook cannot be dislodged from the seventh rank : 22 . . . .t.c8 23 : xa 7!) 1 7 .t.a6 (on 1 7 d5, Black has 1 7 . . . .t.b2, when neither 1 8 'D. b1 .t.xf3 19 gf 'D. c2 20 'D. fd 1 .t.c3 nor 1 8 : xc8 'D. xc8 1 9 'D. b 1 .t.xf3 20 gf 'D. c2 21 'D. d l .t. c 1 presents him with any danger) 1 7 . . . : xc 1 1 8 'A xe l .t.xf3 19 gf .t.xd4 20 'D. c7. White regains the sacrificed pawn, but cannot achieve more than that.

16 : ci .t.g4 (51) Again a straightforward decision. Perhaps it was worth prepar

ing an exchanging operation with 1 6 . . . a6!? 1 7 'D.bl ( 1 7 : c7 .t.b5 1 8 .t.xb5 ab 19 d5 lDc4 20 .t.b4 .t.f6 2 1 .t.xe7 .t.xe7 22


: xe7 : xa2 leads to complications, since 23 10d4 b4 24 10c6? b3 2S 10xd8 b2 looks too hazardous for White, while 23 : b 1 fS !? 24 lOgS fe 2S 10xe4 10d2 should lead to a draw) 1 7 . . . .tbS 1 8 .txbS ab 19 : xbS 10c4 20 .tb4 : xa2 21 : c 1?! .txd4! 22 : xc4 .txf2+ 23 �fl :d 1 + 24 10e 1 .txe1 2S .txe1 : aal . Nor does 1 7 .tf4 produce the expected effect: 17 . . . .tbS 1 8 .txbS ab 19 .tc7 :t d7 20 .txb6 10c4. Or if 1 7 .tgS, then 17 . . . h6! 1 8 ..txe7 : es, and Black restores the balance. However, it is worth considering 1 7 dS!? .tbS 1 8 :fe 1 : ac8, with possibilities for both sides.

1 7 d5 10b7

5 1 w

Devices such as 1 7 . . . e6 1 8 h3 .txf3 ( 1 8 . . . ed 1 9 hg de 20 .txaS ef 2 1 .txf3 ba 22 .txa8, and Black has a lost position) 1 9 .txf3 ed 20 ed fail to help Black - he i s still faced with the same inferior ending.

18 h3 White would be wasting time with 1 8 .tb4 e6, while 1 8 : c7?!

is a mistake owing to 1 8 . . . tOeS! ( 1 8 . . . .l:t d7 19 : fc 1 , with the initiative) 19 : xe7?! .tf6 20 : c7 10xe4.

18 . . . .txf3 Nor is 1 8 . . . .td7 1 9 .tb4 e6 20 .ta6 any better for Black. 19 .tx/3 10c5 20 .te3 : ac8 21 .tg4! :t b8 Or 2 1 . . . : c7 22 .tf4 :t b7 23 :t fd 1 , and White has a large

plus, since 23 . . . 10xe4? loses at once to 24 : cS : xeS 2S .txc8. White also retains a substantial advantage with 22 f4 (with a view to 22 . . . hS 23 .tf3, and the pawns creep forward) 22 . . . fS !? 23 ef gf (23 . . . .l:txdS 24 fg hg 2S .l:t fd 1 .l:t xd 1 + is also a grim outlook for Black) 24 .txfS : xdS 25 : xeS! : dxcS (if 2S . . . : exeS, then 26 ..te6+ �f8 27 .txcS : xeS 28 :t d 1 ! , and Black is in a difficult

position after either 28 . . . .tb2 29 .l: d8+ �g7 30 g4 �f6 3 1 .tb3, o r 2 8 . . . .ih6 2 9 .l: d8 + �g7 3 0 g3! �f6 3 1 .tb3) 26 .te6+ �f8 27 .l: d 1 �e8 (27 . . . .l: c8?? fails to 28 .txcS, and 27 . . . .tb2 28 .txcS be 29 .l: d8 + �g7 30 g4 �f6 3 1 .tc4 brings Black no relief either) 28 .txcS be 29 .1: b 1 ! (astonishing domination with full material equality) 29 . . . .td4+ 30 �h2 �f8 3 1 .l: b8+ <t;g7 32 g4 <t;f6 (he can't save himself with 32 . . . .l: c6 33 .l: g8 + �f6 3 4 .tdS .l: d6 3S gS+ �fs 36 : f8+ �g6 3 7 .te4+ �g7 38 : e8 .l: e6 39 .tdS .l:e2+ 40 <t;g3 hS 41 �h4, or 40 . . . h6? 41 .l: g8 + �h7 42 g6 mate) 33 .ic4 e6 34 .l: e8 .l: e7 3S gS+ �f7 36 .l: c8 �g6 (36 . . . �g7? 37 fS ! , or 36 . . . .te3 37 �g3) 37 �g3 h6 38 h4 hg 39 hg eS 40 : c6+ cJi;g7 41 fS, with a simple win.

The reader will agree that these variations are pretty, although rather long. But now to return to the game, i.e. the 9th match game :

22 .l: c4 A mistake would be 22 .txcS be 23 .1: xeS .1: b2 24 .1: aS .id4,

when White's material advantage is neutralised by Black's pressure against f2.

22 . . . h5 A difficult ending results from 22 . . . lDxe4 23 .l: xe4 fS 24 .txfS !

gf 2S .l: xe7, or 22 . . . fS 23 .txcS be 24 ef gf (not 24 . . . .l: xdS? 2S f6!) 2S .txfS .id4 26 .te6+ �g7 27 �h2 (27 .l: a4 .l: b7 28 .l: d 1?! is less convincing in view of 28 . . . .l: b2 29 .l: xa7 .txf2+ 30 �h2 .th4, when neither of the alternatives bring a clear result : 3 1 d6 .tf6! 32 d7 - not 32 de?? .teS+ 33 �g1 .l: xd 1 mate - 32 . . . .teS+ 33 cJ;g1 �f6 34 ..tg4 e6 3S : a6 .td4+ , or 31 .l:fl .tf6 32 .l: xe7+ .txe7 33 .l: f7+ �g6 34 .l: xe7 .l: xa2 3S d6 .l: a6! 36 d7 .l: xe6 37 .l: xe6+ �f7 38 .l: h6 �g7 39 .l: d6 c4) 27 . . . .l: b2 28 .l: a4 .l: b7 29 f4, with a technically won position for White.

23 .t/3 e6 24 'IJ. dl ed 25 ed Stronger than 2S .l: xdS .l:e8! , when Black is defending actively. 25 . . . .ie5 (52) At first sight it seems that on blocking the d6-square, Black will

have constructed an almost impregnable fortress. 26 g4! This breaking-up of the pawn chain dispels any illusions of

equality. The endgame was and still is unattractive for Black. 26 . . . hg 27 hg lDb7?! Perhaps 27 . . . .1: bc8 or 27 . . . .1: b7 was more logical, but even

52 w


so, by deploying his forces in the same way as in the game -g4-g5, ..;.gl-g2, and the transfer of the rooks to the h-file - White would clearly have the better chances.

28 TJ. a4 For good measure, Black's knight is now driven onto a bad

square, since 28 . . . TJ. a8 fails to 29 d6. 28 . . . li:Ja5 29 g5!? It is also worth considering 29 TJ. e4!? 29 . . . TJ. bc8 30 J:.e2 J:.d6 If 30 . . . TJ. c2, a good reply is 3 1 J:.d3 TJ. c3 (3 1 . . . TJ. b2? 32 J:.c1 ,

and the rook is trapped) 32 J:.a6, after which White again carries out the plan beginning with �g1 -g2 .

31 �g2!? More accurate than 3 1 J:.a6 TJ. c2 32 ..;.g2, which could be met

by 32 . . . J:.c5 33 J:.d3 TJ. c3; whereas now, 3 1 . . . TJ. c2 32 J:.d3 would immediately put Black in a critical position.

3 1 . . . J:.c5 (53)

53 w


32 J.d2?? An absurdity! I make the second move of the variation 32 J.f4

J.d6 33 J.d2, which would give White somewhat the better chances. If instead 33 J.c1 J.fS 34 J.b2 : cS ! (bl,lt not 34 . . . J.g7 35 J.f6!, when 35 . . . J.xf6 36 gf : d6 37 : h4 : xf6 3S .J:I. dh 1 , or 37 . . . g5 3S .J:I. h6, leads to catastrophe) 35 .J:I. h4 J.g7 36 J.f6 .J:I. dxd5! 37 .l:l. dh 1 .J:I. xg5 + ! 3S J.xg5 : xg5+, the play is quite sharp. White's best line is 32 J.f4 J.d6 33 J.xd6 (33 J.a6 : c7 brings White to a dead end) 33 . . . : xd6 34 .J:I. h4 : cS (34 . . . : c2 loses to 35 : dh 1 �fS 36 : hS+ �e7 37 : e 1 �d7 3S J.g4+ �c7 39 : cS + ) 35 .l:t dh 1 �fS 36 : hS + �e7 37 J.f3, with a considerable endgame advantage.

32 . . . : xd5 33 J.f3 : dd8 34 J.xa5 i-J. It is not without good reason that we have so thoroughly

examined the first duel with this variation. It is useful for the reader to understand that the assessment of the endgame plays a key role in this particular Griinfeld line.

Instead of S . . . 0-0 or S . . . cd, Black frequently chooses 8 • . . "if aS 9 .J:I. bl cd (the prophylactic 9 . . . a6, defending against .J:I. b5, can be met by 1 0 : c 1 !?. White sets up the positional threat of d4-d5 - since the c3 point is defended! - and compels his opponent to go into an ending with 10 . . . cd 1 1 cd "ifxd2+ 1 2 �xd2. This ending is even more favourable to White, since the point b6 in Black's camp has been weakened. Epishin-Gutman, Frankfurt 1 990, saw instead 10 . . . li'ld7 1 1 li'lf3 0-0 1 2 J.d3 li'lb6 1 3 J.h6! cd 14 J.xg7 �xg7 1 5 cd "ifxd2+ 16 �xd2, and White had the advantage). The game Kozul-Polajzer, Ptuj 1 9S9, continued : 10 cd "ifxd2+ 11 �xd2 0-0 12 li'lf3 e6 ( more precise than the immediate 1 2 . . . : dS 1 3 J.d3 li'lc6 14 d5 li'la5 1 5 �e2 b6 1 6 .J:I. hc 1 e6 1 7 J.g5 f6 1 S J.d2, when White's characteristic bishop manoeuvre to shake up the black position brings tangible results) 13 J.d3 lLic6 14 : bet .l:t d8 15 eS f6!? ( 1 5 . . . f5 1 6 h4 h6 1 7 �e2 J.fS 1 S g3 lLib4 19 J.c4 b6 20 li'le1 ! is less effective for Black; Damjanovic-Kozul, Belgrade 19S9. Also 1 S .J:I. c4 !?, recommended by Zaitsev, deserves serious attention. Black cannot develop his queenside freely : 1S . . . li'le7 19 .J:I. g 1 ! li'ld5 20 g4 b5 21 : xeS : axeS 22 gf ef 23 J.xf5, and White has more than enough for the exchange) 16 �e2 fe 1 7 de li'lxe5 1 S li'lxe5 J.xe5 19 J.e4 : bS 20 : cs J.d6?! (according to Informator 47, 20 . . . J.d4 2 1 : as e5 22 J.xd4 .J:I. xd4 23 f3 gives White a big advantage, but this


verdict is obviously exaggerated. After the simple 23 . . . ..te6 24 .l:t xa7 ..tc4+ 25 �e3 b5 26 f4 .l:t bd8, threatening 27 . . . .l:t xe4+ 27 �xe4 ..td3+, Black is out of danger. If instead 24 . . . .1:1. bd8 25 .1:1. bxb7! .l:t d2 26 �e3, the position remains unclear; neither 26 . . . ..tc4 27 f4!, nor 26 . . . .l:t xa2 27 .l:t g7+ �f8 28 .l:t xh7, nor 26 . . . .l:t xg2 27 .l:t g7 + �f8 28 .l:t xh7 gives Black a cast iron draw) 21 l:l. a5 a6 22 ..tb6 .l:t d7 23 h4! ? ..tc7 24 �e3 ..txb6+ .l:t xb6, and White succeeded in realising his positional advantage.

We can see that this result was due to the somewhat unfortunate 20 . . . ..td6?! . However, at move 1 6, White had an opportunity to strengthen his play : 16 ef (the difference between 1 5 . . . f6 and 1 5 . . . f5 now disappears) 16 . . . ..txf6 1 7 �e2 il:lxd4+ (the attempt to develop with 1 7 . . . e5 leads to a worse position : 1 8 de il:lxe5 -not 1 8 . . . .txe5? 1 9 .l:t xc6 and wins - 19 il:lxe5 .txe5 20 .te4 .l:t b8 21 ..txa7! .l:t a8 22 .tb6 .l:t xa2+ 23 �el .l:t e8 24 .td5+ .te6 25 .l:t c8! ) 1 8 il:lxd4 ..txd4 1 9 .txd4 .l:t xd4, and now White has the better chances after either 20 �e3 e5 21 .l:t b5, or 20 .l:t c7 b6 (20 . . . e5 2 1 .1:1. bel .l:t d8 22 .l:t e7) 2 1 .1:1. be l .td7 22 �e3 e5 23 f4 ! .l:t d8 24 .tc4+ �h8 25 fe.

Evidently it is due to considerations of this kind that Black has recently been giving preference to the move-order starting with 8 . . . 0-0.

9 il:lf3 Earlier, one of the main lines here was 9 d5, but then it was

discovered that Black obtains a fine game with 9 . . . e6!, for example: 1 0 .txc5?! 'iWc7! 1 1 .td4 .txd4 1 2 cd ed 1 3 e5 il:lc6 14 .te2 f6 1 5 ef .1:1. xf6 1 6 il:lf3 .tg4 1 7 0-0 .1:1. e8, and already Black has the initiative. Perhaps an even more accurate method is 1 1 . . . ed!? 1 2 .txg7 �xg7 1 3 ed .l:te8 + 14 .te2 .tg4 1 5 'iWd4+ �g8 16 f3 ( 1 6 d6 'iWc6) 16 . . . .tf5, and it is obvious that Black's initiative compensates for the sacrificed pawn.

9 .l:tcl leads to the same kind of position as in the game, if Black replies 9 . . . 'iWa5. It acquires original overtones only in the case of 9 . . . llld7 (54) .

There can follow : 10 ..td3!? Better than 10 d5 il:lf6 1 1 f3 ( 1 1 .td3 b6 1 2 c4 ll:lg4 1 3 .tf4

e5!) 1 1 . . . e6 (after 1 1 . . . b6 1 2 .tb5 il:le8 1 3 .tc6 .l:t b8 14 .tf4?! ll:ld6 1 5 e5 il:lc4 1 6 'iWe2 il:la5 1 7 e6 il:lxc6 1 8 de .1:t a8 1 9 ef+ .l:t xf7 Black obtains a good position, but the correct line for White is


54 w

1 3 .i.xe8!? .l:l. xe8 14 ll'le2 e6 1 5 c4 ed 1 6 .i.a6 1 7 �f2 1Wh4+ 1 8 g3 1Wh3 19 ll'lf4 1Wd7 20 .l:l. he l . Now, a s Vaiser demonstrates, Black should play 20 . . . .l:l. xe3! 21 �xe3! id4+ 22 �e2 g5 23 ll'ld3, with complications. In Vaiser-Romanishin, Sochi 1983, Black neglected this opportunity and ended up in a difficult position after 20 . . . .l:l. e7 2 1 �g2 .l:l. ae8 22 d6! .l:l. e5 23 ll'ld5) 1 2 c4 .l:l.e8 1 3 .!Oe2 b6 14 ll'lc3 ( 1 4 e 5 .!Oxd5!) 14 . . . .i.a6 1 5 ie2 (or 1 5 e 5 ll'ld7 16 f4 f6, and White can't maintain his centre) 15 . . . .!Od7 16 0-0 .!Oe5 1 7 llld 1 ed 1 8 cd ixe2 1 9 'ifxe2 f5! and Black firmly seizes the initiative; Spassov-Tseshkovsky, Moscow 1985.

10 . . . e5! This pawn-break is the point of Black's whole set-up. If now 1 1

d5, Black plays 1 1 . . . f5! and obtains strong counterplay. 1 1 ll'lf3 ed 12 cd cd 13 ll'lxd4 An interesting alternative is 1 3 ixd4!? ixd4 ( 1 3 . . . ll'lf6 14

h3 .l:l. e8 1 5 0-0 ll'lxe4 1 6 ixe4 ixd4 1 7 .i.xg6 ixf2+ 1 8 'ifxf2 hg 1 9 ll'le5 leads to a double-edged position) 14 ll'lxd4 lllf6 1 5 ib1 'it"b6 16 0-0 .l:l. d8 1 7 .l:l.fd 1 ig4 f3 .l:l. d7 1 9 fg .l:l. ad8 20 .l:l. c4 ll'lxg4, with unclear play (Yusupov).

13 . . . ll'le5 14 ie2 'it"h4!? Or 14 . . . ll'lg4 1 5 ixg4 .i.xg4 1 6 0-0, which is also good for

Black. 15 0-0 1Wxe4 16 lbb5 ie6 1 7 .1:1./dl! .i./6 18 llld6 'it"h4 19

i/4 .1:1./dB 20 'it"e3 ll'lg4 21 'tW/3 ie5!? 22 ixe5 ll'lxe5 23 'it"xb7 ll'lg4 24 .i.xg4 'if xg4

The chances are equal; Yusupov-Timman, Belgrade 1 989. These fairly authoritative examples show that in the current

state of theory, 8 1Wd2 0-0 9 .l:l. c 1 is inadequate to give White an advantage.


9 'iF aS Iri the seventeenth match game, Kasparov refrained from the

queen excursion to aS, preferring 9 • • . .tg4 (see Game No. 1 5). I should also mention the variation 9 . . • cd 10 cd .tg4 l t : c1

.txf3 1 2 gf e6 1 3 .tb5 ll:lc6 (the pawn sacrifice is justified, since after 1 3 . . . a6 14 .ta4 it is not simple for Black to complete his development) 14 .txc6 be 1 5 : xc6 11Vd7! 1 6 : c4 (or 1 6 : a6 11Vb5 17 'iFe2 11Vb1 +, and White achieves nothing) 16 . . . f5! (precisely calculated - 17 0-0 fe 1 8 fe e5! , and to avoid coming under attack White has to permit a draw with 19 d5 ! 11Vg4+ 20 �h1 11Vf3 + etc.) 1 7 d5 fe 1 8 fe ed 19 11Vxd5+ 11Vxd5 + 20 ed : fd8, and Black has a satisfactory endgame, as 2 1 : cs can be answered by 2 1 . . . .tf8 ! 22 : bS a6; Vilela-Armas, Bayamo 1 988. White might consider the prospects offered by II : bl!? or II ll:lg5!, on the lines of Game No. 1 5.

10 :ci e6 (55) Another popular line here is the transition to an ending with

10 . . . cd 1 1 cd 11Vxd2+ . Compared with the variation 7 ll:lf3 c5 8 : b l , the rook on c 1 is more mobile, which increases White's possibilities. Play might continue :

12 �xd2 More precise than capturing with the knight. Let us take a close

look at the latter : 1 2 ll:lxd2 e6 1 3 ll:lb3 (after 1 3 .tbS .td7! 14 .txd7 ll:lxd7 1 5 �e2 : fc8 16 ll:lc4 .tf8 17 .tf4 : c6 1 8 ll:la5?! : a6 19 ll:lxb7 : xa2+ 20 �f3 e5! 21 de :e8 22 : a t ! : b2! White had some difficulties to overcome in Hiibner-Adorjan, Bad Lauterberg 1 989) 1 3 . . . : d8 14 .tgS f6 1 5 .te3 .tf8 (or 1 5 . . . f5 1 6 ef gf 1 7 .t b5 lhd7 1 8 0-0 ll:lf6 19 .tg5 .td7 20 .tc4 b6 21 lhd2 : e8 22 ll:lf3 .tc6 23 llJeS .tdS 24 .tbS : ec8 25 a4 ll:le4 26 .tf4 .tf8 27 f3, with advantage to White; Karpov-Hiibner, Tilburg 1980) 16 .tc4 llJc6 1 7 �e2 .td7 1 8 : hd l �g7 1 9 f3 : ac8 20 d5 ed 2 1 .txd5 .te8 22 ll:lc5 .txc5 23 .txc5 .tf7, and the game heads towards a draw; Johansen-Hart, Martigny 1986.

12 . . . : dB

Another well-known continuation is 1 2 . . . lhc6. 13 .tb5 .tg4 14 : c7 lhc6 15 d5 : ab8! Black does badly with 1 5 . . . e6? 1 6 tLlgS! ed 1 7 llJxf7 : dc8 1 8

: xc8 + : xc8 19 ll:ld6 : c7 20 : c l ! d e 2 1 ll:lxb7 : d7+ 2 2 �e 1 : xb7 23 .txc6 :e7 24 .td5 + .te6 25 .tgS! .txd5 (25 . . . :e8 26 : c8 !) 26 .txe7 .txa2 27 : c8+ , and White's win is not in doubt; Ftacnik-Smejkal, TrenCianske Teplice 1 979.


16 �f4! In Mikh. Tseitlin--Schmidt, Lodz 1 980, White played the

inadequate 1 6 ..t.xc6 be 1 7 : xc6 e6 18 d6 .l: b2+ 19 �d3 : xa2 20 .l: hc l (or 20 h3 �xf3 2 1 gf �e5, and Black has no problems) 20 . . . e5! 2 1 l:tlc2 ..t.d7! 22 J:l. 6c5 J:l. xc2 23 �xc2 f6, with the more pleasant position for Black.

16 . . . e6 1 7 ltlg5 ed 18 ltlxf7 .I: deB!? 19 ed :Z. xc7 20 ..i.xc7 .l: c8 21 de .l: xc7 22 ..i.c4!? �f8 23 ltld8 be 24 ltle6+ �xe6

Again, the draw cannot be avoided. In practice, as a rule, Black nonetheless steers clear of this

ending, since White does after all dictate the play.

55 w

1 1 ..i.h6! A new, and quite effective, idea of Zaitsev's. By exchanging

bishops, White attempts to exploit the weakening of the dark squares in his opponent's camp. I proceeded less convincingly in the 1 3th game of the match, in which the diagrammed position arose by a different move-order (9 .l: c l 1Wa5 10 ltlf3 e6). Now is the best time to pause and examine that game.

l l d5 Practice has also seen 1 1 ..i.e2 cd 1 2 cd 1Wxd2+ 1 3 ltlxd2 ltlc6

14 d5?! ed 1 5 ed ltld4, with equality; or 1 1 �d3!? .l:t d8 1 2 ..t.g5 f6 13 ..t.e3 cd 14 cd 1Wxd2+ 15 �xd2 ltlc6 16 .l:t c4 ..t.d7 17 : b l ..i.f8 1 8 �e2 b6 1 9 ..i.d2 : ac8, with unclear play; BleesMikhalchishin, Budapest 1 990.

11 . . . ed 12 ed .l: e8!? A game Lputian-Tukmakov, USSR 1989, went 1 2 . . . ..t.f5 ( 1 2 ___ /

. . . ltld7 1 3 ..i.d3 b5! 1 4 c4 1Wxd2+ 1 5 ltlxd2 be 1 6 ..t.xc4 should have led to equality in Ftacnik-Hartston, Skara 1980 : 16 . . . ..t.b7


1 7 0-0 llle5. But after 1 3 c4!? 1fxd2+ 1 4 lllxd2 b6 1 5 .i.e2, White has a minimal edge) 1 3 .i.e2 llla6!? 14 0-0 ll. ad8 1 5 .i.xa6 1fxa6 16 .*.xeS ll. fe8 17 c4 ll. c8 18 ..id4 ll. xc4, with equality.

13 .i.e2 .i./5 14 0-0 llld7 15 h3 A useful prophylactic move. Instead, Piket-Korchnoi, Wijk aan

Zee 1 990, went 1 5 1fb2 lllf6 ( 1 5 . . . lllb6 is also sufficient for equality; Seirawan-Olafsson, Novi Sad 01 1 990, continued 16 1fb3 c4 1 7 ..ixc4 ll. xe3 1 8 fe 1fc5 19 .i.b5 a6 20 ..ie2 lllxd5 2 1 1fxb7 ll. e8 22 ..ixa6 .i.xc3 !-!) 1 6 .i.c4?! ( 16 1fxb7 i s more logical : 1 6 . . . 1fxa2 1 7 .i.b5 ll. ab8 1 8 1fa6 1fxa6 19 .i.xa6 lllxd5 20 ..ixc5 lllxc3 2 1 llld4, maintaining the balance) 16 . . . 1fb6! 1 7 1fa3 .i.f8 1 8 ll. cd 1 .i.d6 1 9 h 3 a 6 2 0 .i.h6 1fc7 2 1 1fc 1 1fd7 22 ..if4 ..ie4 23 lllg5 b5 24 ..ixd6 1fxd6 25 lllxe4 lllxe4 26 ..id3 c4 27 ..txe4 II. xe4, with advantage to Black in the queen and rook ending.

15 . . . lllb6 The prophylactic reply 15 • . . h5!? is also worth considering. In

reply, 16 c4?! is dubious : 1 6 . . . 1fxd2 17 lllxd2 ..id4! 1 8 .i.xd4 ll. xe2 19 ..ie3 ll. e8, with the highly unpleasant threat of 20 . . . .: 8xe3. Therefore White would have to content himself with exchanging bishops : 1 6 ..id3 ..ixd3 17 1fxd3, which allows 1 7 . . . 1fxa2 1 8 1fb5 lllf6 1 9 c4 b6 20 rHd1 with complex play.

After 15 • • . lllf6 16 c4 1fxd2 1 7 lllxd2 llle4 1 8 lllxe4 ll. xe4, White would still retain a microscopic plus.

16 g4 ..id7 17 c4 1fxd2 18 lllxd2 llla4 If 1 8 . . . ..id4, then 19 .i.xd4 II. xe2 20 .i.e3 f5 2 1 II. fe 1 II. xe 1 +

22 .: xe 1 fg 23 hg .i.xg4 24 ..ixc5, and White's game is somewhat preferable.

19 ..i/3 lll c3 (56) The more restrained 1 9 . . . b6 is also playable.

20 .J:l xc3!? .i.xc3 21 li:::Je4 � xe4! An opportune counter-stroke. After 2 1 . . . .i.g7 (2 1 . . . ..td4?

22 ..txd4 cd 23 li:::Jf6+ and wins) 22 li:::Jxc5, with a pawn for the exchange, White would retain a powerful initiative.

22 ..txe4 .J:l e8 23 J..d3 b6 24 �g2 /5!? 25 gf ..tx/5 26 ..txf5 gf 27 : dl

Holmov recommends 27 .J:l c l ..tf6 (27 . . . ..td4? 28 ..txd4 cd 29 .J:l d l .J:l e4 30 f3, and endgame difficulties arise for Black) 28 �f3 .J:l e4 29 .i.f4, and the bishop threatens to penetrate to Black's rear. Hence a more accurate reply is 27 . . . ..tg7 28 �f3 .J:le4, preventing 29 .i.f4? because of 29 . . . .J:l xf4+ ! . If 29 d6 (29 .J:l c2 ..te5) 29 . . . �f7 30 .J:l d l , then 30 . . . ..td4 is playable.

27 . . . <i;f7 Not 27 . . . .J:le4? 28 ..tg5! 28 .J:ld3 J..fl 6 29 l:!. a3 a5 30 .J:l b3 ..td8 31 .J:l c3 ..tc7 32 a4

<i;f6 33 �fl /4 34 J..cl �!5 35 .J:l c2 .J:lg8 36 .J:l e2 ..te5 37 ..tb2! ..td4! 38 ..txd4 cd 39 : e7! d3!

The immediate 39 . . . .J:l c8 is weaker: 40 �e2 .J:l xc4 41 �d3, and Black is faced with problems (4 1 . . . .J:lxa4? 42 d6).

The amusing 39 . . . f3 40 .J:l xh7 .J:l g l + 41 �xgl d3 fails to 42 .J:l f7+ .

40 </;elf .J:l c8! One more precise move that is indispensable to Black. 41 <i;d2 .J:l xc4 42 �xd3 -1--J. After the forced 42 . . . .J:l xa4 43 d6 : a t 44 �c2 (44 �c4 .J:l d l )

44 . . . .J:l a2+ 4 5 <i;c3 : a t , the draw i s unavoidable. We now return once again to the 1 5th match game.

57 B

1 1 li:::Jc6 12 h4! (57)


This is the whole point; White forces his opponent to go into an ending, as the threat of h4-h5 is quite unpleasant. A factor of considerable importance is that with the dark-squared bishops exchanged, the white king will find a comfortable central post on e3.

12 13 -*.xg7 14 cd

cd �xg7

14 h5 is hazardous : 14 . . . de 1 5 J:l xc3 l:l d8 16 .i.d3 �g8! , but not 1 6 . . . l£lb4 1 7 hg l£lxd3+ 1 8 l:l xd3 ( 1 8 �e2! is interesting) 1 8 . . . 'iFxd2+ 1 9 l:l xd2 l:l xd2 20 �xd2 hg 2 1 e5, with approximate equality.

58 w

14 15 �xd2 16 �e3

'iFxd2+ :t d8 -*.d7 (58)

The only way. Postponing development with 1 6 . . . h6 (aiming to meet 1 7 h5 with 1 7 . . . g5) could be costly to Black after 1 7 .i.b5 .i.d7 1 8 -*.xc6 .i.xc6 1 9 ll:le5.

17 :t b1 As the answer . . . b7-b6 would be bad for Black on account of

-*.a6, which gains control of the c-file, this transfer of the rook is very useful. Incidentally, 1 7 ..tb5 would fail to the unexpected 1 7 . . . l£lxd4! 1 8 l£lxd4 e5! with complete equality.

17 :t abS 18 .i.d3

It was worth considering an immediate 1 8 h5!?, since in the game Black could now have played 1 8 . . . h6 to frustrate this plan.


18 -t'i:Je7 Kasparov probably rejected 1 8 . . . h6 because of 19 g4 .i.e8

(nor can he rid himself of his troubles with 19 . . . f6 20 g5 hg 2 1 hg 22 l:t h6 .ll h8 23 l:t xh8 �xh8 24 d5, when White has the better prospects) 20 g5 h5 21 J: hc 1 , and relying on his powerful centre White can proceed with a long-term plan for pressure on both flanks.

19 h5 19 J: hc l , recommended by Krasenkov, is answered by 1 9 . . .

J:[ bc8 ! and Black has already solved most of his problems.

59 8

19 f6 20 hg hg 21 J: h2 (59)

Rather slow. White could have set his opponent more problems with 21 g4!?. The threat to break up Black's kingside structure is highly unpleasant. Replies which fail to help are 21 • . . g5 (2 1 . . . .i.c6 22 .i.c4!) 22 e5! t'i:Jd5 + 23 wd2 l:t h8 24 J:[ xh8! wxh8 25 .l:t h 1 + wg8 26 J: h6, and 21 . . . .ll h8 22 .l:txh8 wxh8 23 g5 f5 24 l'i:Je5 .i.e8 25 ..ta6 b6 26 J: c 1 or 26 f3.

It seems Black would have to play 21 • . . e5 22 de ..txg4 23 ef+ wxf6. White would then carry out a similar idea to one which will arise in the actual game (with the difference that the black pawn is on b6 and the white rook on h 1 ) : 24 t'i:Jd4 J: d7 25 : h4 : bd8 26 e5 + (26 : xg4 : xd4 27 e5 + wxe5 28 : xd4 J:[ xd4 29 f4+ J:[ xf4 30 J: b5 + t'i:Jd5+ 31 n xd5 + wxd5 32 wxf4 draws) 26 . . . wxe5 27 J: b5 + l'i:Jd5+ 28 l:t xd5+ J:[ xd5, but this plan loses its force because the c6 square has not been weakened. There is likewise no danger to Black in 24 e5+!? wg7 (24 . . . we6 25


llJg5+ !) 25 llJg5 llJd5 + (stronger than 25 . . . .I:Xd3 + 26 �xd3 .tf5 + 27 �d2! .txb l 28 : xb 1 ) 26 �e4 (26 �d2? llJf4 27 : b3 .tf5, and it is Black who wins) 26 . . . llJc3 + ! (after 26 . . . .tf5 + 2 7 ri>f3! .txd3 2 8 : h7 + �g8 29 : bh 1 , there i s n o defence against the mate threats) 27 �e3, and everything ends happily as in a good vaudeville : 27 . . . llJd5+ 28 �e4 llJc3+ etc.

21 b6 One can understand Black's wish to avoid the unattractive

endgame that results from 21 . . . : h8 22 : xh8 �xh8 23 .ta6! b6 (23 . . . .tc8 24 .tb5! or 24 .tc4) 24 :c 1 , or 23 . . . .tc6 24 .tc4 ..td7 (24 . . . f5 25 : h 1 + �g7 26 llJg5 ri>f6 27 �f4 fe 28 : h7 : f8 29 g4, threatening 30 llJxe4+ .txe4 3 1 g5 mate; the only defence is 29 . . . llJd5 +, but then Black loses his rook after 30 .txd5 ed 3 1 : c7!) 25 g4.

22 g4 Kasparov now plays very ingeniously :

60 w

22 e5!? 23 de .txg4 (60)

24 ef+ 24 : bh 1 has been recommended as stronger. But let us see

whether it is. 24 . . . : xd3 + ! (inadequate alternatives are 24 . . . .txf3 25 : h7 + ri>f8 26 : h8 + llJg8 27 .tc4! .txh 1 28 ef 'i>e8 29 .te6, and 24 . . . fe 25 llJxe5 .te6 26 : h7+ ri>f6 27 f4 : bc8 28 ttJf3! .tg8 29 ttJg5! .txh7 30 : xh7 : xd3 + 3 1 'i>xd3 : f8 32 �e3 llJc6 33 :c7 llJe7 34 llJh7+) 25 'i>xd3 .txf3 26 : h7 + �f8 (61) .

27 : 1 h3 (forced, since 27 ef llJg8! 28 : l h3 .txe4+! draws; nor does 28 : 1h4 help, in view of 28 . . . g5!, and when the rook moves,


61 w

Black again has 29 . . . -*.xe4+ !) 27 . . . -*.g4 (the attempt to weave a mating net for the white king, with 27 . . . -*.g2 28 J:l. g3 -*.fl + 29 �e3 fe 30 .IH3 + �g8 3 1 J:l. xe7 J:l. d8, fails to the simple 32 J:l. xe5) 28 J:l. g3 f5 29 �e3 J:l. e8 ! 30 J:l. h8+ ll'lg8, and Black is out of danger. It remains for us to consider 28 ef J:td8 + ! 29 �e3 (29 �c3 J:l. c8 + ! 30 �b2 -*.xh3 3 1 fe+ �e8 32 J:l. xh3 �xe7 33 J:t h7 + �e6 34 J:l. xa7 J:t f8, and Black has everything in order) 29 . . . ll'lg8 30 J:l. g3 ll'lxf6 3 1 J:t h8+ �e7 32 J:l. xd8 �xd8 35 e5 ll'ld5+ 34 �d4 -*.e6 35 J:l. xg6 �d7, and Black's defence cannot be breached.

So the move suggested by the annotators, 24 J:t bh 1 , fails to justify their hopes.

24 �xf6 25 ll'ld4

25 e5+ �g7 (25 . . . �e6?! 26 ll'lg5+) is not dangerous for Black.

62 w

25 .l:t b7?! (62)


This move was probably based on miscalculation in the main variation. But even after the more precise 25 • • • .l:l. h8 (25 . . . �g7 leads to an unattractive minor piece ending after 26 .l:l. bh l .l:l. h8 27 .l:l. xh8 .l:l. xh8 28 .l:l. xh8 c;;,xh8 29 c;;,f4) 26 l:t bh l .l:l. xh2 27 .l:l. xh2 c;;,g7, White can play, for example, 28 f3 (28 �f4 is premature in view of 28 . . . l:t d8 29 c;;,es g5! 30 lDe6+ ..txe6 3 1 c;;,xe6 .l:l. xd3 32 c;;,xe7 .l:l. d2!), and it is not easy for Black to save himself.

Has Black no other suitable defence, then? Zaitsev suggests 25 . • • l:t e8!?, with the idea of answering 26 .tb5 (26 c;;,f4 .1:1. bd8) with 26 . . . lDf5 + ! 27 lDxf5 (risky alternatives are 27 c;;,f4 g5 + ! 28 �xg4 .l:l. xe4+ , and 27 c;;,d3 .l:l. ed8! 28 .l:l. b4 .l:l. xd4+ ! 29 .l:l. xd4 .te2 + !) 27 . . . .l:l. xe4+ 28 c;;,xe4 .txf5 + . A complete demonstration of the geometric co-ordination of the black pieces.

26 f3 At this point, White could have put his opponent in a critical

position by playing 26 .l:l. h4!. How is Black to defend? The obvious-looking 26 • • . .l:l. bd7 (26 . . . .te6 loses to 27 e5 +)

suffers a fiasco after 27 e5+ ! �xeS (27 . . . �g5 28 .l:l. xg4+ ! �xg4 29 .l:l. g l + is crushing) 28 .l:l. b5 + lDd5+ 29 .l:l. xd5 + c;;,xd5 (Black comes out a piece down after 29 . . . .l:l. xd5 30 lDc6+ c;;,f6 3 1 lDxd8 �g5 32 .l:l. h8) 30 .l:l. xg4. It is notable that this line was indicated by Mephisto, the World Microcomputer Champion. True, Black can still struggle on : 30 . . . .l:l. e8 + !? 3 1 �f3 (after 3 1 ..te4+ �c4! , the black king slips across to the queenside, creating counterchances; while after 3 1 �d2 .l:l. f7 32 f3, Black is not devoid of hope although he clearly stands worse), and now 3 1 . . . .;;,cs 32 lDb3 + ! or 3 1 . . . �e5 32 lDc6+ allows White to re-group effectively, so Black must play 3 1 . . . .l:l. f7 + or 31 . . . .l:l. f8+; though in either case, after 32 �g3, the two pieces are much stronger than the rook and pawn.

Another try is 26 • • • lDfS+. If now 27 lDxfS gf 28 f4 (if 28 ef, then 28 . . . .l:l. xd3 + , while after 28 .l:l. g1 .l:l. bd7 29 .tb5 .l:l. e7 30 f3 c;;,g5 3 1 .l:l. hh 1 Black is guaranteed a draw with 3 1 . . . .l:l. xe4+ ! 32 fe f4+ 33 �f2 .l:l. d2 + etc.) 28 . . . fe 29 .txe4 ..tf5! , there follows a lengthy combination with a happy end : 30 .l:l. h6+ ! c;;,g7 3 1 ..txf5 �xh6 3 2 .l:l. h 1 + c;;,g7 33 .l:l. h7+ �f6 3 4 .l:l. xb7 �xf5 3 5 .l:l. xa7 .l:l. e8 + ! 3 6 �f3 .l:l. e 1 3 7 : f7+ c;;,e6 3 8 .l:l. b7 c;;,fs 3 9 .l:l. xb6 .l:l. fl + 40 �e3 .l:l.f3 + ! 4 1 c;;,d4 (capturing the rook leads to a classic stalemate) 4 1 . . . .l:l. a3 42 .l:l. b2 c;;,xf4 43 c;;,c5 : as 44 .l:l. b4+ �e5 45 a4 .l:l. c8+, and the draw is elementary.

Nonetheless there is a way for White to win. He should answer 26 . . . lDfS + with 27 ef l:l.e7+ 28 .i.e4! gf 29 l:l. xg4 etc.

26 l:l. bd7 If 26 . . . .i.d7 or 26 . . . .i.c8, then 27 f4 is strong.

27 l:l. b4 .te6 The best defence. 27 . . . .txf3? is refuted by 28 l:l.f2, and 27 . . .

lDc6 by 28 lDxc6 l:l. xd3+ 29 �f4 .thS 30 eS + ! . If 27 . . . l:l. e8, then 28 l:l. a4. The matter is more complicated after 27 . • • .th5 28 f4! lDc6 29 eS+ �g7 (29 . . . �f7 30 .i.c4+ is bad for Black) 30 lDe6+ �g8 3 1 lDxd8 lDxb4 32 .i.c4+ lDdS + 33 �e4 l:l. xd8 34 .txdS + <i;g7 3S l:l.c2 l:l.d7 36 e6 and 37 �eS. At move 30, if 30 . . . <i;h8, then 3 1 .i.xg6, or if 30 . . . <i;h7, then 3 1 l:l. xhS + . One final possibility i s 3 0 . . . �h6 3 1 l:l. b3 l:l. e8 3 2 .txg6! <i;xg6 33 <J;e4 lDxeS 34 l:l. g3 + lDg4 (34 . . . .tg4 3S fS+ <J;f6 36 l:l. h6+ <J;e7 37 </;xeS) 3S fS + �h6 36 l:l. xhS + <i;xhS 37 l:l. h3 mate.

28 l:l.c2 White could try playing for the win with 28 f4, but Black has

one satisfactory reply : 28 . . . .tg8! (if 28 . . . aS?, then 29 l:l. h7? is a mistake on account of 29 . . . lDfS + 30 ef ab, but 29 eS+ ! �f7 30 l:l. h7+ �g8 3 1 l:l. xe7! l:l. xe7 32 l:l. xb6, or 3 1 . . . ab 32 l:l. xe6 l:l. xd4 33 l:l. xg6+, gives White a clear plus) 29 eS+ �g7 30 .i.e4, and the chances are about level.

28 .tbS! looks a much stronger try. There follows 28 . . . l:l. c7 (63)

63 w

29 lDxc6 l:l.c3+ 30 �f4 �xe6 3 1 .tc4+ �d6 (3 1 . . . �f6 is dangerous : 32 l:l. h7 gS+ 33 �g4! and the black king comes under a strong attack) 32 l:l.d2+ �c7 (more precise than 32 . . . <i;cS 33 l:l. xd8 �xb4 34 .te6! gS+ 3S �xgS l:l. xf3 36 l:l.d7 lDc6 37 l:l. c7,


and Black's troubles continue) 33 l:lxd8 cRxd8 34 ..te6 g5+ (34 . . . .!Llc6 can be met by 35 J:lc4! J:l xc4 36 ..txc4 �e7 37 �g5, when White's chances are better) 35 cRxg5 J:l xf3 36 J:l d4+ �e8 37 J:l d7 (also 37 ..tg4 J:l g3 38 �f4 J:l a3 39 ..th5 + cRf8 only leads to equality) 37 . . . J:l g3 + (37 . . . .!Llc6 is unconvincing in view of 38 J:l c7 .!Lld4 39 ..td7+) 38 �f6 (an equal position results from 38 cRf4 J:lg6 39 ..tf5 .!Llxf5 40 ef �xd7 41 fg �e7 42 �e5 �f8! 43 �d5 �g7 44 cRc6 �xg6 45 a4 cRf6 46 cRb7 aS) 38 . . . J:l g6+ 39 cRe5 J:lg5 + (Black loses after 39 . . . .!Llc6+ 40 �d5 .!Llb4+ 41 cRd6 .!Lld3 42 J:l xa7, and has a difficult ending after 40 . . . .!Lle7+ 41 cRd6 .!Llc8 + 42 cRe5) 40 cRd6 .!Llc8 + 41 cRc7 J:l c5+ 42 �b7 J:l e5 43 ..th3 .!Lle7 44 cRxa7 J:l xe4 45 cRxb6 J:l e3, with a clear draw.

28 aS 29 J:l a4 gS! 30 ..tbS J:ld6 31 ..tel ..td7 32 J:l ac4 J:l e8 33 J:l b2 .!LidS+

t-t Black now has a tiny pull, but neither side can seriously count

on winning.

Game No. 1 5 Karpov-Kasparov

World Championshp (17th game) Lyon 1990

I have said that in my last match with Kasparov the Griinfeld Defence was played in four games. The first three of them, though containing fascinating struggles, ended in draws. We have covered all three together in the context of Game 14. In the seventeenth match game, which concluded our dispute in this opening, I finally succeeded in winning. That is why this example merits a separate number in the book.

1 d4 .!Llf6 2 c4 g6 3 .!Llc3 d5 4 cd .!Llxd5 S e4 .!Llxc3 6 be ..tg7 7 ..te3 c5 8 11rd2 0-0

1 9 �n ..t� !Up to here, everything coincides with the fifteenth match game.

From the notes to that game, the reader will derive a large amount of information about this opening line. In that game Kasparov


played 9 . . . "ifa5, and encountered distinct problems - as we saw. Consequently, this time he prefers a different method.

64 8

10 ll:lg5!? (64)

Perhaps White can hardly to without this thrust if he is to count on an advantage. But then I dare say the debate about this opening line will be continued further, and the assessment will undergo many alterations.

10 cd 11 cd ll:lc6

If 1 1 . . . h6, then 1 2 h3 is obligatory. 12 h3 ..td7

An inferior choice is 1 2 . . . ..txd4 1 3 ..txd4 ifxd4 14 "ifxd4 ll:lxd4 15 hg ll:lc3 + 16 ..t>d2 ll:lxa l 17 ..td3.

13 J:r b1 J:[c8! The subtle tactical point lies in the highly attractive, though

not complicated, line 14 J:r xb7 ll:lxd4! 1 5 ..txd4 ..txd4 16 "ifxd4 J:[c1 + 1 7 �d2 (or 1 7 �e2 ..tb5 + 1 8 ..t>e3 J:[ e 1 +) 1 7 . . . J:r d 1 + ! 1 8 ..t>xd 1 ..ta4+, and White loses his queen. Hence he has to spend some time bolstering his centre.

14 ll:lf3 ll:la5 15 ..td3 ..te6

The black pieces have securely fastened onto the c4-point. It is a rule in many Griinfeld variations that the domination of c4, in the absence of organic pawn weaknesses, guarantees Black counterplay.

16 0-0 17 J:[ fd1

..tc4 b5?! (65)

65 w


This might seem to be quite in order; Black strengthens his hold on c4. But in so doing, he commits a major positional error. From now on, his queenside pawn chain becomes vulnerable and causes him a great deal of worry. The modest 1 7 . . . b6! would have been more appropriate.

18 ..tg5 The first encouraging sign. The threat is not only 1 9 ..txc4

.!Dxc4 20 "it"b4 .!Dd6 2 1 e5, but also, in some variations ..txe7, deflecting the black queen.

18 a6 19 .Z: bc1

I believe 1 9 : de l !? was also worth considering. Then, for example, after 1 9 . . . ..txd3 20 : xeS 1Wxc8 2 1 "it"xd3 "it"b7, White would immediately have the highly effective break 22 a4, and if 22 . . . b4, then 23 ..td2, winning material.

19 ..txd3 20 .l:t xc8 "it"xc8 21 'ilfxd3 .l: e8?!

A passive move. 21 . . • "it"b7 22 "it"a3 .!Dc4 23 "it"xe7 1Wxe7 24 ..txe7 J:le8 was more flexible. It is possible that I would then have had to revert to thy plan of .l:l d 1-b 1 and a2-a4. After 22 .l: b1, the reply 22 . . . h6 (ls bad on account of 23 ..td2! . The alternative 22 .z:t cl is less convincing, as after 22 . . . h6 23 ..tf4 tllc4, neither of White's options promises him anything to speak of: 24 tlld2 .l:t c8 ! 25 d5 .!Db2, or 24 d5 J:l c8 25 .!Dd4 tllb2 26 l:t xc8 + 1Wxc8 27 W'd2 .!Dc4 28 .!Dc6 tllxd2 29 tllxe7 + �f8 30 .!Dxc8 .!Dxe4.

Another interesting variation is 21 . . . tllc4 22 J..xe7 .l: e8 (22 . . . .!Db2 23 W'b3 tllxd 1 24 J..xf8 1Wc1 25 .ta3 'ilfa1 26 "it"c2!) 23


..ta3 tt.Jxa3 24 'ilfxa3 l:txe4 25 d5 ..tf8 26 d6 'iii'd7 27 tt.Jg5, with complications.

22 l:tcl 'iii'b7 23 d5 tt.Jc4 24 tt.Jd2

A key factor in White's overall strategy. Evicting the last piece from c4, he seizes the vital file.

24 tt.Jxd2 It was worth considering 24 . . . h6 25 ..tf4 e5 (25 . . . g5 26 ..tg3

tt.Jxd2 27 l:l. c7 'iii' b6 28 'iii'xd2 favours White, as does 27 . . . l:l. c8 28 l:l. xb7 l:l.c l + 29 �h2 tt.Jfl + 20 'ilfxfl l:l. xfl 3 1 l:l. a7; another unsatisfactory line is 25 . . . tt.Jxd2 26 'ilfxd2 g5 27 .:t c7 'iii' b6 28 l:t c6 'ilfd4 29 W'xd4 ..txd4 30 ..td2 .:t a8 31 ..tb4 �f8 32 d6! ed 33 l:l. xd6) 26 ..te3 (avoiding the sly trap 26 de?! l:l. xe6 27 tt.Jxc4 'ilfxe4 ! with instant equalisation) 26 . . . tt.Jxe3 27 'ilfxe3 l:l. c8 28 tt.Jb3 ..tf8 29 tt.Ja5 l:l. xc l + 30 'ilfxc l 'iii'b6 31 tt.Jc6 f6, and Black can resist stubbornly.

66 w

25 ..txd2

26 l:l. c6!

: cs (66)

In view of the variation 26 . . . l:t xc6 27 de 'ilfxc6 28 W'd8 + ..tf8 29 ..th6, Black's position at once becomes critical. Of course, 27 . . . W'c7 28 'iii'd7 ..te5 29 ..th6! would also lead to a simple win for White : 29 . . . 'ilfxd7 30 cd ..tc7 3 1 e5! aS 32 �fl b4 33 �e2 a4 34 �d3 ..td8 35 �c4, etc.

26 ..te5 27 ..tc3 ..tb8


In the event of 27 . . . ..txc3 28 'ii'xc3 .l:!. xc6 29 de iilc7 30 g3 a5 3 1 'ii'c5 b4 32 iild5, or 30 . . . f6 31 'ii'b3 + �f8 32 'ii'e6, the white queen penetrates decisively to d7.

28 'ii'd4 f6 29 ..ta5

Even now White has to play carefully; not 29 .1:1. b6?? .l:!. xc3! 30 .l:!. xb7 .l:!. c1 + and mates.

29 ..td6 30 'ii'c3 .l:!. e8 31 a3

Why hurry? The fruit will ripen of its own accord. 31 �g7 32 g3 ..te5 33 'ii'c5 h5 34 ..tc7 .tal 35 ..tf4 'ii'd7 36 .l:!.c7 'ii'd8 37 d6!

The death agony of the black pieces, suffocating on the edge of the board, now commences.

37 38 39 40

d7 ..td2 .l:!. b7

l-0

g5 .l:t£8 ..te5

On 40 . . . h4, Zaitsev gives 41 ..ta5! 'ii'xa5 42 "it'xe7+ �g6 (42 . . . .l:!. f7 43 'ii'xf7+) 43 "it'h7 + ! �xh7 44 d8('ii')+ . A paradox - to win Black's queen, �ite must first sacrifice his own!

4 Russian System

Game No. 1 6 Ehlvest-Ernst Tallinn 1989

1 d4 ltlf6 2 ll'lf3 g6 3 c4 J..g7 4 ll'lc3 d5 5 iib3 de 6 "ifxc4 0-0 7 e4 J..g4

The queen move to b3 (or the rarely played "ifa4+) constitutes what numerous theoretical manuals call the Russian System. But it incorporates various subsidiary systems which depend on Black's seventh move. At present, the Ragozin (or Prins) Variation, 7 . . . ltla6, enjoys the greatest popularity. This is evidently due to its appearances in the 1986 and 1987 World Championship matches, and the subsequent rapid development of its theory. The Smyslov Variation, 7 . . . J..g4, also occurred in two of the KarpovKasparov match games - I shall come to these presently. Among the rarer and somewhat passive continuations, I would menion 7 . . . c6 (The Boleslavsky Variation), and 7 . . . a6 with the idea of . . . b7-b5 (the Hungarian Variation).

8 J..e3 The most flexible move. The continuations 8 ll'lg5 and 8 ltle5

cause Black no harm. 8 ltlfd7 (67)

The critical position of the Smyslov Variation. White's most popular moves are 9 J: d 1 , 9 J..e2, 9 0-0-0, and the one he plays in this game.

9 1Wb3

67 w

Russian System 109

The most active continuation. Instead of 9 � b3 White also has an interesting idea of placing his queen on c5: 9 l:t dl il'lc6 10 .tel il'lb6, and now 1 1 'ii'c5 'ii'd6 (68) .

68 w

A large number of moves have been seen here : 12 "ifxd6, 12 il'lb5, 1 2 d5 , 12 h3 , 1 2 0-0. I would remind you of the famous game Botvinnik-Fischer, Varna 1 962, which went 12 h3 ..txf3 1 3 gf .ll fd8 14 d 5 il'le5 1 5 il'lb5 'ii'f6 1 6 f4 il'led7 1 7 e5, and now Fischer gave his opponent an unpleasant shock with 1 7 . . . "ifxf4 ! . After some errors . by both sides, however, this fascinating encounter was eventually drawn.

But White proves to have one other interesting move at his disposal. Without having been tried out before, it occurred twice in the return match for the World Championship in 1986, and for that reason I have decided to discuss it :

12 e5 A paradoxical decision at first sight. White not only leaves his

pawn on e5 hopelessly weak, he allows a queen exchange too. But it is not as simple as all that.

110 Russian System

12 . . . "it'xe5 13 de 'DeB The more natural-looking 1 3 . . . 'Dd7 is met by 14 h3! ..txf3

1 5 gf ! , and the 'doomed' pawn on e5 is immune - since if either knight takes it, 16 f4 ! wins a piece. Otherwise White will protect the e-pawn with its neighbour, gaining a clear advantage.

14 h3! This was played in the seventeenth match game (Karpov

Kasparov, Leningrad 1986). In the fifteenth game, where the novelty 12 e5 was used for the first time, I continued instead with 14 'Db5, and after 14 . . . l:t b8! 1 5 'Dxc7 e6! Black managed to obtain fully equal chances. The point is that the threat to surround the knight with 16 . . . a6 compels White to lose a tempo with 1 6 'Db5, whereupon the black knight on c8 quickly transfers itself to a convenient post. The game concluded : 16 'Db5 ttJ8e7 17 J:.d2 b6 1 8 cb ab 1 9 ..tg5 'Df5 20 b3 h6 2 1 ..tf6 ..txf3 22 ..txf3 'Dxe5 (only now does Black recover the pawn) 23 ..txe5 ..txe5 24 0-0 J:!. fd8 25 .!:tfd 1 J:. xd2 26 J:. xd2 :t c8 27 g3 :t e l + 28 <;t;>g2 <;t;>f8 29 ..te4 <;t;>e7 -!---!-.

We now return to the seventeenth game : 14 . . . ..txj3 If 14 . . . ..te6, the sortie 1 5 'Dg5 is unpleasant : 1 5 . . . 'Dxe5 1 6

'Dxe6 fe 1 7 f4. 15 ..txf3 ..txe5 After 1 5 . . . 'Dxe5 16 ..txb7 .:. b8 1 7 c6 'Dc4 1 8 J: d7 'Dxb2,

White can choose between two attractive possibilities : 19 'Dd5 and 19 'Db5.

16 ..txe6! be 17 ..td4 .i.f4 18 0-0 It is worth considering 18 'De2. A correspondence game Riim

mele-Brummer ( 1 988) continued 18 . . . e5 19 ..tc3 f6 20 l:!. d7 J:. f7 2 1 l:t d8+ <;t;>g7 (White also has a clear plus after 2 1 . . . J:. f8 22 J:. xf8 + <;t;>xf8 23 'Dxf4 ef 24 ..txf6) 22 'Dxf4 ef 23 <;i;>e2, with a won position. In Rummel's view Black can play more precisely with 18 . . . ..th6 19 ..tc3 l:. b8 20 J:t d7 ..tg7 21 ..txg7 (2 1 'Dd4 ! i s interesting) 2 1 . . . <;t;>xg7 22 b3 : b5 23 0-0 .:. xc5 24 .:. c 1 .:. xc 1 + 25 'Dxc1 'Dd6 26 .l hc7, with a minimal edge for White.

18 . . . a5? The correct line is 1 8 . . . e5 19 ..te3 ..txe3 20 fe 'De7, as

demonstrated in the game Karpov-Timman, Tilburg 1986, which continued :21 .l:d7 'Df5 22 J:. xc7 J:. fc8! 23 l:td7 l:td8 24 l:t fd 1 J:. xd7 25 : xd7 'Dxe3 26 l:t c7 .:. b8 27 b3 : d8! 28 'De4 .:. d4 29 'Df6+

Russian System 111

rt;g7 30 J:l. xc6 J:l. d2 3 1 g4 lDc2 32 �fl lDd4 33 J:l. a6 lDf3 !-} . 1 9 'IJ.fe1 a4 20 'IJ. e4 J.h6 2 1 J.e5 a3 22 b3 lDa7 23 J:l. d7 J.cl

24 J:t xc7 J.b2 25 lDa4! The material balance is restored, but White's threats are

mounting up. 25 . . . lDb5 26 J:l. xc6 J:l.fd8 27 'IJ. b6! J:l. d5 Cunningly thought out; if 28 lDxb2, then 28 . . . J:l. xe5! 29 J:l. xe5

ab 30 J:l. e 1 lDc3, and the b-pawn unexpectedly brings victory to Black. But avoiding the trap is quite simple.

28 J.g3 lDc3 29 lDxc3 J.xc3 30 c6 J.d4 31 J:l. b7 1-0 Black resigned in view of the inevitable c6-c7.

9 lDb6 Alternatives are 9 . . . c5 and 9 . . . J.xf3 .

10 J:l. d1 lDc6 Practice has also seen 10 . . . e6, and 10 . . . e5 1 1 de lD8d7 with

sharp play. With 10 . . . lDc6, Black provokes the advance d4-d5. 1 1 dS lDeS 12 J.el lDxf3+ 13 gf J.hS

As practice has shown, 1 3 . . . J.h3 is less convincing in view of 14 J:l. g 1 ! , and Black's bishop will soon have to retreat.

14 a4 14 lDb5 and 14 f4 are not dangerous for Black. We shall say a

word about 14 .l: g l presently. 14 'ii'd7 (69)

In such situations Black usually plays . . . 'ifc8 to defend his bpawn, but here this is not necessary, since after 1 5 aS lDc8 1 6 'ifxb7 lDd6 1 7 'ifc6 'ifh3! Black obtains good counterplay.

69 w

1 12 Russian System

15 J:. gl ! A n interesting idea. I t used to be thought that White should

not allow the queen onto h3. Hence the rook manoeuvre to g 1 was carried out a move earlier; then after 14 J:.g 1 'it'd7 (or 14 . . . 'it'c8), White could reply 1 5 J:. g3. In the present game, on the contrary, White 'tempts' the black queen to h3 and then even further - to h2. The result is mind-bending complications, which in fact were what attracted me to this game.

Formerly, the normal move from diagram 69 was 1 5 h4, for example : 1 5 . . . a5 1 6 lt:lb5 lt:lc8 1 7 .td4 (Uhlmann-Kozma, Zinnowitz 1 967, went 1 7 f4 .txe2 1 8 �xe2, and now 18 . . . lt:ld6! would have equalised) 17 . . . .txd4! 18 lt:lxd4 lt:lb6 19 .tb5 'it'd6, and the chances are level. I should mention that the exchange on d4 was suggested by Euwe. Long ago, in a game SmyslovBotvinnik from the return World Championship Match of 1958, Black played instead 17 . . . lt:ld6, and White gained a distinct advantage.

15 Black could also go after a different pawn - the one on a4 -

but after 1 5 . . . .txc3 + 1 6 be 'it'xa4 1 7 'it'xa4 lt:lxa4 1 8 �d2, White has quite enough initiative for it.

16 f4 'it'xh2 17 �d2!

White's intention becomes clear : having lured the enemy queen into his camp, he aims to exploit its awkward position in order to work up a kingside attack.

17 .txe2 The sharp variation 1 7 . . . .txc3+ 1 8 be .txe2 19 �xe2 'it'h5+

20 �d3 J:. ab8 2 1 : b 1 lt:ld7 22 'it'dl i s also possible. l8 lt:lxe2 c6

Black has to play energetically. If 18 . . . 'it'h3, then 19 f5 with unpleasant threats.

19 a5 lt:ld7 20 lt:lg3

White's plan of surrounding the black queen is quite consistent. After 20 'it'xb7 lt:lf6! Black would have a promising position.

20 cd Black sacrifices a piece, but obtains a large number of pawns

in return. 21 J:!. h1 'it'g2

22 ..te2! de

Russian System 113

The only move, in view of the threatened 23 J:t dgl . 23 J:t xd7 e5! 24 f5! 'iff3+

Ehlvest recommends 24 . . . gf at once. Play could continue : 25 .i..c5 (it must have been to prevent this move that Black gave check) 25 . . . f4 26 .i..xf8 J:t xf8 27 J:tfl (27 J:t hd l e3) 27 . . . .i..h6! (of course not 27 . . . fg 28 "ilfxg3, with a won ending for White) 28 J:t h l .i..g7, and White cannot improve the placing of his pieces.

70 w

25 ..tel gf (70)

26 J:tgl ! Once again the rook occupies the g-file, this time with even

greater effect. 26 f4

If 26 . . . J:t ad8 (26 . . . J:t ac8 27 "ilidl ! ), then 27 Wd5 ll xd7 28 "ifxd7 f4 29 lZ:lf5 fe 30 fe is decisive.

27 lZ:lf5 28 'ifxe3! 29 "ifh6 30 ll xg6 31 'ifh4!

fe 'ifxf5 'ifg6 hg

Now there is nominal material equality, but Black's pieces are very passively placed and he doesn't succeed in co-ordinating them.

31 b6 32 a6! e3 33 fe e4 34 b3 b5

1 14 Russian System

34 . . . -*.c3 + 35 �e2 -*.b4 doesn't help : 36 1Wxe4 -*.c5 37 1Wd5!

35 ""el -*.c3 36 1We7 -*.a5 37 .l:l. xa7 .l:l. ae8 38 1Wa3 -*.c3 39 1Wc5 b4 40 .l:l. e7 .l:l. c8 41 1Wd5 1-0

71 w

Game No. 1 7 M. Gurevich-Kasparov

Moscow 1988 1 d4 lt.:lf6 2 c4 g6 3 lt.:lc3 d5 4 lt.:lf3 -*.g7 5 1Wb3 de 6 1Wxc4 0-0 7 e4 lt.:la6 (71)

Introduced into practice by Ragozin as long ago as the 1930s, this variation has occurred three times in the encounters between Kasparov and myself. The nineteenth game of the return match in 1 986 ended in a beautiful win for White, but in Seville (games 1 5 and 2 1 ) Kasparov succeeded in obtaining equality, and both games were drawn. I do not know if its occurrence in our matches was the reason, but since then the popularity of this variation has noticeably increased, and a good many games have been played enriching its theory. In the notes to the present game we shall examine the most interesting practical examples, including all three of the World Championship games to which I have referred.

In the diagram position, Gurevich developed his bishop on e2. Strange though it may seem, no fewer than ten other continuations

Russian System 1 15

have also been seen! The moves 8 'ifa4, 8 b4, 8 ..te3, 8 h3, 8 e5, 8 l: b 1 and 8 ..td3 have completely gone out of use. 8 'it'b3, 8 ..tg5 and 8 ..tf4 are also employed only rarely, and in any case they often amount merely to a transposition - the bishop is brought to e2 a little later. I shall now given one example of each of these three alternatives that I have said to be the most common.

G. Georgadze-Tukmakov, Odessa 1 989, went 8 'it'b3 c5 9 d5 'it'b6! (9 . . . e6 is the usual continuation. The queen exchange gives Black a comfortable game) 10 'ifxb6 (perhaps White shouldn't be in a hurry to exchange; 10 ..tc4 or 10 e5 should be tested) 10 . . . ab 1 1 ..tc4 e6 12 de ..txe6 1 3 ..txe6 fe 14 0-0 lt:lb4 1 5 h3 lt:ld7 16 l: d l lt:lc2 17 l: b l lt:ld4 18 lt:le1 (better 18 lt:lb5 e5 19 a3, with unclear play. Now Black seizes the initiative and quickly achieves a won position) 1 8 . . . lt:le5 19 a4 : ad8 20 ..tg5 l: d7 21 ..th4 h6 22 b3 g5 23 ..tg3 lt:lec6! 24 �f1 l: fd8 25 lt:le2 lt:lxe2 26 �xe2 lt:ld4+ 27 �fl b5! 28 ab lt:lxb5 29 : xd7 : xd7 30 e5 lt:lc3 3 1 :b2 b5! 32 : c2 b4 33 : c l l: d2 34 : at c4 35 lt:lf3 l: d8 0-1 .

Flear-Ftacnik, Belgrade 1989, went 8 ..tgS h6 9 ..th4 c5 10 d5 b5 ! 1 1 lt:lxb5 (ECO gives 1 1 'ifxb5 : b8 1 2 'ife2 l:l xb2! 1 3 11Vxb2 lt:lxe4 14 :c 1 'ifa5, with advantage; if now 15 ..txe7 :e8 16 d6 lt:lxd6 17 'it'd2 : xe7 + 1 8 ..te2, Black has the very strong 1 8 . . . ..txc3 19 'ifxc3 'it'b5! 20 lt:lg1 lt:lb4 2 1 �fl :xe2! 2 1 lt:lxe2 ..ta6, and although White is two exchanges up, Black's attack is irresistible) 1 1 . . . 'ifa5+ 12 lt:ld2 l: b8 (theory recommends 12 . . . lt:lxe4, but the rook move is more precise) 1 3 0-0-0 (after the familiar 13 ..tg3 lt:lxe4 14 'ifxe4 : xb5 1 5 ..txb5 'ifxb5 16 'ife2 'ifxb2 1 7 0-0 lt:lb4, Black has excellent compensation for the minimal material deficit. But in castling queenside, the English grandmaster comes under a strong attack) 1 3 . . . lt:lg4! 14 1Wb3 (14 ..tg3 ..td7!) 1 4 . . . c4 15 ..txc4 ..td7 16 a4 : res 17 �b1 lt:lc5 1 8 11Va3 1Wxa4 ! 1 9 b3 ..txb5! 0-1 . After 20 ba ..txc4+ 2 1 �c 1 ..tb2+ 22 1Wxb2 lt:ld3 +, White suffers big material losses.

Eingorn-H. Olafsson, Reykjavik 1990, saw 8 ..tf4 c5 9 de 11Va5 (9 . . . ..te6 is also frequently played) 10 e5 lt:ld7 1 1 a3 1Wxc5 (72) . This position was recently the object of a minor theoretical debate, associated with three games.

12 1We4 (in Eingorn-Gavrikov, Tallinn 1989, White played 1 2 lt:ld5, and after the inferior reply 1 2 . . . : e8 he quickly reached a won position : 1 3 l: d 1 h6 14 h3 g5 1 5 1Wxc5! lt:ldxc5 1 6 ..te3 etc. Soon afterwards, in Piket-Ivanchuk, Tilburg 1 989, Black

1 16 Russian System

72 w

innovated with 1 2 . . . ttlb6 and seized the initiative : 1 3 1i'xc5 ttlxc5 14 ttlxe7+ �h8 15 J:t c l ttle6! 16 .te3 .td7 17 .te2 : ae8! 1 8 .i.xb6 a b 1 9 ttld5 .i.c6 20 l: d 1 f6! with clearly the better position. This time, it is Eingorn who himself changes the complexion of the game) 12 . . . ttlb6 13 .i.e3 1i'c7 14 :c 1 .tf5 1 5 1i'h4 : ac8 1 6 .i.xa6 ba 1 7 0-0 1i'b1 1 8 ttld4 ttlc4 19 .th6 .txh6 20 1i'xh6 ttlxeS 21 l: fe1 f6 22 b4 .l:l. fd8 23 ttlxf5 gf 24 l: xe5 .l:l. xc3 25 .l:l. c5 (but not 25 .l:l. xf5 1i'e4 26 .l:l. c5 1i'c2!) 25 . . . .l:l. xa3 (the chances here are level) 26 .l:l. xf5 .l:l. ad3 27 .l:l. g5 + fg 28 "it'xg5+ �f8 29 11t'f5+ �e8 30 "it'h5+ �f8 31 11t'f5+ -!-! .

8 .tel c5 9 d5

9 e5 is met by 9 . . . ttlg4 1 0 h3 cd 1 1 hg de 1 2 be "it' aS, with a good game for Black.

White gains nothing from 9 de .te6 10 1i'b5 l: c8! 1 1 1i'xb7 ttlxc5 12 1i'xa7 l: c7 13 1i'a3 (if 1 3 1i'a5 ttlcxe4, White can't play 14 ttlxe4 because of 14 . . . : xc1 + !) 1 3 . . . ttld3 + 14 .txd3 1i'xd3, with a very strong initiative for Black. This whole variation is given in the opening monographs. In M. Gurevich-Gavrikov, Moscow 1990, White varied with 1 1 c6 l: xc6, and then 12 1i'xb7, but after 1 2 . . . l: xc3! 13 1i'xa6 ttlxe4 14 0-0 l: c7 1 5 .tf4 ttld6! 1 6 .te5 1i'b8 17 .txg7 �xg7 18 b3 1i'b4 19 l: ad 1 ttlf5 20 l: fe 1 l: b8 2 1 .tfl l: b6, the game was level.

9 e6 10 0-0

The bishop' sortie 10 .tg5 (73) should also be mentioned. After 10 • • . ed (stronger than 10 . . . h6 1 1 .txf6 1i'xf6 1 2 e5

1i'd8 13 d6!, or 10 . . . 1i'b6 1 1 .txf6 .txf6 12 e5 .tg7 1 3 0-0,

73 B

Russian System 1 17

when practice has shown White's chances to be better) 1 1 ltlxd5, Black has two options : 1 1 . . . ..te6 and 1 1 . . . '1Va5 + .

Belyavsky-Kasparov, Belfort 1988, went 1 1 . . . ..te6 1 2 0-0-0 ..txd5 1 3 .:t xd5 '1Vb6 14 ..txf6 '1Vxf6 1 5 e5 '1Vf5 ! (in FlearKorchnoi, Lugano 1 988, after 1 5 . . . '1Ve7 1 6 .:t hd 1 : adS 1 7 : xd8 .l:l. xd8 18 .:t xd8+ �xd8 19 e6 White obtained the better chances, although his initiative was gradually neutralised) 1 6 ..td3 11t'c8! 17 .:t d 1 (allowing his opponent to work up a dangerous attack. The right move was 1 7 .:t d6, with complex play) 17 . . . b5! 18 '1Vh4 ltlb4 19 ..txg6 fg 20 .:t d7 '1Ve8! 21 : e7 ..th6+ ! 22 b1 :Z. d8 23 l:. d6 '1Vc6! 24 a3 .:t xd6 25 ed '1Vxd6 26 ab cb 27 '1Ve4 b3 0-1 .

Farago-Dorfman, Budapest 1 988, continued instead 1 1 . . .

11t'a5+ 12 ..td2 "it'd8 1 3 ltlxf6+ ..txf6 14 e5 ..tg7 1 5 ..tc3 (Gutman recommends 15 0-0 ..te6 16 '1Vc1 ! '1Vb6 1 7 ..th6, with slightly the better game) 1 5 . . . ..te6 16 11Vb5 '1Ve7 1 7 0-0 ltlc7 1 8 '1Va4 ..td7 19 "it'c2 ..tc6, with equal chances.

10 eel 1 1 ed : es (74)

The move 1 1 . . . ..tf5 might well transpose after 1 2 ..tf4 J:. e8; alternative ideas after 1 1 . . . ..tf5 are discussed in Games No. 1 8 and 1 9.

After 1 1 . . . b5 12 '1Vxb5! .:t b8 1 3 '1Va4 .:t b4 14 '1Vd 1 ! Black has no compensation for the pawn.

1 1 • • . b6 is interesting -- the knight on a6 is at once freed from its unwanted duty of defending the c-pawn. TukmakovChiburdanidze, Biel 1988, went 12 .:td1 ( 1 2 ..tg5 h6!) 1 2 . . . ltlb4 1 3 a3 ( 1 3 '1Vb3 ..tf5! ) 1 3 . . . ..ta6 14 '1Vb3 ..txe2 1 5 ltlxe2 ltlbxd5 16 ltlf4 c4 17 11t'xc4 '1Vc7 1 8 '1Vxc7 ltlxc7, and the position was

118 Russian System

completely equal. In Naumkin-Henkin, Moscow 1 989, White instead played 1 2 ll:le5, but again achieved nothing : 1 2 . . . ll:lb4 ! 1 3 ll:lc6 ll:lxc6 14 de .t.e6 1 5 1Wa4 1Wd4 1 6 1Wxd4 cd 1 7 ll:lb5 ll:ld5! , and after some sharp play the endgame concluded peacefully.

In Vladimirov-Fette, Groningen 1 989, White answered 1 1 . . . b6 with the valuable novelty 12 1Wh4!. After just three moves, Black was in trouble : 1 2 . . . ll:lb4 1 3 .t.g5 .t.f5 ( 1 3 . . . h6 14 .t.xh6 .t.xh6 1 5 1Wxh6 ll:lbxd5 16 l:t ad 1 1We7 17 l:t xd5! is also bad for Black) 1 4 l:t ad l . Vladimirov gives another variation which likewise favours White : 1 2 . . . ll:lxd5 1 3 ll:lxd5 (or 1 3 .t.g5!) 1 3 . . . 1Wxd5 14 ll:lg5 h6 15 .t.£3 hg 16 .t.xg5.

74 w

12 .t.f4 The fashionable development of the bishop on e3 (though with

Black having played 1 1 . . . .t.f5 instead of 1 1 . . . .l:l. e8) is covered in Game No. 1 9.

After 12 l:tdl (to prepare a quick advance of the d-pawn) 12 • . .

.t.£5, we reach a position that occurred twice in the Seville match (75) . This is the appropriate moment to recall those games.

75 w

Russian System 119

The 1 5th game of the 1987 match continued : 13 d6 h6 Not 1 3 . . . llJe4? 14 d7! 14 h3 In the 2 1 st match game, I played 14 .i.f4 (see below). A

suggestion of Gutman's is also interesting : 14 a3! llJd7 ( 14 . . . llJe4 1 5 .i.e3!) 1 5 1Wa2 ! llJb6 1 6 .i.e3! .i.e6 1 7 1Wbl . The queen has escaped from pursuit, and the d-pawn continues to fetter the opponent's pieces.

14 . . . llJb4 15 .i./4 llJd7 16 : d2 a6 1 7 1Wb3 b5 18 1Wd1 c4 19 a4! llJc5 20 ab llJbd3 21 .i.xd3 llJxd3 22 : xd3! cd

The bishop on f5 has to be left where it is, to restrain the d-pawn; a weaker choice is 22 . . . .i.xd3 23 : xa6 : xa6 24 ba 1Wa5 25 1Wa4! 1Wxa4 26 llJxa4.

23 llJd5! ab! 24 llJe7+ <i;h7 25 : xa8 1Wxa8 26 llJxf5 gf 27 1Wxd3 'ile4 28 1Wxb5 : a8 29 .i.d2 :dB 30 1Wc5 1We6 31 .i./4 .i.xb2 32 llJh4 .i.f6 33 1W xf5+ 1Wxf5 34 llJxf5 h5! 35 g4 hg 36 hg �g6 37 �g2 .i.b2 38 llJe7+ <i;f6 39 llJc6 : d7 40 llJb8 : dB 41 d7 <i;e6 42 <i;f3 .i.a3 :f--J.

From diagram 75, the 2 1 st game of the Seville match went : 13 d6 h6 14 .i./4 llJd7! More accurate than 14 . . . llJhS 1 5 .i.e3, when the white bishop

has enticed the enemy knight onto the edge of the board and itself settled in a good position.

15 : d2 Or 1 5 1Wb3 llJb4 1 6 :d2, transposing; alternatively 1 6 .i.c4

1Wf6! 1 7 :d2. 15 . . . llJb4 16 1Wb3 .i.e6 Better than 16 . . . a6 1 7 a3 llJc6 1 8 llJdS! . 17 .i.c4 llJb6 18 .i.xe6 :xe6 (76)

76 w

120 Russian System

19 a3? An unfortunate move. 19 tLlb5 is also bad : 19 . . . .l:. e4 ! 20 ..te3

tLlc4 and Black has an excellent position. Instead, 19 ..tg3 tLld3 20 tLlb5 (or 20 tLld5 c4 21 "it"b5 tLlxd5 22 "it"xd5 tLlxb2 23 Wxb7 c3 24 J:l. c2) 20 . . . c4 2 1 "it"a3 leads to double-edged play. In a simultaneous display with clocks, against six of the strongest American juniors (New York 1 988), Kasparov reached the diagram position again in his game against Rao. But in answer to 1 9 ..tg3, he didn't jump to d3 with his knight; he chose the modest 1 9 . . . Wd7, and after 20 a3 tLlc6 2 1 Wb5?! .l:. c8 ! 22 .l:. ad1? (22 tLld5 i s stronger) 22 . . . ..txc3 23 be tLle5! 24 Wxd7 tLlxf3 + 25 gf tLlxd7, Black achieved a won endgame.

19 . . . tLld3! The appearance of the knight at d3 causes a certain amount of

confusion in the white camp. Capturing it, obviously, loses to the fork 20 . . . c4 !

20 ..tg3 c4 21 Wc2 .l:. c8 22 .l:. ad1 Wd7 Not 22 . . . tLlxb2, which fails to 23 Wxb2 tLla4 24 tLlxa4 ! ..txb2

25 J:l. xb2 Wa5 26 .l:. b4. 23 h4 f5 Perhaps 23 . . . .l:. c6 or 23 . . . l:t c5 was more accurate. The

exchange sacrifice on d3 would then be ineffective; White would have to play 24 "it"b1 and then 25 tLlel .

24 .l:. xd3 cd 25 "it"xd3 tLlc4 26 "it"d5 tLlb6 Evidently, Kasparov's weakened king position is not to his

liking, and he decides not to risk anything. And yet with 26 . . . <it>h7! Black would have retained the better chances. You may ask what difference it makes whether the king goes to h8 or h7. In fact the difference is highly significant. After 26 . . . c;t>h8 (26 . . . tLlxb2 27 .l:. e 1 .l:. ce8 28 .l:. xe6 .l:. xe6 - 28 . . . "ifxe6? 29 d7! - 29 tLlb5 <it>h7 30 tLle5! ..txe5 3 1 ..txe5 tLlc4 32 f4 ! a6 33 tLld4, with advantage) 27 tLlb5 tLlxb2 28 .l:. b 1 tLla4 29 tLlc7 .l:. xc7 30 de "it"xd5 31 c8(W)+ c;t>h7 32 .l:. xb7, it is White who has the winning chances. With the king on h7, the c-pawn would queen without check and the whole variation would fail.

27 Wd3 More precise than 27 Wb3 Wf7! 28 tLld5 (28 <it>h2 g4) 28 . . .

J:l. d8, with the better chances for Black. 27 . . . tLlc4 28 "if d5 tLlb6 i-f.

Russian System 121

12 .i.f5 Again 12 . . . b6 is possible. In Annageldiev-Arbakov, Uzhgorod

1988, there followed 13 d6 ( 1 3 a3 li:Jc7 14 d6 li:Je6 gives White nothing, but 1 3 .:. fd 1 li:Jb4 14 Wb3 .i.f5 15 .:. ac1 is worth trying) 13 • • • li:Jb4 14 li:Jg5 Wd7 15 'ifb3 (77) .

77 8

The position looks unpleasant for Black ; after 15 • • • .i.a6 1 6 .i.xa6 li:Jxa6, his knight has been thrown back again. But the Moscow master found a powerful resource: 15 • • • .:. xe2! (Black gives up the exchange but activates his bishops) 1 6 li:Jxe2 .i.a6 1 7 li:Jc3 (White takes the correct decision t o return the exchange. 1 7 .:. fe 1 i s risky : 1 7 . . . li:J d 3 1 8 .:. ed 1 h6! 19 li:Jf3 li:Jh5! with a fine game for Black, for instance 20 Wd5 .:. e8! 21 .:. xd3 .:. xe2 22 li:Je5 .i.xe5 23 .i.xe5 .:. xe5! 24 'ifxe5 .i.xd3; but not 17 . . . li:Jh5 1 8 li:Jg3! li:Jxf4 19 .:. e7 c4 20 iff3, and Black comes to grief) 1 7 . . . .i.xfl 1 8 .:. xfl .:. e8 19 h3 h6 20 li:Jf3 iff5. A complex position has arisen, with chances for both sides.

13 .:. ad1 li:Je4 1 3 . . . iVb6 has been played; White continues 14 ifb5 or 14

li:Jh4, with mutual chances. 1 3 . . . li:Jd7 appears to be weaker. Ivanchuk-Kotronias, Lvov

1988, continued 14 Wb3 li:Jb4 1 5 .:. d2 li:Jb6 16 .i.b5 .i.d7 1 7 .i.g5 ifc8 18 .:.c 1 a5 19 li:Ja4 c4 20 'it'd l li:Jxa4 2 1 'it'xa4 .i.xb5 22 'it'xb5 'it'f5 23 h4 c3 24 be .:. ec8 25 c4, and Black came away emptyhanded.

14 ..td3 Before proceeding further with the main game, we must recall

the 19th game of the return World Championship match ( 1 986), in which I employed a prepared variation here.

14 li:Jb5! (78)

122 Russian System

78 B

The basic strategic conflict revolves round the d-pawn. If White succeeds in utilising its potential energy, the initiative will be his. If Black manages to blockade it securely, his position will be the more promising. In advancing my knight to b5 I was prepared to part with the b-pawn, aiming to develop an attack in the centre after 1 4 . . . J.xb2 1 5 d6 J.f6 1 6 J.d3! and 1 7 ll fe l . Kasparov refuses the gift.

14 . . . fll/6!? Taimanov recommended 1 4 . . . g5, but after 1 5 J.c1 g4 1 6 �e1

White has a clear positional advantage. 15 J.d3 Considering what happened later in the game, many people

gained the impression that Black's opening set-up had been refuted outright. But in chess such things rarely happen, and soon after the match Kasparov himself was to go back to this variation. In fact, in Belyavsky-Kasparov, Moscow 1 987, it was White who first departed from precedent by advancing his d-pawn at once. The ensuing fierce struggle brought about a rapid depletion of forces : 1 5 d6 J.d7 1 6 g3 g5 1 7 J.e3 h6 18 �c7 �xc7 1 9 de J.c6 20 �d2 ll ac8 2 1 �xe4 ll xe4 22 f/Jxc5 ll xc7 23 ll d6 b6 24 ll xf6 be 25 ll xc6 ll xc6 26 J.f3 ll xe3 1-1 .

15 . . . �b4 Not wishing to be condemned to passive defence, Black plays

va banque, but the exchange sacrifice fails to justify itself. It may have been worth taking the pawn with 15 . . . f/Jxbl, even though this involves a definite risk. On the other hand, Black has a way of sacrificing a pawn himself. This improvement was worked out by Kasparov, who used it in a clock simultaneous display at a

Russian System 123

Young Pioneers' tournament. His opponent in this game was the only master in the contest :

Dzhandzhgava-Kasparov, Baku (Simul) 1 987 : 15 • • • J:t ad8! (possibly it was because of this move that Belyavsky refrained from 1 5 .id3 and played 1 5 d6) 1 6 J:t de 1 ( 1 6 J:tfe 1 'ifxb2 1 7 J.xe4 J:t xe4 1 8 J:t xe4 .ixe4 19 'ifxe4 'ifxb5 20 d6 deserved consideration. Also worth noting is 16 llJxa7 'ifxb2 17 llJb5 llJb4 18 J.b1 , when 1 8 . . . llJxd5 19 J:t xd5 J:t xd5 20 'ifxd5 'ifxb5 appears to fail after 2 1 llJe5! However, 2 1 . . . J.xe5 22 .ixe5 llJd2! is better for Black ed.) 16 . . . 'ifxb2 1 7 llJc7 llJxc7 1 8 J.xc7 llJd2! (a pretty resource, eliminating all inconveniences. In the sharp endgame, Kasparov easily outplays his young opponent) 19 J:t xe8 + J:t xe8 20 llJxd2 'ifxd2 21 J.xf5 gf 22 g3 J.d4 23 d6 J:te 1 24 �g2 J:t xfl 25 'ifxfl 'ifxa2 26 'ifb5 �g7 27 �h3 'ifd5 28 'ife8 'ifc4 ! 29 �g2 W"c2 30 W"e1 W"a4! 31 W"d2 1Wc6+ 32 f3 J.f6 0-1 .

At move 1 5, Black also has the interesting 1 5 . . . J.d7. This occurred in lvanchuk-Dorfman, Lvov 1988. The further course of the game was fascinating : 1 6 J.e5 J.xb5 1 7 W"xb5 J:txe5 1 8 llJxe5 llJd6! 19 llJg4 'iff4 20 1Wd7 c4 2 1 g3 llJc5 22 W"c7 W"xg4 23 'ifxc5 'ifd7 24 J.e2 .ixb2 25 1Wb4 c3 26 J.d3 a5 27 W"f4 J.a3 28 J:tfe 1 J.c5 29 �g2 b5 20 h4 h5 31 'iff6 'ifd8 32 'ifxc3 J.b4 33 1We5 J.xe 1 34 J:t xe l . The concentrated tactical crossfire has at last culminated peacefully; the players agreed a draw a few moves later.

16 llJc7 llJxd3 1 7 llJxe8 J:t xe8 18 'ifxd3! To Black's misfortune, 18 . . . llJxf2 or 18 . . . llJg3 would fail to

the counter-stroke 19 'ifb5. 18 . . . 'ifxb2 For the exchange Black has a pawn and an active position, but

the passed pawn on d5 has yet to reveal its full potential. 19 J:t de1 The straightforward 19 d6 J:t d8! 20 'ife3 h5 21 J:t b 1 W"xa2 22

J:t xb7 W"d5 23 J:txa7 llJxd6 24 .ixd6 W"xd6 25 llJg5 is unpleasant for Black and was undoubtedly worth playing. But I decided to pin the knight, restricting the activity of the enemy pieces. White incidentally threatens 20 g4. Of course, 19 J:tfe1 doesn't work in view of 19 . . . W"xf2+ 20 �h1 "ifxe 1 + ! 2 1 J:t xel llJf2+ 22 �g1 J:t xe1 + and 23 . . . .ixd3.

19 . . . 'ifb4 The decisive mistake. After 1 9 . . . llJf6 20 J:t xe8+ llJxe8 2 1 'ife3,

124 Russian System

White can realise his material advantage without trouble, but with 1 9 . . . 'lfxa2 20 11Vb5 (20 g4 c4!) 20 . . . l: d8 2 1 Wxb7 11Vxd5, Black can hold out. In his notes to the game, Kasparov pointed out that after 22 11Vxd5 l: xd5 23 g4, Black has the neat 23 . . . ltlf6! .

20 ltld2! Playing to exploit the pin is the chief factor in White's strategy. 20 . . . '1Va4 21 Wc4 Forcing the queens off, whereas 2 1 ltlxe4 is not so clear : 2 1 . . .

l: xe4 22 l: xe4 .txe4 23 11Vd2 c4, with counterplay. 21 . . . '1Vxc4 22 ltlxc4 .tc3 Black would lose at once with 22 . . . b5 23 ltld2 ltlf6 24 l: xe8+

ltlxe8 25 l:e1 etc. 23 ltld2 .txd2 24 .txd2 .td7 (79)

79 w

To many onlookers during the game, Black seemed to have 'wriggled out', and indeed done more than that. The bishop on d2 is attacked, and Black threatens . . . ..tb5. But White is able to give the exchange back . . . .

25 J..f4! .tb5 26 f3! g5 After 26 . . . .txfl 27 �xfl ltlf6 28 l:xe8 + ltlxe8 29 .te5 ! , the

d-pawn finally shows its full capability : 29 . . . f6 30 d6! and Black has to give up a piece.

27 .txg5 .txf1 Nor is 27 . . . ltlxg5 any better : 28 l:[ xe8 + .txe8 29 h4 ! , winning

the knight. 28 �x/1 ltld6 29 J..e7 ltlc8 30 .txc5 l: d8 31 J: e5 f6 32 l:/5

b6 33 J..d4 ltle7 34 .txf6 J: xd5 35 J:g5+ J: xg5 36 J..xg5 ltlc6 37 �e2 �!7 38 �d3 �e6

After 38 . . . ltlb4+ 39 �c4 ltlxa2 40 �b3, the black knight is trapped.

Russian System 125

39 �c4 ttJe5+ 40 �d4 ttJc6+ 41 �c4 Here the game was adjourned, and Black resigned without

resuming. After 41 . . . ttJe5+, White can achieve his ends with either 42 �b5 or 42 �d4 ttJc6+ 43 �e4, and the white pawns are not to be stopped.

We now return to the main game, which concluded very quickly : 14 ..txc3 15 be b5 16 1hb5 ttJxc3 17 1ha6

The correct plan, involving an exchange sacrifice, was demonstrated by Gurevich in another game : 1 7 'it'c4!? ttJxd 1 1 8 J:l xd 1 ..txd3 19 J:l xd3 'it'b6 20 J:l b3 'it'f6 2 1 g3 ! J:l ad8 22 ..tg5 1Wd6 23 ..tf4 'it'f6 24 a3!; M. Gurevich-Kotronias, Reykjavik 1988. The black knight is badly placed, and the activity of White's pieces assures him of more than enough for the exchange.

17 ..txd3 18 'it'xd3

If 1 8 J:l xd3, then 1 8 . . . ttJe2 + 19 �h1 ttJxf4 is good for Black. 18 ttJe2 + 19 �h1 ttJxf4 20 'it'c4 'it'd6 (80)

In an earlier game Ptin-Lau, played in a postal tournament ( 1 975), Black continued with 20 . . . 'it'f6, leading to equality.

It is strange that thirteen years later (this game was played in the 55th USSR Championship), Grandmaster Gurevich - who is one of Kasparov's seconds, too - should bring about this position again. It is only now that Black plays a new move, and the resulting ppsition can already be assessed as good for him.

80 w

126 Russian System

21 .l:l. fel After 21 g3 (2 1 lt)g5 'ife5! 22 lt)f3 'iff5) 21 . . . lt)h5 22 .l: fe l (22

lt)d2 lt)f6 23 lt)b3 .l:l. ac8) 22 . . . lt)f6 23 .l: xe8 + .l: xe8 24 .l:c l .l: c8 25 .l: d l .l: d8, Black has a slight advantage.

Apparently, 21 lt)d2 lt)xd5 22 lt)e4 (but not 22 lt)b3 .l: ad8) 22 . . . 'iff4 (22 . . . 'ife5) 23 'ifxc5 .l: xe4 24 'ifxd5 .l: ae8 would have maintained equality (Kasparov). Incidentally, after 2 1 lt)d2 a bad line for Black is 21 . . . .l: ad8 22 lt)e4 'ife5 23 f3 �g7 24 d6 'ifb2 25 :t g l 'ifb4 26 'ifcl ! lt)e2 27 'ifal + lt)d4 28 .l: b l ! 'ifa5 29 : bel 9b6 30 .l: gd l .l:l. e6 3 1 lt)xc5 .l:exd6 32 .l:c4, and White wins; Annageldiev-Chudinovskikh, Yalta 1 989.

White's unfortunate rook move leads to a quick catastrophe. 21 : xe1 + 22 lt)xe1

White has also a difficult position after 22 l:t xe1 lt)xd5 23 .l: d 1 lt)b6 2 4 9xf7+ �xf7 2 5 .l:l. xd6 �e7, but now Black succeeds in increasing his advantage decisively.

22 .l: b8! 23 a3

23 lt)c2 loses to 23 . . . 9xd5! . If 23 g3, Black has 23 . . . : b4 ! , while 23 lt)d3 is met by 23 . . . �xd3 24 9xd3 : b2.

23 . . . : b2 24 f3 9e5! 25 'ife4 9g5 26 g3 9h5 27 h4 lt)e2 28 9e8+ �g7 29 d6 lt)xg3+ 30 �g1 lt)e2+ 31 �fl 9f5 32 9xe2 9h3+ 0-1

Game No. 1 8 Belyavsky-Tukmakov

Odessa 1989 1 d4 lt)f6 2 c4 g6 3 lt)cJ d5 4 lt)f3 J..g7 5 9b3 de 6 9xc4 0-0

7 e4 lt)a6 8 J..e2 c5 9 d5 e6 10 0-0 ed 1 1 ed J..f5 12 J..f4 (81)

The more 'modest' bishop development with 1 2 J..e3, which has been seen frequently of late, is examined in Game No. 1 9.

A standard position in this variation. The contemporary line 12 • . . .l: e8 was examined in detail in Game No. 17 (where the move-order was 1 1 . . . l:t e8 1 2 J..f4 J..f5). Before demonstrating the valuable novelty employed by Tukmakov in the present game, let us recall the 'classical' continuation 12 . . . 'ifb6, which in our day has been undergoing further development. An entertaining

81 B

Russian System 127

instance of this occurred in the 1987 USSR Championship in Minsk, where Grandmaster Gavrikov, one of the most noted specialists on the black side of the Griinfeld, employed this queen move three times.

Belyavsky-Gavrikov continued 13 h3? (an unsuccessful innovation) 1 3 . . . 'ifxb2! 14 g4 i,c2! 1 5 .l:t ac1 �d7 1 6 �b5 i.a4! 1 7 �d6 (Black also has the advantage after 1 7 .l:t b 1 �b6 1 8 .l:t xb2 �xc4 19 .1:1. bb 1 , but now the White position becomes highly critical) 17 . . . �b6 1 8 'ife4 ( 1 8 'ifd3 �b4) 1 8 . . . .l:t ae8 etc.

A more logical plan was tried in Gurevich-Gavrikov from the same event : 13 i.e5.

This manoeuvre is the most dangerous to Black. 1 3 �a4 can be met by 1 3 . . . 'if b4 or 1 3 . . . 'ifa5, and 1 3 .l:t ad 1 by 1 3 . . . 'ifxb2.

13 . . . .l:t ad8 (82)

82 w

Alternatives are 13 . . . .l:t fe8 and 13 • • . �e8, but taking the bpawn is risky; after 13 • • • 'ifxb2 14 �e4 'ifb6 15 �d6, or IS i.xf6 .i.xf6 1 6 .l:t ab l , Black is in great danger. On the other hand, in Ivanchuk-Lputian, Irkutsk 1986, White played the mistaken IS

128 Russian System

J: ab1? t'Llb4 1 6 ..txf6, and now with 16 . . . ..txe4 ! 1 7 J..xg7 ( 1 7 J..e7 J..xd5!) 1 7 . . . J..xb 1 1 8 J..xf8 J..xa2 1 9 'it'xc5 J: xf8, Black could have achieved a winning position.

14 J:fd1 Perhaps 14 J: ad1 was more exact. 14 . . . J:fe8 Against Lputian (this is the last in the trio of examples),

Gavrikov played 14 . . . t'Lle8 1 5 t'Lle4 'it'a5 1 6 J..xg7 <l;xg7 1 7 J: ac 1 t'Lld6 1 8 'it'b3 (it was better t o play a t once 1 8 'it'c3+ 'it'xc3 19 t'Llxc3, with equality) 18 . . . J..d7 19 'it'c3 + 'it'xc3 20 t'Llxc3, and now after the retreat of the knight to b8 or c7 Black would not be threatened with anything. Instead he continued with 20 . . . J: fe8?, when 2 1 .i.xa6 enabled White to obtain a big endgame advantage.

15 'it'h4 The immediate 1 5 b3 is more accurate. 15 . . . t'Lld7 16 J..xg7 <l;xg7 1 7 b3 'it'a5 18 J: acl t'Llf6

Black has no problem whatsoever, and after 19 ..tb5 .i.d7 20 ..txa6 'it'xa6 21 t'Lld2 h6 22 'it'c4 'it'a5 23 fl 'it'b4 24 t'Llde4 'it'xc4 25 be t'Llxe4 26 t'Llxe4 b6 27 �f2 f5 28 t'Llc3 f4 29 g3 the players agreed a draw.

After these three games by Gavrikov, the impression was formed that in the 1 2 . . . 'it'b6 line White has no particular prospects of gaining the initiative. An attempt to refute this opinion was undertaken in the game Vladimirov-Popovic, Moscow 1 989, which made a substantial contribution to theory. From diagram 82, play went : 14 d6! 'it'b4 (Black goes into a sharp endgame; 14 . . . 'it'xb2 loses to 15 t'Lld5) 15 J: ad 1 'it'xc4 1 6 J..xc4 t'Llb4 (this game has been thoroughly annotated by the winner and by Grandmaster Makarichev and International Master N. Andrianov. I shall quote some of the interesting variations they give. At this point, Black could perhaps have got off with a slight scare by continuing 1 6 . . . t'Lle4 1 7 J..xg7 <l;xg7 1 8 t'Llxe4 J..xe4 19 J: fe 1 J..xf3 20 gf) 1 7 h3 ! (not prophylaxis, but preparation for an attack) 1 7 . . . a6 (it is now too late for 1 7 . . . t'Lle4, in view of 1 8 g4! t'Llxc3 19 be J..e4 20 .i.xg7, and White wins) 1 8 a3 t'Llc6 19 g4 .i.d7 ( 19 . . . ..te6 i s worse : 20 ..txe6 fe 2 1 J: fe 1 t'Llxg4 22 hg J: xf3 23 ..txg7 <l;xg7 24 t'Lle4 J: f4 25 t'Llxc5 J: xg4+ 26 �fl . with a solid advantage to White) 20 J:fel b5 2 1 ..td5! (Black has all the time been refraining from . . . t'Llc6xe5 in view of the dangerous pressure

Russian System 129

against f7. But his attempt to drive the bishop off its diagonal or block it in with 2 1 .ta2 c4 does not succeed. White is prepared to exchange this bishop only if it can be replaced on its active square by a knight) 2 1 . . . ll:lxe5 (21 . . . b4 would be answered by 22 ab cb 23 .txf6 .txf6 24 ll:la4 and 25 ll:lc5, but 2 1 . . . J:t fe8! was stronger) 22 lLixe5 b4 23 lLixd7! (the point of White's plan lies in the variation 23 . . . J:t xd7 24 lLia4 J:t fd8 25 lLixc5 J:t xd6 26 lLib7 ll:lxd5 27 ll:lxd8 J:t xd8 28 .l:l d3 .txb2 29 ab followed by J:t ed 1 , winning. However, after 24 . . . lLixd5! 25 J:txd5 J:t fd8 26 ll:lxc5 J:t xd6 27 J:t xd6 J:t xd6 28 J:te8+ .tf8 29 ab �g7, Black would have every chance of holding on) 23 . . . ll:lxd7? (now Black has no counterplay at all) 24 ab cb 25 ll:la4 .tf6 26 J:t e4 a5 27 .tc6 ll:le5 28 .tb5 J:t b8 29 J:t d5 (Black's extra pawn on the queenside has no significance and White wins the ending, creating quite a good textbook example in the process.) 29 . . . lLif3+ 30 �g2 lLig5 3 1 J:t c4 lLie6 3 2 d 7 g 5 33 �fl 'li1g7 3 4 .tc6 .td8 35 J:t f5 b 3 (35 . . . f6 was more tenacious) 36 ..td5 ll:lf4 37 J:t c8 .tf6 38 J:t xf8 J:t xf8 39 .txb3 J:t d8 40 ll:lc5 J:t b8 41 ll:le4 ..txb2 42 J:t xf7 + �g6 43 ll:ld6 1 -0.

Returning to diagram 8 1 , we observe that the Odessa grandmaster excluded himself from the debate on the variation 12 . . . "ifb6 1 3 ..te5, and instead employed a very valuable novelty, which, it appears, he had prepared specially for this USSR Championship.

83 w

12 ll:ld7! (83)

By this means (the rook on f8 and queen on d8 are both left where they are for now), Tukmakov succeeds in drawing his opponent into a tactical whirlpool.

130 Russian System

13 d6 Although this advance fits into White's plans, in the present

situation it is rather committal. It was worth considering 1 3 .1g5, or 1 3 lHd 1 .!l:lb6 14 1i'b3 .!l:lb4 1 5 .J:l d2!

13 tL!b6 14 1i'b3 .!l:lb4 15 .J:lacl

Perhaps the queen should have returned home with 15 1i'dl . 15 .i.e6

Black has many possibilities - 15 . . . a5, 15 . . . .!l:ld3, 15 . . . .i.d3 - but the move played is not bad either.

16 'ifa3 Again it was safer for the queen to return to its own camp : 1 6

Wd1 .!l:l6d5 ( 16 . . . .i.xc3 1 7 be .!l:lxa2 1 8 .J:l c2 .!l:la4 1 9 Wd2) 1 7 .!l:lxd5 .!l:lxd5 1 8 .i.g3 .i.xb2 19 .J:l xc5 .!l:lc3 20 'ifc2 .!l:lxe2+ 2 1 'ifxe2 .i.a3 22 .J:l c3 .i.b4 2 3 .J:lc2, and White's chances are a little better (Tukmakov).

16 17 .i.xc4 18 .J:l fd1

White's hope of 18 . . . .!l:ld3 disappointed.

19 tL!e4

.!l:lc4 .i.xc4 b6

19 .J:l xd3 .i.xd3 20 'ifxc5 is

Here again, White had to play 1 9 'ifa4 a6 20 .J:l d2, and then 'ifa4-d l . It is becoming clear that the white queen is out of the action.

19 .i.d5! 20 .!l:lfg5

Black also has a big advantage after 20 .!l:lc3 .i.xf3 2 1 gf 1tf6 22 .i.g3 'ifxf3 23 .!l:ld5 'ifxa3 24 ba .!l:lc6! 25 .!l:le7 + .!l:lxe7 26 de .J:l fe8 27 .J:l d7 .i.d4. Now after 20 . . . h6, White is counting on bringing his queen across to h3 (2 1 Wh3 .J:le8 ! 22 d7 .J:le7).

20 Wd7! (84) After this cool rejoinder, White's pieces turn out to be most

precariously placed. 21 .i.d2

Against the piece sacrifice 2 1 .!l:lxc5 be 22 .J:l xc5, Black has the decisive 22 . . . 'ifb7!, since 23 .J:l cxd5 tL!xd5 24 Wf3 fails to 24 . . . 1i'b4 !

21 .!l:lxa2

84 w

22 lla1 h6 23 ltlc3 ltlxc3 24 ..txc3 ..txg2 25 ..txg7 �xg7 26 �xg2 hg

Russian System 131

As well as two extra pawns Black has the better position, as the d-pawn presents no threat.

27 1tg3 llfe8 Black could also calmly play 27 . . . g4.

28 1txg5 .l:!.e4 29 h3 .l:l. ae8 30 lla3 .l:t 8e5

30 . . . a5 is simpler. 31 1tg3 .l:t e8 32 ll f3 a5 33 lld5 ll 4e6 34 .1:1. £4

This loses at once - the f7 point is invulnerable, and Black wins the d-pawn. 34 h4 was more stubborn.

34 ll d8! 35 1tb3 1tc6 36 1Wf3 .l:!.d7 37 �g1 .l:t exd6

0-1

Game No. 1 9 Bareyev-Lputian

Lvov 1990 1 d4 ltlf6 2 c4 g6 3 ltlc3 d5 4 ltlf3 ..tg7 5 1tb3 de 6 1txc4 0-0

7 e4 ltla6 8 ..te2 c5 9 d5 e6 10 0-0 ed 1 1 ed ..tf5

132 Russian System

12 .i.e3 The alternatives 1 2 .i.f4 and 1 2 l:t d 1 were explained in the

context of Game No. 1 7. 12 .i.g5 is also occasionally seen. An old game Brinck-Claussen

Jakobsen, Denmark 1 970, continued 12 . . . h6 1 3 .i.xf6 11fxf6 1 4 l:t ad 1 l:t ad8 1 5 -*.d3 l:t fe8 1 6 .i.xf5 11fxf5, and White had no more than a minimal plus.

85 w

12 l:t e8 (85)

13 l:t acl The alternative is 13 l:t ad1 ; let us look at a few interesting

examples. A game Ree-de Boer, played some time ago (Amsterdam 1983),

went 13 • • • .!De4 14 .!Dxe4 l:t xe4 1 5 11fc1 .!Db4 16 a3 .!Dc2 17 .i.g5 f6 18 11fxc2 fg 19 .i.d3 lH4 20 11fxc5 g4 21 11fe3, and White gained a clear advantage. An improvement is 1 5 . . . 11rb6 1 6 b3 l:t ee8 1 7 .!Dd2 .!Db4 1 8 .!Dc4, with a minimal plus for White.

Farago-Kozul, Montpellier 1989, went 13 . . . 1i'b6 14 b3 l:t xe3 !? ( 14 . . . 11t'b4 leads to unclear play. In Ree-Chandler, Helsinki 1984, Black ended up in a difficult position after 1 4 . . . .!Dg4 1 5 -*.d2 l:t ad8 1 6 l:t fe 1 .!Db4 1 7 .!Da4 11rd6 1 8 .i.f4 ! 1Wf8 19 .i.c7 l:t d7 20 d6. Black also stands worse after 14 . . . .!De4 1 5 .!Dxe4 l:t xe4 1 6 11t'b5) 1 5 fe .!Dg4 1 6 .!Da4 11rd6 ( 1 6 . . . 11fa5 1 7 e4 .!De3 1 8 11rd3 .!Dxd 1 19 ef .!Dc3 20 .!Dxc3 is in White's favour) 1 7 11rf4 ! (better than 1 7 e4? -*.d7 1 8 11fc1 b 5 1 9 .!Db2 .!Db4, with sufficient compensation for the exchange, for example : 20 .!Dd3 .i.d4+ 21 <oth1 .!Dxa2, with advantage to Black) 17 . . . 11fxf4 1 8 ef .!Db4 ! (but not 18 . . . .!De3, in view of 1 9 .i.xa6 ba 20 .!Dxc5) 1 9 l:t d2 .!De3 20 l:tc 1 .!Dbxd5

Russian System 133

2 1 i.d3 lL!xf4 22 i.xf5 lL!xf5 23 lL!xc5 i.h6 24 J: e 1 , and the sharp play eventually led to a peaceful conclusion.

In another game, the Yugoslav grandmaster Kozul played the white side, and had to contend with the novelty 13 • • • b6. There followed 14 h3 J: c8 (better than 14 . . . 1Vb6 1 5 b3 1Vb4 1 6 lL!a4, or 14 . . . lL!e4 15 lL!xe4 l: xe4 16 1Vc1 , hitting c5 and h6) 15 J: d2 ( 1 5 1Vb3! is stronger) 1 5 . . . lL!e4 1 6 lL!xe4 l: xe4 1 7 1Vcl c4! 1 8 lL!d4 i.d7 1 9 lL!e6!? fe 20 de J: xe6 2 1 i.xc4 1Ve8 22 J: fd l i.b5? (22 . . . l: xc4 ! is correct; Kozul gives the variation 25 1Vxc4 i.b5 24 1Vb3 i.a4 25 "ii'xb7 i.xd 1 26 J: xd 1 J: xe3! 27 fe 1Vxe3 + 28 �h1 lL!c5 29 1Vxa7 �h7 30 J:fl 1Ve2 31 J:f7 1Ve1 + , with a draw) 23 i.xe6+ 1Vxe6 24 1Vxc8+ 1Vxc8 25 ll d8 + 1Vxd8 26 ll xd8 + �f7 27 b3 lL!b4 28 a4 i.c6 29 i.xa7, and White won the ending; Kozul-Popovic, Yugoslavia 1 989.

Van der Sterren-Miralles, Lyon 1990, went 13 • . . J: c8 (again Black innovates - this time successfully) 14 h3 lL!b4 1 5 lL!e1 lL!e4 16 lL!xe4 (better is 1 6 a3 lL!xc3 1 7 be lL!c2 1 8 lL!xc2 i.xc2 19 ll d2 i.e4, with unclear play - Miralles) 1 6 . . . J: xe4 1 7 1Vb3 a5! 1 8 g4 i.d7 19 l: c 1 a4 20 1Vd1 a3 ! 2 1 ba lL!xa2 22 i.d3 J: xe3 23 fe lL!xc 1 24 1Vxc 1 c4 25 i.c2 (Black also has the advantage after 25 i.e4 1Ve7 26 .i.f3 i.h6 27 lL!c2 c3. Now the c-pawn settles matters) 25 . . . 1Vg5 26 1Vd2 c3 27 1Vd3 J: c5 28 .i.b3 c2! etc.

13 lL!e4 The immediate 1 3 . . . 1Vb6 deserves attention.

14 1Vb3 Another possibility is 14 J: fd 1 lL!d6 ( 14 . . . 1Vb6 1 5 lL!a4) 1 5 1Vf4,

with complex play. 14 1Vb6! 15 i.b5

A level game results from 15 J: fd1 1Vxb3 16 ab J: ad8 1 7 lL!xe4 i.xe4 18 i.xa6 ba 19 J: xc5. The chances are approximately equal after 15 i.xa6 ba 1 6 lL!xe4 i.xe4 1 7 l: xc5 1Vxb3 18 ab i.xb2.

15 J: ed8 16 lL!h4 (86)

Rather a risky move; 1 6 J:fd 1 , defending the central pawn, was safer.

16 lL!xc3 17 be i.e4 18 c4 lL!c7

134 Russian System

86 B

19 -*.a4 'lff6! The awkward position of the knight on the edge of the board

makes itself felt. 20 l0f3

Not 20 f3, on account of 20 . . . -*.xd5! 20 -*.xf3 21 gf b6 22 -*.d2

22 �g2 'lfh4, followed by . . . -*.e5, is bad for White. 22 l0e6?!

This looks pretty, but now White extricates himself from a tricky situation. His defence would have been more difficult after 22 . . . l0e8! 23 .*.xeS (otherwise 23 . . . l0d6) 23 . . . l:l. xe8 24 l:l. fe l 'lfh4.

23 de 24 "ife3! 25 "ife4 26 ef+ 27 l:l. cdt 28 'lfg4!

.I:Xd2 l:l.d6 l:l. b8 �xf7 'lfg5+

More exact than 28 �hl -*.d4 29 -tc6 �g7 30 -*.d5 l:l. f8. A level endgame now arises.

28 • • . 'lfxg4+ 29 fg l:l. bd8 30 l:l. xd6 l:l. xd6 31 l:l. dt l:l. xdl + 32 -txdt g5 33 �g2 �e6 34 �f3 -*.d4 35 �e4 -*.xf2 36 h3

t-t.

5 4 ..tf4 System

Game No. 20 Lukacs-Ftai!nik

Stara Zagora 1990 The reader will no doubt recall that for the eleventh game of

the return World Championship Match in 1 986, a brilliancy prize was awarded - to both players. Indeed, that game was one of the most fascinating played with the Griinfeld in recent years. We shall, of course, examine it here (in addition to other important games with the same variation) in the notes to the principal game.

1 d4 ll:lf6 2 c4 g6 3 ll:lc3 d5 4 J.f4

Theory sometimes refers to this variation as the Classical Line. 4 .ig7 5 e3 c5

This system was used in the first game of the 1 986 match, but with a different move-order (e2-e3 was postponed for three moves) : 4 ll:lf3 .ig7 5 .if4 c5 6 de "ifa5 7 l:l c 1 (previously, 7 cd had been played; Kasparov responds to the novelty very precisely) 7 . . . de! 8 e3 "ifxc5 9 'ira4+ ll:lc6 10 .ixc4 0-0 1 1 0-0 .id7 12 'irb5 'irxb5 1 3 J.xb5 l:l ac8 14 l:l fd 1 l:l fd8 1 5 h3 h6 1 6 �fl a6 1 7 J.e2 J.e6 1 8 l:l xd8 + l:l xd8 19 ll:le5 ll:lxe5 20 J.xe5 l:l d2

!-! . 6 de

Winning a pawn with 6 J.xb8 l:l xb8 7 'ira4+ .id7 8 'irxa7 is too risky: 8 . . . cd 9 'irxd4 0-0 10 cd 'ira5 1 1 'ird2 b5! 1 2 .id3 b4 1 3 ll:lce2 'irxd5.

6 'ira5 7 l:ld (87)

136 4 J.f4 System

87 B

Here too, the pursuit of material gains brings trouble : 7 cd ltlxd5 8 'irxd5 J.xc3 + 9 be 'irxc3+ 10 �e2 'irxa 1 1 1 J.e5 'irb1 1 2 J.xh8 J.e6 1 3 'ird3 'irxa2+, with a very strong initiative for Black.

Instead, Timman-Kasparov, Belfort 1988, continued as follows : 7 'ira4+ 'irxa4 8 lt!xa4 0-0!? After 8 . . . ltle4 9 J.xb8 .*.d7 10 f3 .i.xa4, the game is level.

But Kasparov is capable of finding new ideas even in seemingly peaceful positions.

9 lt!f3 Salov-Korchnoi, Brussels 1 988, went 9 .l:l c 1 .i.d7 10 ltlc3 de

1 1 J.xc4 lL!a6 1 2 lL!f3 lL!xc5 1 3 �e2 .l:l ac8 14 .te5 a6 1 5 a3 lL!a4 1 6 lL!xa4 .txa4 1 7 ltld4 ltle4 1 8 .txg7 �xg7 !-! .

9 . . . lt!e4 10 J.e5 J.d7 11 lt!c3 lL!xc3 12 be de 13 .txe4 .l:l e8 14 .td4 (88)

88 B

An improvement. In an earlier game between the same opponents (Amsterdam 1988), White played the ineffective 14

4 .i.f4 System 137

.t.d5? .t.c6 15 .t.xc6 : xc6 ( 15 . . . tllxc6! 16 .t.xg7 �xg7 17 0-0-0 tlld8 18 ll f8 1 9 ll hd 1 �e8 20 ll 7d5 tlle6 would, in Kasparov's view, have given Black a plus) 1 6 .t.xg7 �xg7 17 : b1 llc7 1 8 tlld4 tlla6 19 c6 b6 20 f4 tllb8 21 : b4 tllxc6 22 ll c4 ll ac8 23 tllb5 ll d7 24 tlld4 ll dc7 !-t .

14 . . . e5! 15 .t.xe5 llxe5 16 .t.xg7 �xg7 1 7 .t.b3 llxc3 18 0-0

White could have kept the game equal with 1 8 �d2 or 1 8 ll d l . Now Black seizes the initiative.

18 . . . tlla6! 19 tlle5 .t.e8 20 .t.d5 ll e7 21 :ab1 tlle5 22 e4 ll d8 23 llfcl llde8 24 g4 f6 25 tllf3 b6 26 tlld4 J.d7 27 f3 tlld3! 28 ll xe7 ll xe7 29 ll d1 tllf4 30 �!2 �f8 31 .t.b3 �e7 32 tlle2 tllxe2 33 �xe2 ll c3

Black has an obvious endgame plus, and he quickly converts it to a win.

34 h4 h6 35 e5? J.b5+ 36 �!2 fe 37 ll d5 ll e5 38 llxe5 be 39 g5 hg 40 hg J.d3 41 J.g8 .t.f5 42 .t.b3 .t.e6 43 J.e2 .t.xa2 44 .t.xg6 a5 45 �e3 a4 0-1.

7 de 0-0 8

9 10

J.xc4 tllf3 .t.b3

Wxc5

10 tllb5 will be mentioned in the notes to Game No. 2 1 . 1 0 tllc6

Game 2 1 also introduces the idea of playing the knight to a6.

89 w

1 1 0-0 WaS (89)

I should explain that the specific game which we are examining reached the diagram position by way of 7 tllf3 0-0 8 llc 1 de 9 .t.xc4 Wxc5 10 .t.b3 WaS 1 1 0-0 tllc6 . For convenience in the

138 4 J.f4 System

layout of material, I have given the opening moves in a different order.

12 h3 J.f5 12 • • • 1Wa6!? is also interesting. It was first played by Mecking

against Ribli in Manila 1 976. The idea is familiar - Black tries to eploit the weakness of the d3-square in the White camp. If 13 ll)d4, Black should not play 13 . . . l:.d8 14 J.c7 l:l. d7 1 S J.xf7+ �xf7 1 6 1Wb3 + �e8 1 7 ll)e6! with the unanswerable threats of ll)cS and ll)xg7+ . In Farago-Gavrikov, Amsterdam 1 987, Black replied correctly with 13 . . . J.d7, and after 14 ll)dbS a draw was agreed; in the event of 14 . . . l:l. ac8, the position js completely equal.

Another possibility (after 12 . . . 1Wa6) is 13 ll)a4 l:l. d8 14 ltlcS l:l. xd 1 1 S ll)xa6 l:l. xfl + 16 �xfl J.fS 17 ll)cS. In a game SchneiderItkis, USSR 1 987, White gained a big advantage after 17 . . . b6? 1 8 g4 ! J.c8 1 9 ll)e6!, but a new move 1 7 . . . ll)aS! was played in Mikhalchishin-Gavrikov, Lvov 1 987. There followed 1 8 J.c2 J.xc2 19 l:l. xc2 ll)dS 20 J.eS ll)b4 21 l:l. c3 b6 22 J.xg7 �xg7 23 ll)d3 ll)xd3 24 l:l. xd3 1-1 .

Finally, there i s a third option for White : 13 ll)e5 ll)xeS 14 J.xeS J.e6 1 S J.xe6 1Wxe6 1 6 J.d4 ( 1 6 J.h2 l:l. ac8 1 7 1Wa4 a6 1 8 iraS ll)e4 is good for Black; Sinkovics-Krasenkov, Budapest 1 988) 1 6 . . . l:l.fc8 17 1Wf3 1Wc6 18 1Wxc6 l:l. xc6 19 ll)bS l:l. xc 1 20 l:l. xc 1 a6, with an equal game; Levitt-Gulko, St John 1 988.

13 1We2 The ninth game of the 1986 match went 13 ll)d4 J.d7 14 Wel

ll)xd4 15 ed e6 16 J.dl "ii'b6, resulting in equal chances. Another four moves - 1 7 l:l. fd 1 J.c6 1 8 J.e3 iraS 19 ..td2 Wb6 20 J.e3 WaS - and we agreed a draw.

An attempt at strengthening White's play was made in Petursson-Ivanchuk, Reggio Emilia 1 989/90 :

16 J.e5 J.c6 17 l:l.fd1 l:l.fd8 18 1We3 l:l. d7 19 1Wg5 It is worth considering the sharp 19 d5 J.xd5 20 J.xf6 J.xf6

2 1 ll)xdS ed 22 Wf3 ..txb2 23 l:l. xdS! Ivanchuk examines two alternatives for Black: 19 . . . ed 20 Wd4 ll)e8 21 ..txg7 .!L\xg7 22 J.xdS l:l. ad8 23 WeS J.xdS 24 l:l. xdS l:l. xdS 2S .!LlxdS, and 19 . . .

.!Llxd5 20 ..txdS J.xdS 21 J.xg7 �xg7 22 .!Lle4! l:l. e8 23 Wf4 Wd8 24 .!LieS l:l. de7 (but not 24 . . . eS? 2S 1Wd2 l:l. d6 26 .!L\xb7 J.xb7 27 1Wxd6 irgS 28 f3! J.xf3 29 1Wd2, with advantage to White) 2S ll)e4 l:l. d7 - with equality in both cases.

4 j_f4 System 139

19 . . . "ild8 20 liJa4 h6 21 "flg3 liJh5 22 "fle3 "flh4 23 liJc5 j_xe5 24 de

The correct line was 24 11t'xe5 : e7 25 liJxe6!? j_xg2 26 oi>xg2 fe 27 j_xe6+ oi>h7 28 : c7 liJf4+ 29 oi>fl : ae8 30 d5 liJxh3 3 1 Wg3 Wxg3 3 2 : xe7+ : xe7 3 3 fg liJg5, although Black's chances would still have been slightly better.

24 . . . : xd1 + 25 j_xd1 liJf4 26 g3 1Wxh3 27 j_f3 liJd5 And Black went on to win.

13 liJe4 14 liJd5 eS (90)

In Rizhkov-Epishin, Leningrad 1 986, Black played 14 . . . liJc5 at once, which led to equality after 15 j_c4 e6 16 b4 1i'a3 1 7 liJe7+ liJxe7 1 8 be j_e4.

90 w

In the 'main' game which we are examining, the white queen's bishop calmly retreats to h2 - the familiar theoretical continuation. In the 'sensational' encounter in London ( 19th game, 1 986), I played the unexpected 1 5 : xc6!? In the digression which now follows, we shall examine that game as well as the current theoretical verdict on White's exchange sacrifice.

15 : xc6 ef So Kasparov decides to decline the sacrifice. Many commenta-

tors thought that accepting it would virtually lose outright : 15 . . . be 16 liJe7+ �h8 17 liJxe5 j_xe5 1 8 liJxc6, but now after 1 8 . . . "ild2! 19 j_xe5+ f6 it is White who suffers material loss. I was intending 17 liJxc6, and if 17 . . . "ilb6 ( 1 7 . . . Wc5 1 8 liJcxe5 "fle7 19 liJd4), then 18 liJcxeS. After 18 . . . j_e6, White has three choices : 19 j_xe6 1i'xe6 20 b3, with two pawns for the exchange; 19 liJc4

140 4 J.f4 System

.i.xc4 20 .t.xc4 lDc5, with unclear play; or 19 11fc2 .t.xb3 20 11fxe4 .i.e6, as in Gavrikov-Kochiev, Tallinn 1 987.

16 l:t c7 .i.e6 If 1 6 . . . fe 1 7 11fxe3 .i.xb2, then 1 8 lDd4 is strong; or if 1 7 . . .

lDd6, White has 1 8 l:t d 1 , with mounting pressure. 1 7 1i'e1! The natural-seeming 17 : xb7 .i.xd5 1 8 l:t b5 would give Black

the better game after either 18 . . . 11t'a6 19 .i.xd5 l:t ae8 or 1 8 . . . lDc3 19 be .i.xf3 20 gf 11fxc3. In a later game Szihigyi-Schmidt, Hungary 1986, Black played the even stronger 17 . . . lDd6! (depriving the rook of the b5-square) 18 lDe7 + ..th8 19 lDc6 'it'c5 20 .i.xe6 11fxc6, and White came out a piece down.

The queen move to e 1 was planned in my pre-game analysis, but later it was established that 17 lDe7+ is stronger : 17 . . . ..th8 1 8 l:t fc 1 -*.xb3 19 ab, with the initiative ( 1 9 . . . fe 20 11fxe3 lDd6 21 'it'f4).

1 7 . . . 1i'b5! The endgame after 17 . . . 11fxe1 1 8 l:t xe 1 .i.xb2 19 lDe7+ �h8

20 .t.xe6 fe 21 ef l:t xf4 22 l:t xb7 .t.c3 23 l:t e2 is unattractive for Black.

18 lDe7+ �h8 19 -*.xe6 fe 20 11t'b1! The white queen has made an unusual tour, 1i'd1-e2-e1-b1 ,

and now unexpectedly aims a t the opponent's kingside. 20 . . . lDg5! The only move. 20 . . . 11t'b6 is simply answered by 2 1 : re t , while

if the knight goes anywhere else, 2 1 lDxg6+ is decisive. 21 lDh4! (91)

91 B

After 2 1 lDxg5 (2 1 lDd4 11fe5) 2 1 . . . 11fxg5 22 ef J: xf4 23 l::t xb7 l:t e8 24 lDc6 11fc5 25 lDxa7 .i.d4, Black is dangerously active.

4 j.,f4 System 141

21 . . . toxh3+!? In this extremely sharp duel, each of us was hoping to outwit

the other. No doubt for that reason, Kasparov refrains from 2 1 . . . fe, after which I could have forced a draw with 2 2 tohxg6+ hg 23 toxg6+ �g8 24 toe7+ .

22 �h2 Of course not 22 gh 'lfg5 + 23 tog2 f3. 22 . . . 'lfh5? A serious error. 22 . . . toxf2! 23 l:t xf2 fe 24 .l:l xf8 + .l:l xf8 25

�h3 e2 26 'lfe4 'lfh5! 27 .l:l c4 g5 28 'lfxe6 gh leads to a draw. 23 toexg6+ If 23 �xh3, Black replies 23 . . . g5. Many observers thought that capturing with the other knight

was stronger : 23 tohxg6+ hg 24 'lfxg6. As we shall see, this variation is indeed unpleasant for Black, but it happens to be unimportant which knight takes on g6 first.

23 . . . hg 24 'lfxg6 An inaccuracy throwing away the win! 24 toxg6+ �g8 25

toe7 + �h8 26 'lfg6! (rejecting the perpetual check with 26 tog6+) would have given White a clear advantage; this is the same position that was arrived at in the previous note. I shall now quote the interesting variations given by Halifman : 26 . . • 'lfeS (26 . . . .l:lf5 is met by 27 gh fe 28 'lfxh5+ .l:l xh5 29 tog6+ �h7 30 fe!) 27 �xh3 fe (27 . . . .l:lf6 28 �g4!) 28 'lfg4! .l:l f6 29 'lfh4+ j_h6 30 f4 ! 'lfxb2 3 1 .l:l b 1 ! , etc. ; a more stubborn defence is 26 • . . 'lfh7, but here again, after 27 gh .l:lf6 (27 . . . j.e5 28 .l:l c5 ! ; or 27 . . . fe 28 'lfxe6 .l:lf6 29 'lfxe3 with two extra pawns) 28 'lfg4 (28 'lfxh7+ �xh7 29 .l:l g 1 is also good), Black is in a critical position since taking on e3 is no good: 28 . . . fe 29 fe! .l:l xfl 30 tog6+ �g8 3 1 'lfxe6+ .l:l f7 3 2 'lfxf7 mate.

24 . . . 'lfe5! An ingenious retort which gets Black out of danger. White

cannot now play 25 �xh3, as the rook on c7 is en prise (whereas with the knight on e7, the capture 25 . . . 'lfxc7 would be unplayable : 26 'lfh5 + and mates).

25 .l:l/7 Not 25 .l:l xg7 fe+ 26 f4 'lfxg7 27 'lfh5+ �g8 28 gh 'lfxb2+ 29

�h1 .l:l f7! , or 26 'lfg3 'lfxg7 27 tog6+ �g8 28 toxf8 tog5! 29 tod7 .l:ld8 30 toe5 e2 3 1 .l:l e l .l:l d l 32 tod3 'lfh7+ 33 �gl 'lfxd3 34 'lfxg5 + �f7 etc. On the other hand, 25 'lfc2 would have forced

142 4 �f4 System

a draw : 2S . . . fe+ 26 cot>xh3 'iPg8 27 f4 e2 28 fe ef('if 29 .l: xg7+ cot>xg7 30 'ifg6+ etc.

25 . . . .l: xf7! 26 'ifxj7 fiJg5! After the game I discovered that the computer had indicated

26 . . . 'ifbS here, but there would follow 27 fiJg6+ .,Ph7 28 fiJe7 'ife8 29 11t'xe6 fOgS 30 11t'fS + cot>h6 31 .l: h 1 ! and in spite of his extra peice Black has no defence.

27 fiJg6+ cot>h7 28 fiJxe5 liJxf7 29 liJxf7 'it>g6 30 fiJd6 fe The position has become simplified, the tension has abated, and

Black has even emerged with a slight edge, though its significance is purely symbolic.

31 fiJc4 ef 32 .l: xf2 b5 33 fiJe3 a5 34 .,Pg3 a4 35 .l: c2 .1:/8 36 'it>g4 �d4 37 .l:e2 �xe3 38 l:txe3 .1:/2 39 b3 .l: xg2+ 40 .,Pf3 .l: xa2 41 ba J-J.

15 .i.h2 fOeS 15 • • • fiJf6 is very strongly answered by 16 .l: xc6! fiJxdS 17 .l: d6

fiJe7 1 8 e4, with a clear advantage; Pinter-Rogers, Szirak 1986. But a more popular continuation is 15 • • • �e6 16 .l: fd1 (92) . Let us look at some interesting games with this line from recent years.

92 B

Farago-Smejkal, Baden-Baden 198S, went 16 • • • .,PhS 1 7 �c4 .l: ad8 1 8 b4 fiJxb4 1 9 fiJxb4 .l: xd 1 + 20 .l: xd 1 11t'xb4 2 1 �xeS, with a slight advantage to White.

Gleizerov-Vakhidov, Uzhgorod 1988, went 16 • • • .l: fe8!? 1 7 11t'c4 fiJf6 1 8 e4 .l: ac8 1 9 11t'c3 11t'xc3 20 fiJxf6+ �xf6 2 1 .l: xc3, and a draw was agreed.

Basin-Kozlov, Naberezhnie Chelny 1988, continued 16 • • . .l: fd8 1 7 11t'c4 (the exchanges with 1 7 11t'e 1 .l: d7 1 8 11t'xaS fiJxaS 19 fiJc7 fiJxb3 20 fiJxa8 fiJxc1 21 .l: xc 1 f6 are to Black's liking; Rajna-

4 J..f4 System 143

Nikoloff, St John 1 988) 17 . . . llld4? (the correct line is 1 7 . . . lllf6 1 8 e4 .l:l. ac8 1 9 lllg5 llld4 20 llle7+ �f8 2 1 tllxe6+ �xe7 22 lllxd8 .l:l. xc4 23 J..xc4 J..h6, and Black's chances are no worse; Huzman-Dorfman, Lvov 1 988) 1 8 ed J..xd5 19 "it'e2 J..h6 20 l:t c2 J..f4 2 1 J..xd5 "it'xd5 22 lllxe5 J..xe5 23 J..xe5 lllg5 24 "it'g4 llle6 25 "ifh4 g5 26 "it'h6 "it'e4 27 "it'f6 �f8 28 .l:l. cd2 "it'g6 29 d5 lllc5 30 d6 1 -0.

16 e4!? An interesting novelty. After 1 6 J..c4 .:. adS? 1 7 b4! lllxb4 1 8

llle7+ �h8 19 lllxf5 gf 20 lllxe5, White had a big advantage in Vainerman-Epishin, Norilsk 1 987, but Black has the much better 16 . . . e4 ! , with chances of seizing the initiative.

16 : adS! (93) 16 . . . J..xe4 loses to 1 7 J:t xc5 J..xf3 1 8 "ife3 J..h6 19 "it'xh6.

White also has the initiative after 16 . . . J..e6 17 J..c4, with a2-a3 to follow, or 16 . . . lllxb3 1 7 ab J..e6 1 8 .l:l. a 1 "it'c5 19 .l:l. fc1 "it'd6 20 .l:l.d1 "ifb8 2 1 b4. But 16 . . . tllxe4!? leads to sharp play.

93 w

17 "it'e3 An inaccuracy. In Ftacnik's view, an improvement is 1 7 .l:l. fd 1

J..xe4 1 8 .l:l. xc5 J..xf3 1 9 "it'e3 J..xd1 20 .l:l. xa5 J..xb3 2 1 ab lllxa5 22 llle7+ �h8 23 J..xe5 lllc6 24 J..xg7+ �xg7 25 "it'c3+ f6 26 lllxc6 be 27 "it'xc6, with somewhat the better ending.

17 lllxb3 18 ab J..xe4 19 J:t c5 "it'xc5 20 "it'xc5

20 lllf6+ J..xf6 2 1 "ifxc5 .l:l. d5 22 "ife3 was more exact.

144 4 J.f4 System

20 : xd5 21 'lfe4 .i.xf3 22 gf : fd8 23 J.g3 :dl 24 'lfb5 l0d4 25 'lfxb7 l0e2+

Black could have retained the better chances with 25 . . . Jlxb2 26 'lfe7 : f8 27 'lfxa7 l0xf3+ 28 �g2 : xb3.

26 �gl l0xg3 27 �xg3 J.b6 l8 'lfxa7 : xb2 29 'lfb6

The position has levelled out, and the game gradually heads towards a draw.

29 : d3 30 : at �g7!

Precisely played. Not 30 . . . e4? 3 1 'lfb8 + J.f8 32 : aS : xf3+ 33 �h4. Also 30 . . . J.f4+ 3 1 �g2 : dxb3 is dangerous in view of 32 : aS + �g7 33 'lfd8, with an attack (Ftacnik).

31 : a6 J.g5 32 'lfe7 J.f4+ 33 �gl : dl 34 'lfb6

l-t

Game No. 2 1 Belyavsky-Gavrikov

Moscow 1988 1 d4 l0f6 2 e4 g6 3 l0e3 d5 4 J.f4 J.g7 5 e3 e5 6 de 'lfa5 7

: ct de In the fifth game of the return World Championship Match,

London 1986, Kasparov played 7 . . . l0e4 here. The game was quite interesting, so let us take this opportunity to give it in full :

7 . . . l0e4 8 cd l0xc3 9 'lfd2 'lfxa2 10 be This variation came into use after the famous game Petrosian

Fischer, Candidates final 197 1 , which continued : 10 • . . 'lfa5 1 1 J.e4 l0d7 12 l0e2 l0e5 1 3 J.a2 J.fS 1 4 J.xe5! J.xe5 1 5 l0d4 'lfxc5 16 lOxfS gf 1 7 0-0, with a dangerous attack. Afterwards, various alternatives were proposed for Black ; the most effective of them belongs to Mikhalchishin : 12 • • . l0xe5 1 3 0-0 0-0 14 f3

4 J.f4 System 145

e5! 1 5 J.g3 b5 1 6 J.a2 1Wb6! 1 7 �h1 a5, with adequate counterplay.

Recently the investigations have switched to 12 10f3, which so far has brought variable success. In Agzamov-Gulko, Frunze 1985, the interesting continuation was 1 2 . . . /Oxc5 ( 1 2 . . . 0-0 1 3 0-0 /Oxc5 14 J.e5 J.xe5 1 5 /Oxe5 f6 16 .J:I. a 1 10e4 led t o sharp play in Razuvayev-Mikhalchishin, Minsk 1985) 1 3 J.e5 J.xe5! (a game between the same opponents a few months earlier in Sochi had ended quickly with 1 3 . . . 0-0 14 0-0 f6 1 5 .J:I. a 1 1Wd8 16 J.c7! 1Wd7 17 d6+ e6 18 /Od4 1Wf7 19 .J:I. a5 b6 20 .J:I. xc5! be 21 /Ob3 1Wd7 22 1Wd3! .J:I. d8 23 1We4 1-0) 14 /Oxe5 f6 15 10f3 0-0 1 6 /Od4 /Oe4 1 7 1Wb2 /Od6 1 8 J.a2 J.d7 19 0-0 .J:I. ac8 2 0 e4! 11'c5 21 .J:I.fe1 , with the initiative.

10 . . . 'llxd2+ 11 �xd2 10d7 12 J.b5 After 12 c6 be 1 3 de ltlb6 (or 13 . . . ltlf6), White achieves nothing. 12 . . . 0-0 13 J.xd7 This time, 1 3 c6 would be met by 1 3 . . . ltlc5. Now Black obtains

the two bishops, but it soon becomes clear that they will both be constricted.

13 . . . J.xd7 14 e4 f5 The alternative 14 . . . .J:I. ac8 is weaker : 1 5 J.e3 f5 16 f3. 15 e5 e6! A critical moment. The line that had previously been tested was

1 5 . . . .J:I. ac8 16 c6 ( 1 6 e6 J.a4 17 c4 .J:I. xc5 18 J.e3 .J:I.c7 19 /Of3 .J:r. fc8 20 �d3 b5! favours Black) 16 . . . be 17 d6 ed 18 ed .J:r. f6, with the advantage; Schmidt-Gross, Naleczow 1984. Why does Kasparov reject it? The answer is that after 1 5 . . . .J:I. ac8 White has the much stronger 1 6 c4 ! .J:I.xc5 17 J.e3. In Seirawan-Adorjan, New York 1 987, there followed 17 . . . .J:I. c7 ( 1 7 . . . .J:I. a5 18 f4 e6 19 d6) 1 8 10f3 b6 1 9 c5 be 20 .J:I. xc5 .J:I. xc5 2 1 J.xc5 .J:I. c8 22 J.xa7,

. and White won. 16 c4 .J:r.fc8 After 16 . . . g5 1 7 J.xg5 .i.xe5 1 8 10 f3 J.g7 1 9 .J:I. b 1 and

.J:I. he1 , White would complete his development while maintaining powerful pressure.

1 7 c6! The extra pawn cannot be retained, but in returning it White

extracts the maximum profit : he creates a passed pawn and limits the scope of Black's rooks and light-squared bishop.

146 4 j_f4 System

1 7 . . . be 18 d6 c5 Now the light-squared bishop acquires a little freedom, but the

dark-squared one begins to suffocate. Perhaps the fate of the bishops should have been decided 'the other way round', with 1 8 . . . g5 19 j.xg5 j.xe5 20 c5 .l:l. cb8, and Black has hopes of counterplay.

19 h4! h6 20 li:Jh3! (94)

94 B

Kasparov must have been examining the more natural-seeming 20 li:Jf3, which he could answer with 20 . . . j_c6!, leading to complex play with mutual chances. But I arrived at a solution to the position on mathematical lines. The white knight is heading for its ideal square, d3, by the only suitable route! Once the knight reaches its destination, the black bishop on g7 will be interned for good inside the cage formed by the pawns on d6 and e5 and the bishop on f4. At the present moment, White forestalls the freeing advance . . . g6-g5.

20 . . . a5 21 f3 a4 22 .l:l. hel! Over-protecting the pawn on e5. The hasty 22 li:Jf2 could be

met by 22 . . . g5! 23 hg hg 24 j_h2 f4 25 li:Jd3 .te8 26 j_gl , when the situation i s fairly obscure.

22 . . . a3 23 li:Jf2 a2 24 li:Jd3 .l:l. a3 25 .l:l. al g5 The alternative 25 . . . .1:1. b8 is inadequate on account of 26 :ee l

g5 27 hg hg 28 li:Jxc5 (but not 28 .txg5 .1:1. bb3 29 li:Jxc5 .1:1. b2+) 28 . . . j.a4 29 li:Jxa4 : xa4 30 j_xg5 j.xe5 31 d7.

26 hg hg 27 j_xg5 (95) Now after 27 . . . .1:1. b8, the contest could have ended in a study

like draw : 28 j_f4 .l:l. bb3 29 li:Jxc5 .l:l. b2+ 30 �c1 l:l xg2 3 1 j.d2 j_h6! 32 j.xh6 .l:l. c3 + 33 �d1 .l:l. d3 + ! 34 li:Jxd3 j.a4+ 35 �c1 .l:l.c2+, with perpetual check. But there is quite a simple

95 B

4 J.f4 System 147

refutation of the 'study' in 28 �e2! l:t bb3 29 .lt.:lxc5 ll b2+ 30 �fl . and White wins.

27 . . . �f7 28 J..f4 l:l. b8 29 l:l. ecl J..c6 30 l:l. c3 l:l. a5 31 l:l. c2 l:l. ba8 32 .lt:lcl ]-{)

8 J.xc4 0-0 9 .lt.:lf3

A couple of years ago, when Black was achieving good results with this Griinfeld variation, the rare move 9 .lt.:le2 was tried out in a game Barlov-Gulko, New York 1 988. There followed 9 • . •

1t'xc5 10 1t'b3 .lt.:lc6 1 1 .lObS 1t'h5 1 2 .lt.:lc7, and now after 1 2 . . . 1t'a5 + 1 3 1t'c3 1t'xc3 + 14 .lt.:lxc3 Black equalised.

In lnformator 45, Gulko considers 1l • • • l:t b8, and judges that White stands better on the basis of 1 3 J.xf7+ l:t xf7 14 l:t xc6 1t'a5 + 1 5 .lt.:lc3 .lt.:le4 1 6 .lt.:ld5! e6 1 7 l:t c7. But this assessment was later refuted by Krasenkov in his game against Toth, Mazatlan 1 988 : 16 . . . .lt.:lxc3 1 7 .lt.:lxc3 ( 1 7 l:t xc3 J.xc3 + 1 8 be J.e6 19 J.xb8 J.xd5 20 1t'b2 J.xg2 2 1 l:tg 1 was more tenacious, although even then Black would be better) 1 7 . . . be 1 8 J.xb8 ( 1 8 1t'xb8 J.xc3+ 19 be 1t'xc3 + 20 �e2 1t'c4+ 2 1 �f3 1t'e6! 22 �e2 1t'xa2+ , and White is in a bad way) 1 8 . . . e6!, and in view of the threatened 19 . . . l:t b7 and 20 . . . l:t xb2, White loses his bishop on b8. A game Zlochevsky-Krasenkov varied with 1 7 .lt.:lxe7+ �f8 1 8 l:t xc3 J.xc3 + 19 be l:t xf4! 20 ef �xe7 21 0-0 J.e6 22 l:t e 1 1t'b6, and White resigned after ten more moves.

A little later, Krasenkov ascertained that all these variations are redundant, as the simple 1 5 . . . e5 (attacking the bishop and threatening 16 . . . be 1 7 1t'xb8 1t'xc7) settles matters at once. After 1 6 l:t xf6 J.xf6 1 7 .lt.:ld5 ef 1 8 .lt.:lxf6+ �g7, White's position is completely hopeless.

148 4 i./4 System

9 1i'xcS 10 i.b3

10 ll:\bS looks like an active move, for example : 10 . . . 11t'b4+ 1 1 ll:\d2 ll:\a6 ( 1 1 . . . ll:\e4 is more precise) 1 2 a3 11t'a5 1 3 b4 11t'd8 14 i.e5 with advantage; Huzman-Tseshkovsky, Tashkent 1987. However, Black has the powerful retort 10 • • • .i.e6! (96) .

96 w

1 1 ll:\c7 (after this, Black seizes the initiative. Unclear play results from 1 1 .t.xe6 11t'xb5 1 2 .i.c4 11t'xb2 1 3 0-0; if 1 2 .i.b3, then 1 2 . . . ll:\e4) 1 1 . . . .i.xc4 1 2 b3 (in Popchev-Lalic, Bosna 1988, White opted for 1 2 ll:\d2, but after 1 2 . . . b5 1 3 b3 ll:\d5 14 ll:\xa8 ll:\c3 1 5 .l:l. xc3 i.xc3 1 6 be 11t'xc4 ! 1 7 11t'e2 .i.xd2+ 1 8 11t'xd2 ll:\c6 19 11t'e2 11t'b4+ his position was hopeless) 1 2 . . . 11t'a5 + 1 3 11t'd2 11t'xd2+ 1 4 ll:\xd2 i.d3 1 5 ll:\xa8 ll:\d5 1 6 ll:\c7 .l:l. c8 1 7 ll:\f3 ll:\xf4 1 8 ef i.b2 19 �d2 i.xcl + 20 .l:l. xc 1 i.e4 21 ll:\b5 .l:l. xcl 22 �xc1 i.xf3 23 gf ll:\c6, and Black realised his advantage in the ending; InkiovLputian, St John 1 988.

10 11t'aS 1 1 0-0 ll:\a6!?

A novel idea. The knight . heads for c5, so as to eliminate the bishop on b3 in some variations. Instead, 1 1 . . . ll:\c6 1 2 h3 i.f5 produces a position which we have already examined in Game No. 20. Gavrikov had introduced 1 1 . . . ll:\a6 a few months earlier, but in a less prominent tournament than the USSR Championship. Hence we have chosen Belyavsky-Gavrikov as the 'main' game.

12 ll:\eS White utilises the fact that the e5-point is undefended. Stationing

his knight there, he takes aim at f7. An equal game results from 12 .i.c4 ( 1 2 a3 i.g4 1 3 h3 .l:l. ad8!) 12 . . . ll:\c5 1 3 a3 i.e6 14

4 i..f4 System 149

i..xe6 ltJxe6 15 i..e5. In a later game Magerramov-Henkin, Podolsk 1989, White played 1l 1t'd4, and there followed : 12 . . . J: d8 1 3 1t'e5 ltJd5 14 1t'g5 h6 1 5 1t'h4 ltJxf4 16 ef ltJc5 1 7 i..c2 i..e6 (but not 1 7 . . . e6? 18 b4, and White wins) 1 8 f5 gf 19 b4 1t'c7? (a serious error; after 19 . . . 1t'a6! 20 be i..xc3 21 i..xf5 i..g7! 22 i..xe6 1t'xe6 23 J: fe l , the game is about equal -Magerramov. Now White obtains a strong attack) 20 be i..xc3 21 J.xf5 i..xf5 22 J: xc3 �g7 23 J: e l J: d7 24 J: ce3 J: ad8 25 ltJe5 J: d 1 26 1t'g3 + i..g6 27 h4 1t'a5 (more stubborn resistance was offered by 27 . . . J: xe 1 + 28 J: xe 1 h5 29 f4 1t'xc5+ 30 �h2 e6 3 1 ltJxg6 fg 32 J: xe6) 28 J: xd 1 J: xd 1 + 29 �h2 1t'xc5 30 J: f3 ! (30 h5? 1t'cl , and it is Black who wins) 30 . . . �h8 3 1 lllxg6+ fg 32 1t'b8+ 1 -0.

97 w

I2

I3 i..xf7 +!

ltJcS (97)

From the material viewpoint this operation benefits Black, but great complications now arise. A quieter move was 13 i..c4.

I3 J: xf7 I4 ltJxf7 �xf7 IS b4!

The consistent follow-up to White's idea. IS 1t'xb4 I6 llldS ltJxdS I7 1t'xdS+ ltJe6

Better than 17 . . . i..e6 18 1t'xc5 1t'xc5 19 J: xc5 i..xa2 20 i..e5! b6 2 1 J: c2 i..b3 22 J: c3 i..xe5 23 .l:t xb3 aS 24 f4 i..g7 25 J: xb6 a4 26 J: f2 a3 27 J: a2 i..b2, with advantage to White (Gavrikov).

IS J:c4 Wb2 (98)

150 4 j_f4 System

98 w

19 J:t fcl? The black pieces are rather awkwardly placed, which compen

sates for White's material deficit. The position may be assessed as double-edged with approximately equal chances, for example : 19 J:t c7! (threatening 20 'it'c4) 19 . . . �f8 20 J:t xc8 + J:t xc8 21 'it'xe6 J:t c l 22 g4, when the resulting situation is completely calm. 19 j_gJ was also possible. The move played allows Black to extricate himself at once.

19 j_d7! The point is that after 20 'it'xd7 J:t d8 21 'it'a4 b5!, Black wins. If

White had played 19 ..tg3, he could have answered 19 . . . ..td7 with 20 'it'xd7 J:t d8 21 J:tf4+. However, in that case Black has 1 9 . . . ..tf6, and i f 2 0 J:t fc l , only then 20 . . . ..td7!

20 h3 White loses not only after 20 'it'xd7, but also after 20 J:t 4c2

..tc6! 20 ll d8

20 . . . ..tc6 is even stronger; then 2 1 J:t xc6? be 22 J:t xc6 fails to 22 . . . 'it' b 1 + 23 �h2 'it'f5.

21 'it'aS Interestingly enough, this very position arose in the first game

that featured 1 1 . . . �a6 : Lukacs-Gavrikov, Debrecen 1988 (we come to it at last). In that game White chose a different queen move : 2 1 'it'f3, but after 2 1 . . . �e8 22 J:t 4c2 'it'b5 23 ..tg3 ..tc6 24 'it'g4 'it'f5 25 'it'b4 J:t d3 26 J:t c4 'it'd5 27 e4 'it'b5, he had to resign. Belyavsky presumably wasn't acquainted with that game; at any rate, the move he plays leads to the same result.

21 a6 22 ..tc7 J:t c8

23 .l:t 4c:Z 'irb5 24 'irxb5 ixb5 25 ib6 .l:t c6 26 .l:t xc6 ixc6

4 i/4 System 151

Black's endgame advantage is obvious, and he exploits it with no particular trouble - although it takes a full thirty moves.

27 f3 ll:lf8 28 �f:Z ll:ld7 29 ia7 ie5 30 e4 .td6 31 �e3 e5 32 �d:Z �e6 33 .te3 .te7 34 �c:Z b6 35 �b:Z �d6 36 .l:tdl + �c7 37 h4 a5 38 h5 g5 39 h6 ll:lf8 40 .td:Z ll:le6 41 .tc3 .td6 42 �at ll:ld4 43 �bl ll:le:Z 44 .tb:Z b5 45 .l:td:Z ll:lf4 46 a3 ll:le6 47 .tc3 b4

Slowly, but surely, Black accomplishes his task. 48 ab ab 49 : xd6 �xd6 50 .txb4+ �d7 51 �c:Z ll:lf4 52 g4

ll:le:Z 53 .tc5 .tb5 54 �d:Z ll:lf4 55 �e3 .te:Z 56 .tb4 .tdl 0-1

6 Fianchetto System

Game No. 22 Karpov-Timman

Candidates Final (2nd game) Kuala Lumpur 1990

The King's Fianchetto System (g2-g3, .ifl-g2) occurred twice in the London/Leningrad match and twice in Seville. The exchange in the centre (c4xd5 c6xd5) which took place in those games has recently become highly popular; it follows that this 'symmetrical variation' is worthy of close examination. In the present book I have included my three most recent games with it. Two of them were played in what may be considered the highest-calibre event in the interval between the last two contests for the world crown : my Candidates Final match with Jan Timman. In those two games I played White, yet in the third example I had the black pieces. With the large amount of supplementary material incorporated in the notes, what is offered here is the most up-to-date survey of the variation in question.

1 d4 .!0(6 2 c4 g6 3 .!0(3 .ig7 4 g3 c6

The advance . . . d7-d5, which actually defines the Griinfeld Defence, is postponed by one move; in the event of an exchange in the centre, Black is preparing to recapture on d5 with the pawn. The other popular line, 4 . . . d5 Sed .!Oxd5, will be considered in Game No. 25.

5 .ig2 d5 6 cd

On a previous occasion I decided to avoid exchanging in the centre. Here is what happened : 6 .!Oc3 0-0 7 1Wb3 e6 8 0-0 .!Obd7 9 .if4 .!Ob6?! (9 . . . b6 is more appropriate) 10 c5 .!Oc4 1 1 1Wc2

Fianchetto System 153

ltlh5 12 b3 ltlxf4 1 3 gf ltla3 14 1Wd2 b5 (on 14 . . . b6, White has 1 5 1Wb2 be 16 1Wxa3 cd 1 7 ltla4 d3 1 8 .l:l. ac1 de 19 .l:l. fe 1 , or 1 5 ltla4 b5 1 6 ltlc3, with the better chances) 1 5 .Z:. fe 1 .l:l. b8 1 6 �h 1 a5 1 7 e3 f5? ( 1 7 . . . b4 1 8 ltla4 ltlb5, or 1 7 . . . 1Wc7 followed by . . . f6 and . . . e5, was safer) 1 8 .tfl .td7 (now 1 8 . . . b4 is less good : 1 9 ltla4 ltlb5 20 ltlb6 ltlc7 21 a3) 1 9 .te2 .tf6 20 .l:l. g 1 �h8 21 .l:l. g3 1We7 22 : ag1 .Z:. g8 23 1Wcl ! b4 24 ltla4 .Z:. g7 25 1Wfl .l:l. bg8 26 1Wh3 .te8 27 ltlb6 1Wd8 28 1Wh6 .l:l. c7 29 ltle5 .l:l.cg7 30 .td3 ltlb5 31 .txb5 cb 32 f3 .th4 (he can't save himself with 32 . . . .te7 33 ltlbd7! J.xd7 34 .l:l. xg6 J.f6 35 .Z:. xg7 .l:l. xg7 36 .Z:. xg7 J.xg7 37 ltlf7+ �g8 38 1Wxg7+ ri;xg7 39 ltlxd8) 33 .l:l. h3 J.f2 34 .Z:. xg6 J.xe3 35 .Z:. xg7 .Z:. xg7 36 .l:l. g3 1We7 37 .l:l. xg7 1Wxg7 38 1Wxg7 + <J;xg7 39 c6 J.xc6 40 ltlxc6 .txf4 41 ltld7 1-0; Karpov-Kir. Georgiev, Wijk aan Zee 1988.

Notwithstanding my success in that game, I still believe that the immediate exchange on d5 promises White a more substantial initiative.

99 w

6 7 ltlc3 8 ltle5

cd o-o e6 (99)

This book was virtually completed when one of the strongest tournaments in chess history took place : Linares, 199 1 . The game Karpov-Gelfand from that event is of interest for the theory of the line we are examining. Play proceeded : 8 . . . J.f5 (a rare move which 'almost' enabled Black to equalise) 9 0-0 ltle4 10 .te3 ltlxc3 1 1 be ltlc6 12 1Wb3 e6 1 3 ltlxc6 be 14 1Wa3 .l:l. e8 1 5 J.f4 e5 16 J.xe5 .txe5 1 7 de .l:l. xe5 1 8 e3 .te4 1 9 .Z:. fd 1 1Wf6 20 .l:l. d4 h5 2 1 h3 .:tb8 22 : ad 1 J.xg2 23 <J;xg2 1We7 24 flxe7 : xe7 25 .1:1. 1d2

154 Fianchetto System

l:t b6 26 c4 de 27 l:t xc4 (for twenty moves we have been very close to a draw, but White has nonetheless extracted everything possible from the position; exchanging most of the pieces, he has obtained a superior rook ending. The remainder of the game provides a good illustration of 'exploitation technique') 27 . . . �g7 2S l:t dc2 l:t c7 29 g4 hg 30 hg �f6 3 1 �g3 �e6 32 a4 �d7 33 g5 : a6 34 l:t d4+ �eS 35 l:t c5 l:t b6 36 �f4 l:t d7 37 : xd7 �xd7 3S �e5 �e7 39 f4 : b4 40 : as : b7 4 1 e4 : c7 42 : c5 : cs 43 : c3 : eS 44 : c4 : cs 45 :b4 : c7 46 a5 �d7 47 : b3 �e7 4S a6 �d7 49 �f6 �cS 50 l:t h3 l:t d7 51 f5 gf 52 ef c5 53 : c3 : c7 54 g6 fg 55 fg �d7 56 g7 : cs 57 l:t g3 1 -0.

9 .tgS Postponing kingside castling for a while. For 9 0-0, see Game

No. 23. 9 1Wb6

9 . . . h6 weakens Black's position and wastes time. In GutmanZysk, Biel 1 9SS, White acquired a big advantage after 10 .te3! 1Wb6 1 1 1Wd2 �h7 12 0-0 tt:Jc6 1 3 : fc l .td7 14 tt:�a4 1Wc7 1 5 tOeS : adS 1 6 b4! 1Wb6 1 7 b5! tt:Je7 l S a4.

10 1Wd2 tt:Jfd7 In Nikolic-Nunn, Amsterdam 1 9SS, Black equalised after 10

••• tt:Jbd7 11 .te3 tOeS! 12 f4 t0d6. Gutman recommends 11 h3! h6 ( 1 1 . . . tOeS 12 .te7, or 1 1 . . . tL!xeS 1 2 de t0d7 1 3 .te3! and f4) 1 2 .te3, with the better chances for White.

The immediate 10 . . . t0c6 has also been seen : 1 1 tt:lxc6 be ( 1 1 . . . 1Wxc6 is also playable : 1 2 :t e l 1Wd7 1 3 0-0 b6 14 .th6 .tb7 15 : c2 : acS 16 : fc l t0e4, and Black maintains the balance; Cvitan-Zysk, West Berlin 19SS) 12 0-0 t0d7 1 3 l:t fd l : bS 14 b3 f6 (after 14 . . . cS 15 de, or 14 . . . e5 15 de, White will play 1 6 .te3 and .td4, with obvious positional gains) 15 .th6 (here too 1 5 .te3 is not bad) 1 5 . . . .txh6 1 6 1Wxh6 cS 1 7 t0a4. This position arose in Haritonov-Ivanchuk, Frunze 1 9SS. In Haritonov's opinion, after 1 7 . . . 1Wb4 or 1 7 . . . 1Wa5, White has a slight advantage. In the game, Black played 1 7 . . . 1Wd6?, which could have had severe consequences after l S e4! . Instead, White replied l S de tt:lxc5 19 tt:lxc5 1Wxc5 20 : ac l 1Wd6 21 1Wd2 f5 22 1Wd4 : f7 23 e3, and still preserved some positional advantages. In the end, Ivanchuk was unable to hold the position, and lost.

1 1 .tel I had already employed the manoeuvre .tcl -g5 against the


Dutch grandmaster four years earlier, but at this point I withdrew my knight to f3.

Karpov-Timman, Bugojno 1 986 : 1 1 ltlf3 ltlc6 12 : d l ltlf6 ( 12 . . . 1t'b4 i s more precise) 1 3 0-0 .i.d7 (and here 1 3 . . . ltle4 was better; then 14 ltlxe4 de 1 5 ltle5 ltlxe5 1 6 de .i.xe5 1 7 .i.xe4 1t'xb2 18 1t'xb2 .txb2 19 : b l gives White no more than a minimal endgame advantage) 14 .i.xf6 .txf6 1 5 e4 1t'a5 1 6 1t'f4 (at this point 16 ed ed 17 ltle5! ltlxe5 1 8 ltlxd5 1t'xd2 1 9 ltlxf6+ �g7 20 : xd2 ltlc4 2 1 ltlxd7 ltlxd2 22 : d l : fd8 23 .txb7 would have given White a large plus) 1 6 . . . .i.g7 1 7 : re t : ad8 1 8 ed ed 19 ltle5 .te6. The chances are now equal; subsequently we both made plenty of mistakes. At first I landed in a difficult position, then Timman gave me the opportunity to restore equality and acquire a decisive initiative.

1 1 ltlc6 The exchange on e5 promises nothing good : 1 1 . . . ltlxe5 1 2 de

1t'a5 1 3 f4 ltlc6 14 0-0 : d8 ( 14 . . . f6 1 5 ef .i.xf6 16 .tf2) 1 5 .tf2 .td7 16 a3! 1t'a6 1 7 b4 ltle7 1 8 : fd l , with a clear advantage to White; Nikolic-Nunn, Brussels 1988.

12 ltlxc6 be (100) Better than 1 2 . . . 1t'xc6 1 3 .i.h6 .i.xh6 14 1t'xh6 1t'd6 1 5 h4!

and Black is in a dangerous position; Shpilker-A. Kuzmin, Moscow 1986.

100 w

13 :ct In Nikolic-Korchnoi, Amsterdam 1988, White played the hasty

13 h4?!, when Black could have gained the better chances with 13 . . • : b8 14 : b 1 c5 1 5 ltla4 1t'b5 16 ltlxc5 ltlxc5 1 7 de d4. But then, the game continuation 13 . . . a5 14 h5 .i.a6 1 5 : d 1 :fb8


16 hg hg 1 7 b3 '1Vb4 1 8 �fl c5 also gave Korchnoi a plus. There followed 1 9 J:l. h4 c4 20 ..th6 ..th8 2 1 be 'lfxc4 22 ..tf3 J:l. c8 23 J:l.c l '1Vb4 24 �g2 J:l. c4 25 a3 'lfxa3 26 J:l. ch l 'lfxc3 27 'lfxc3 J:l. xc3 28 ..td2 ..txd4 ! and Black won.

13 '1Vb4! A good move. If at once 1 3 . . . J:l. b8, then 14 ll:la4 '1Vb4 15 b3

'lfxd2+ 16 �xd2, and White's chances are better, for example : 1 6 . . . J:l. b4? 1 7 J:l. xc6! ..txd4 18 ..txd4 ll xd4+ 19 �c3 J:l. g4 20 J:l. c7 etc.

14 0-0 14 a3? would be premature : 14 . . . '1Vb3 1 5 'lfc2 J:l. b8 16 ll:la4

..ta6! with the initiative. 14 J:l. b8 15 b3 c5

Ridding himself of the backward pawn while he still can. After 1 5 . . . ..ta6 16 J:l.fd l ! J:l. fc8 1 7 ll:la4, Black has difficulty freeing himself.

16 J:l. fd1 Of course not 1 6 de?, because of 1 6 . . . d4 !

16 cd 17 ..txd4 ..txd4

After 1 7 . . . e5? 1 8 ..txa7 J:l. b7 19 ..te3 d4 20 ll:ld5!, or at once 1 9 ll:lxd5!, Black would be left a pawn down.

18 'lfxd4 'lfxd4 19 J:l. xd4 ll:lb6 20 J:l. dd1

Given White's chosen plan, this is indispensable. 20 e4? fails to 20 . . . e5 ! 21 J:l.dd 1 d4, with good play for Black. If 20 e3 , then after 20 . . . ..ta6 2 1 ll:la4 ll fc8 22 ll:lc5 ..tc4! Black similarly has an easy game.

20 ..tb7 21 e4!

Seeing that White's 'heavy artillery' is better mobilised, his decision to open lines in the centre makes perfectly good sense.

21 de 22 ll:lxe4 �g7 23 J:l. c5

Invasion of the seventh rank would be illusory : 23 J:l. c7 ll:ld5 24 ll d7 ll:lb6.

23 J:l. fd8

24 : xd8 : xd8 25 lDc3 J..xg2 26 �xg2 :c8?

White's positional advantage lies in his queenside pawn majority, and Black has to play very accurately to avoid reaching a dangerous ending. The correct move here was 26 . . • ll d2!, and if 27 : c7 :cl 28 lDb5? : xc7 29 lDxc7 �f6, Black is not at all worse. A more precise line for White is 28 a4! a5 29 lDb5 : xc7 30 lDxc7 �f6 31 lDa6 lDd5 32 lDc5 lDb4! 33 f4 (33 �f3 �e5!) 33 . . . lDc6 34 �f3 �e7 35 �e4 �d6 36 lDb7+ �c7 37 lDc5 (37 lDxa5 f5 + !) 37 . . . �d6, with a draw.

101 w

27 : xeS lDxc8 (101)

In this situation White could have utilised his active king position and obtained substantial winning chances by playing 28 ..tf3! In view of the threat to penetrate to the queenside, Black's most natural course is: 28 . . . f5 29 ..te3 e5, and now White has the very strong 30 lDb5!, preventing the stabilising manoeuvre . . . lDc8-d6. It is by no means simple for Black to defend. The variations I shall now quote were indicated by Holmov. 30 . . • ..tf6 31 <.Pd3 <.PgS! (Black's best chance is to activate his king. He now threatens to advance by the route g4-h3 or g4-f3, so White has to take prophylactic measures; but this will lead to a weakening of his kingside, giving Black further possibilities for counterplay. Instead, the pawn endgame after 3 1 . . . <.Pe6 3 2 <.Pc4 lDd6+ 3 3 lDxd6 <.Pxd6 3 4 b4! i s obviously lost for Black, and White can also count on success after 32 . . . g5 33 a4) 32 h3! h5 33 <.Pc4 h4 34 ..td5 hg! 35 fg f4! (after 35 . . . e4 36 <.PeS! Black's defence would be incomparably more difficult) 36 gf+ <.Pxf4! 37 lDc3 (the straightforward 37 ..tc6 e4 38 <.Pb7?


10d6+! would even lose). Now Black has to play very precisely : (a) 37 • • • 10b6+? 38 ¢>c6! (38 ¢>e6? e4 39 ¢>f6 e3! 40 ¢>xg6? ¢>f3 4 1 h4 10d5! etc.) 38 . . . e4 (38 . . . �g3 39 ¢>b7 ¢>xh3 40 ¢>xa7 ll:ld7 4 1 b4 g5 42 a4 g4 43 a5 g3 44 �b7 g2 45 ll:le2, and White wins) 39 ¢>b7 g3 40 ¢>xa7 ll:ld7 41 b4! �f3 42 a4 e2 43 ll:lxe2 ¢>xe2 44 aS ¢>d3 45 ¢>b7 ¢>c4 46 a6 �bS 47 a7 ll:lb6 48 h4! and wins. (b) 37 . . . gS! enables Black to hold the position : 38 ¢>c6 ¢>e3! 39 ¢>b7 ltld6+ 40 ¢>xa7 �d2 41 ltld5 e4 ! (not 4 1 . . . ltlb5+ 42 ¢>b6 ltlc3 43 ¢>c5! e4 44 ¢>d4! ltlxd5 45 ¢>xe4 ltlc3 + 46 ¢>f5 and White wins but 43 . . . ltlxa2! draws-G. Flear) 42 �b6! (less precise is 42 a4? ¢>c2 43 b4 �b3 44 a5 �c4 45 ll:le3 + ¢>xb4 46 �b6 ltlc4+ with a draw) 42 . . . e3 43 ltlxe3 ¢>xe3 44 ¢>c6 .!ticS 45 ¢>d7 ltlb6?! (45 . . . ltla7? is inferior: 46 a4 �d4 47 ¢>e6 �c3 48 ¢>f5 ! �xb3 49 �xgS ¢>xa4 50 h4 ltlc6 5 1 ¢>f6! and wins; or 47 . . . ¢>e4 48 b4 ! ltlc6 49 bS ltla5 50 b6 �f4 5 1 ¢>d5 ¢>g3 52 ¢>c5 ¢>xh3 53 �bS, again winning for White) 46 ¢>e6 �e4! 47 a4 ltld5 48 a5 ltlb4 49 ¢>d6 ¢>d4!, or 49 ¢>f6 ¢>f4, with a draw in either case.

These variations show that although Black achieves a satisfactory result, it is only with immense difficulty.

28 f4? This throws away White's advantage, since the exchange of a

pair of kingside pawns increases Black's chances of a peaceful outcome.

28 fS! 29 ltla4

Having let my winning chances slip, I commit a further inaccuracy. 29 �f3 was simpler.

29 ¢>f6 30 .!ticS

Preventing 30 . . . e5 on account of 3 1 ltld7+. 30 ltlb6 31 ltld3 ltld7 32 ltlb4

I should have brought the king forward : 32 ¢>f3 e5 33 fe+ ltlxe5+ 34 ltlxe5 ¢>xe5 35 h4! with a draw.

32 eS 33 ltlc6 a6 34 fe+ ltlxe5 35 ltld4


The pawn ending would be lost for White. 35 �e7 36 �n

36 �f2 is answered by 36 . . . etlg4+. 36 �d6 37 �e2 �dS 38 etlcl �e4 39 a4 .!Of3 40 b4 etld4+

In the event of 40 . . . .!Oxh2!, White could maintain the balance with 41 b5 ab 42 ab �d5 43 .!Od4! .!Og4 ! 44 b6 �d6 45 .!Oe6! �c6 46 .!Of8 .!Of6 47 �f3 �xb6 48 �f4 �c6 49 �g5 .!Oe4+ (or 49 . . . tt!h5 50 �h4 �d5 51 .!Oxh7) 50 �f4 �d6 51 lt!xh7 �e7 52 g4; alternatively 42 . . . lt!g4 43 b6 tOeS 44 b7 lt!d7 45 ltlb4 lObS 46 .!Od3! and again White draws, as Black cannot win without the use of his knight.

41 lt!xd4 �xd4 42 bS

t-t After 42 . . . ab 43 ab �c5 44 �e3 �xb5 45 �f4 �c4 46

�g5 �d3 47 �h6 �e3 48 h4! �f3 49 �xh7 �xg3 50 �xg6, both sides' resources are completely exhausted.

Game No. 23 Karpov-Timman

Candidates Final (4th game) Kuala Lumpur 1990 1 d4 lt!f6 2 c4 g6 3 lt!f3 J.g7 4 g3 c6 s J.g2 dS 6 cd cd 7 .!Oc3 o-o 8 tOeS e6 9 o-o

In our discussion of this game, we shall scrutinise the fine points of the opening play arising from White's 9th move. The middle game will proceed at a slow pace, yet the ending (after the adjournment) will be exceptionally interesting, and although this does not strictly


belong to the subject of the book, we shall give a large number of entertaining endgame variations.

9 lt)fd7 The best move. After 9 . . . lt)c6 10 li)xc6 be 1 1 �a4 lt)d7 1 2

.tf4 White has a clear positional advantage; AkhmilovskayaChiburdanidze, Women's World Championship match 1986.

102 B

10 f4 (102)

Karpov-Kasparov, 3rd game, London 1 986, continued differently : 1 0 lt)f3 (the exchange 10 lt)xd7 .txd7 gives White nothing : 1 1 e3 lt)c6 1 2 b3 "fle7 1 3 .tb2 .C. fc8 t-t ; Portisch-Nunn, Budapest 1 987) 10 . . . �c6 1 1 .tf4 lt)f6 1 2 lt)e5 .td7 (after 1 2 . . . li)xe5 1 3 .txe5, White's chances are to be preferred) 1 3 "fld2 li)xe5 1 4 .txe5 .tc6 1 5 .l:lfd1 ( 1 5 .C. ac1 , with the prospect of f2-f3 and e2--e4, would have preserved the initiative) 1 5 . . . lt)d7 1 6 .txg7 q;xg7 1 7 .C. ac1 �f6 1 8 "flf4 "fibS 19 "flxb8 .C. axb8 20 f3 .C.fd8 2 1 q;f2 .C. bc8 22 e 3 lt)e8 23 .C. d2 lt)d6 24 .l:l dc2 �f8 2 5 .t n �e7 26 .td3 f5 27 h4 h6 28 b3 (in his notes to the game, Kasparov states that after 28 g4 Black would be in quite a dangerous position. But it seems to me that the continuation 28 . . . .C. f8 29 g5 �e4 + 30 .txe4 fe 3 1 f4 h5, or 30 fe fe+ 31 q;e2 ed + 32 q;xd3 hg 33 hg .C. f5 ! , promises White nothing) 28 . . . g5 2� lt)e2 .td7 30 .C. c5 b6 3 1 .C. c7 .C. xc7 32 .C. xc7 .C. a8 33 lt)g1 �e8 34 .C. c 1 .C.c8 35 .C. xc8 t-t .

10 lt)c6 In game 1 3 of the 1 986 match, my opponent drove tbe knight

back with 10 . . . f6, but after 11 lt)fJ �c6 12 .te3 �b6 l 3 .tf2 f5 14 li)e5 .td7 1 5 "fld2 li)c8 16 "fle3 q;hs 1 7 .C.fd 1 ( 1 7 .C.fc l was


stronger) White generated considerable pressure - although the game was eventually drawn after blunders by both sides.

White can retain the initiative by retreating his knight to d3. The game Hulak-H. Olafsson, Wijk aan Zee 1987, is instructive : ( 10 . . . f6) 1 1 o!Lld3 o!Llc6 1 2 e3! (more precise than 1 2 .i.e3 o!Llb6 1 3 b3 .i.d7 14 o!Llc5 J:l b8 1 5 "it'd2 f5 16 J:l fc 1 .!Llc8 1 7 .i.f2 o!Lld6, with equality; Nikolic-Nunn, Linares 1988) 1 2 . . . f5 (Andersson-Nunn, Brussels 1 988, is worth mentioning : 12 . . . o!Llb6 1 3 b3 ..td7 14 ..ta3 J:l e8 1 5 "it'd2 o!Lle7 16 o!Llc5 J:l b8 1 7 J:l fe 1 f5 1 8 .i.fl o!Llec8 19 J:l ac1 .i.c6 20 o!Lld3 o!Lld6 21 o!Lle5 ..tf8 22 "it'b2! J:le7 23 J:l c2 J:l c7 24 J:l ec1 J:l bc8, and now instead of 25 o!Llb1 , which led to approximate equality, White could have played 25 .i.c5 ! , when according to Nunn the threat of 26 "it'a3 gives him a noticeable plus) 1 3 .i.d2 (a good alternative is 1 3 o!Lle5 o!Lle7 14 b3 o!Llf6 1 5 ..ta3 ..td7 16 J:l c 1 J:l e8 1 7 J:l f2 o!Lle4 1 8 J:l fc2 o!Llc6 19 .i.fl a 6 20 o!Llxe4 fe 21 ..td6 .:l c8 22 h4 o!Llxe5 23 de "it'aS 24 hS; Dzhindzhikhashvili-Mestel, Reykjavik 1990) 1 3 . . . o!Llf6 14 J:tc l ..td7 1 5 o!Lle5 J:l e8 1 6 h 3 o!Llxe5 1 7 de o!Lle4 1 8 o!Llxe4 de 1 9 "it'b3 .i.c6 20 ..tb4, with advantage to White.

Shortly after the London/Leningrad match, Kasparov played this variation himself with White, against Nunn (Brussels 1 986). The exchange on e5 unexpectedly led to a crushing defeat for Black : 10 . . . o!Llxe5 1 1 fe o!Llc6 12 e4! de 1 3 ..te3 (103)

103 8

With Black at the moment a pawn up and not having made any noticeable mistake, it is hard to imagine that he will resign after six ( ! ) more moves. 1 3 . . . f5 14 ef J:l xf6 (the play up to here was still on familiar lines. For example, this position had arisen by transposition in Sveshnikov-Mikhalchishin, Lvov 1983; after


1 S .J:I. xf6 .txf6 16 ll:le2 ll:lb4, White's advantage evaporated. Kasparov doesn't waste time exchanging rooks) 1 S ll:lxe4 .J:I. xfl + 16 1Wxf1 ll:lxd4? 1 7 .J:I. d 1 eS 1 8 lOgS, and Black stopped the clock. The following beautiful variation is possible : 1 8 . . . 1We7 1 9 .tdS+ .te6 20 l:l xd4 ed 2 1 .txe6+ *h8 22 lllf7+ *g8 23 /Od8+ *h8 24 .tgS! 1Wb4 2S /Of7+ *g8 26 ll:leS + *h8 27 /Oxg6+ hg 28 1Wh3 + . Incidentally, all this actually happened in a correspondence game Hjorth-M. Andersson ( 1986), and at this point Black resigned.

Instead of 1 6 . . . ll:lxd4, Black might play 1 6 . . . .txd4 1 7 .txd4 /Oxd4 1 8 l:l e l eS 1 9 1Wf6. Kasparov judged this position to be in White's favour, in view of 19 . . . 1Wxf6 20 lllxf6+ *g7 2 1 /Oe8+ *f8 2 2 .J:I. xeS with the better ending, but Gutman has suggested 19 . . . 1!fb6 20 *hl .td7 21 1!fxe5 l:l f8 22 lOgS .tc6. However, after 23 .tdS+ ! (G. Flear) Black can resign!

12 e4, then, seems quite strong but White can also play the solid 1 2 .te3!?. This position arose in Karpov-Timman, Amsterdam 1986, though by a different move-order ( 10 . . . /Oc6 1 1 .te3 ll:lxeS 12 fe). Let us look at a few moves of that important encounter.

12 .te3 f6 13 ef .J:I.x/6 Better than 1 3 . . . .txf6 1 4 1Wd2 .td7 1 S *h1 .J:I.f7 1 6 .tg1

.te8 1 7 l:l ad 1 .tg7 1 8 l:l xf7 .txf7 1 9 e4, with the advantage; Makarov-Glek, Moscow 1 986.

14 1Wd2 .td7 15 *h1 l:l xf1 + 16 l:lxfl 1We7 (104)

104 w

This is a safer post for the queen than aS ( 1 6 . . . 1WaS 1 7 a3 l:lf8 1 8 l:l xf8+ .txf8 1 9 .tg1 ) or b6 ( 16 . . . 1Wb6 17 .tg1 /Oxd4? 1 8 l:l f4).


1 7 '4 d1 Over-protecting the d-pawn and preparing ..te3-g1 and

e2-e4. But White can also withdraw his bishop to g1 at once. Ribli-Nunn, Dortmund 1987, went 1 7 ..tg1 '4 d8 18 a3 �h8? ( 1 8 . . . ..tc8 1 9 ..te3 '4 f8 i s safer) 1 9 e4! (the fiasco against Kasparov has not put Nunn off this variation, but once again the advance of the centre pawn puts him in a difficult position) 19 . . . de 20 llJxe4 ..tc8 2 1 ..te3 (21 llJg5 ..tf6) 2 1 . . . .l:l£8 22 : xf8+ 1Fxf8 23 b4, with a clear advantage.

17 . . . '4 c8 In Karpov-Chiburdanidze, Bilbao 1 987, the Women's World

Champion played 17 . . . �h8 18 a3, and only then 1 8 . . . : c8. After 19 .*.g5 1Ff8 20 :n 1Fg8 21 e3 h6 22 ..tf6 .*.xf6 23 : xf6 :f8 24 :f2 1Fg7 25 : xf8 + 1Fxf8 26 e4 de 27 lt:lxe4 b6 28 �g1 'ikg7 29 d5 ed 30 1Fxd5 1Fd4+ 31 1Fxd4+ llJxd4, the game had turned from an opening into an ending, with White retaining a minimal edge. But Maya defended precisely, and obtained a draw in spite of all my efforts.

18 a3! White's intention is to retreat his bishop to g1 and seize the

centre. But in reply to the immediate 1 8 ..tg1 , Timman had prepared 18 . . . 1Fb4 ! . Then after 19 e4 de 20 .*.xe4 ..te8, Black can secure his position with . . . .*.e8-f7 (2 1 d5? ..txc3!). If instead 20 llJxe4 1Fxd2 2 1 : xd2, Black has the tactical device 2 1 . . . llJxd4 22 : xd4 (22 ..txd4 .*.xd4 23 : xd4 : c 1 + 24 ..tfl ..t b5) 22 . . . ..txd4 23 ..txd4 :c1 + 24 ..tgl .*.c6, with counterplay in a complex endgame.

18 . . . ..t/6 White easily refutes 1 8 . . . llJa5 with 1 9 llJxd5 ed 20 .*.g5 llJc4

21 ..txe7 llJxd2 22 .*.xd5+ �h8 25 : xd2. 19 .*.g1 ..tg5? A more precise move was 19 . . . 1Fg7, hindering e2-e4 for the

moment, on account of 20 e4 de 21 llJxe4 ..txd4 22 .*.xd4 llJxd4 23 1Fxd4 1Fxd4 24 : xd4 : c 1 + 25 .*.fl ..tb5 ! .

20 1Ve1 llJd8 21 e4 de 22 1Vxe4 b6 23 d5 Black could not prevent this breakthrough, and has to go over

to wholly passive defence. The results of the opening can now be summed up as miserable for Black. The game, however, lasted another 35 moves and ended in a draw.

1 1 .*.e3


At this point the exchange on e5 leads to variations we have looked at already ( 1 1 . . . �dxe5 12 fe f6 etc.). In the main game we are examining, Black drives back the centralised knight and follows with . . . �b6. I should mention that on two occasions in Seville, Kasparov played 1 1 . . . �b6 at once, and equalised easily in both cases. We shall thoroughly discuss the immediate knight move in the notes to Game No. 24.

1 1 f6 12 �d3 �b6 13 b3 'fle7

1 3 . . . ..td7 would transpose into the Nikolic-Nunn game that we have already seen. So it may be said that we are only now breaking new ground.

14 15 16 17 18 19

a4 .tel e3 ..ta3 l:tcl ..txf8

..td7 l:t fd8 ..te8 'flf7 ..tf8 'fixf8 (105)

105 w

As the result of some complex manoeuvres, a certain advantage for White has emerged on the queenside.

20 g4 'fie7 21 'fld2 l:t ac8 22 �e2 l:tc7 23 l:t c5 �c8 24 f5 gS

White is slightly better after 24 . . . �d6 25 �df4 (25 fe? �e4 26 ..txe4 de 27 �df4 'fixc5, and Black wins) 25 . . . gf 26 gf �xf5 27 �xd5 ed 28 : xf5.


25 �g3 2S fe 1i'xe6 26 �c3 (26 �g3 �g6 27 �fS �8e7) 26 . . . �8e7 27

e4 is less clear. Now the threat of bringing the knight to f5 is more dangerous :

2S . . . .i.f7 26 fe 11'xe6 (26 . . . �xe6 27 �h5 .l:l. f8 28 �b4! ) 27 �f5. To avoid this, Black sacrifices a pawn.

106 w

25 e5!? ( 106)

26 Wet An inaccuracy, committed after much thought. The sheer variety

of possibilities put me in rather a quandary, and I decided to decline the pawn sacrifice although Black's position would almost have become critical if I had calmly accepted it. After 26 �xd5+ (26 .l:l. xd5 ed 27 ed .i.f7! 28 .l:l. xd8 + Wxd8 gives Black counterplay for the pawn) 26 . . . �h8, White has a wide choice : (a) 27 �xc6 J.xc6 28 1Wc1 ed 29 e4 ! , or (b) 27 .l:l. fcl ed 28 e4 ! . This strategic operation involving the return of the pawn and the inevitable break with e4-e5 secures White a considerable plus. (c) 27 de �xeS 28 .l:l. xc7 1Wxc7 29 .l:l. c 1 'fle7 30 �xeS 1WxeS 3 1 1Wd4! is also good. (d) 27 1Wcl .l:l. xd5 28 .l:l. xdS b6 (28 . . . �b4 29 l:t cS!) 29 de �b4 30 �xb4 .l:l. xcl 3 1 .l:l. xc l , or 27 . . . b6 28 .l:l. xc6 J.xc6 29 J.xc6 ed (29 . . . .l:l. d6 30 dS .l:l. xdS 3 1 .i.xdS) 30 ed .l:l. xd4 (30 . . . .l:l. d6 31 dS .l:l. xdS 32 .l:l. e 1 !) 3 1 .l:l.e 1 1Wd6 (3 1 . . . .l:l. xd3 32 .l:l. xe7 �xe7 33 Wet �xc6 34 1We8+ �g7 3S �hS+ �h6 36 1Wf8+ and mates) 32 .l:l. e8 + �g7 33 l:te6 - with a very strong initiative for White in either case. (e) 27 .i.g2 (in some of the foregoing variations the white king's


bishop departs from the board, but it is also possible to bring it home before anything else) 27 . . . ed 28 e4 !, followed by e4-e5 (if necessary supported by a rook - : fe l ). If 28 . . . lt:le5, then 29 : xeS! fe 30 f6 and 3 1 'illxg5 is crushing.

The queen move to c1 allows the centre to be blocked; if l had been playing for that, I should have continued 26 : fe l l e4 27 lt:lb4! - possession of the c-file is not without significance. The trouble is that I had 'too many irons in the fire' (trying to control the c-file and f-file at once). Now Black nullifies the danger in the centre and obtains a fully viable game. The ensuing phase, as I have said already, is a trifle dull; the most interesting events occur only after the adjournment.

26 b6 27 : c2 e4 28 lilf2 lt:ld6 29 1i'd2 .l:dc8 30 l:tfcl aS

Black also stands quite well after 30 . . . lt:la5 3 1 : xc7 : xc7 32 : c3 'illd8 33 'illc2 : xc3 34 'illxc3 'illc8!

31 -*.fl If White hadn't made the superfluous queen moves, this bishop

manoeuvre would now be extremely dangerous for Black. The attempt to build up with 3 1 'illc3 would be countered by

3 1 . . . -*.d7 and 32 . . . lt:le8. 31 lt:lb4

3 1 . . . lt:lxd4! looked inviting, but after 32 ed e3 33 -.d3 ef+ 34 �xf2 : xc2+ 35 : xc2 lt:le4+ 36 �f3, the draw is not far away (the exchange of the good knight on c6 for the bad one on f2 is not advantageous to Black). The match situation compelled Timman to maintain the tension in the hope of seizing the initiative.

32 : c3 -.d7 33 lild1 :c6 34 : xc6 : xc6 35 : xc6 -.xc6 36 lilc3 �f8 37 �f2 �e7 38 �e1 �f8 39 �d1 -.cs 40 �el �g7


Under pressure from the clock, we have both been marking time with our kings a little, and I should mention that Black missed quite a good opportunity : 40 . . . hS! 41 gh (41 h3 with equality) 4 1 . . . lllxfS 42 .J..h3 .J..d7!

41 llla2 lllxa2 42 11rxa2 11rc7

A possibility was 42 . . . •c3 + 43 �f2 bS 44 ab lllxbS (but not 44 . . . .i.xbS 45 lllhS+ �f7 46 .i.xbS lllxbS 47 b4!).

43 �f2 �f8 44 1i'b2 �e7 45 .tel �d8 46 �el �c8

46 . . . bS!? is sharper. 47 �d2 �b7 48 •ct 11re7

For drawing purposes, 48 . . . 11rxc1 + was simpler.

107 B

49 �el .i.d7 50 �f2 IlleS 51 11Vhl ! (107)

I didn't feel like settling for a draw, and decided to provoke mind-bending complications by abandoning my queenside to its fate while going ahead on the kingside. Obviously there could be no question of calculating the variations to the end.

51 11Vb4 52 h4! 11Vxb3 53 hg fg 54 11rxh7 •xa4 55 11re7 11rc6


56 1hg5 a4 57 'flle7 'flld6 58 'fllxd6 ltlxd6 59 .t.dl .tb5?!

It was hardly a good idea to block the path of his pawn. After 59 . . . b5, it is most likely that the game would soon have been drawn; at any rate Black would not be risking anything. Here are some sample variations : 59 . . . b5 60 f6 (60 ltle2? b4 61 ltlc1 �c7) 60 . . . .txg4 61 .txg4 ! a3 62 .te6 �c6 63 ltle2 (not 63 f7 ltlxf7 64 ltlxe4 ltlg5! 65 it:lxg5 a2, and Black wins) 63 . . . a2 (63 . . . b4 64 ltlcl) 64 .t.xd5 + �xd5 65 ltlc3 + �c4 66 ltlxa2 b4 (nor does he gain anything from 66 . . . �b3 67 ltlc1 + �b2 68 ltle2 b4 69 d5 b3 70 �g3 �c2 71 ltld4+ �c3 72 ltlxb3 �xb3 73 �f4 �c4 74 �e5) 67 ltlxb4 �xb4 68 �g3 �c4 69 �f4 �d3 (69 · . . . �d5 70 �g5 �e6 7 1 �g6) 70 �e5 ltlf7+ 71 �e6 ltlg5 +, with a draw.

108 8

60 ltle2 a3 61 ltlcl ( 108)

In this position Timman sealed his move after nearly half an hour's thought. Many commentators were misled by Black's queenside activity, and in the local newspapers the following day there were even headlines like "Can Karpov survive?" In actual fact, adjournment analysis revealed that the only winning chances lie with White, whose pawns are much more dangerous than his opponent's. All the same, we shall see that the rightful outcome of this tense struggle would have been a draw.

61 �c7 A good sealed move, but of course we had analysed others too.


Let us look at the main variations discovered by Zaitsev and myself: (a) 61 • • • -*.d7 62 -*.b3! �c6 63 �g3 �c4 64 �f4 �b2 (threatening the deadly . . . �d3+) 65 �g5! �c4 66 -*.xc4 (given White's active king, the creation of another black passed pawn is not dangerous) 66 . . . de 67 �f6 b5 6S �a2 b4 69 �xb4 �b5 70 �a2 �a4 71 �c1 a2 72 �xa2 �b3 73 �c1 + �c2 74 �e7 �xc1 75 f6 c3 76 f7 -*.xg4 77 fS(11t') c2 7S 11t'f4 -*.f3 79 d5 �d2 SO d6 cl (11t') S 1 d7, and Black is defenceless. (b) 61 . . • ..te8 62 �a2 (after the careless 62 �g3, the black knight performs the same nimble triangulation that we have seen before : 62 . . . �c4 ! 63 �f4 �b2, and wherever the bishop goes, Black plays 64 . . . �d3 +) 62 . . . �c6 63 -*.b3 �c4 64 g5 �d6 65 f6 -*.f7 66 �g3 �c4 67 �f4 �d6 6S �c3 ..tg6 69 -*.xc4 de 70 �b5 + �c6 7 1 �xa3 �d5 (7 1 . . . b5? 72 �e5 c3 73 �c2 -*.f7 74 d5+ ! -*.xd5 75 g6) 72 �b5 �c6 73 �c3 b5 74 �e5 b4 75 d5+ �d7 76 �a2 b3 77 �c3 b2 7S �d4, with a decisive advantage. (c) 61 • • • �c6 62 �g3 �c4 (if 62 . . . -*.c4 63 �f4 �d7 64 g5 a2 65 �xa2 -*.xa2 66 g6, Black is mated after either 66 . . . �eS 67 -*.a4+ �e7 6S -*.xeS �xeS 69 �g5 b5 70 �h6 b4 7 1 g7 �f7 72 �h7 b3 73 gS(11t')+, or 66 . . . �e7 67 g7 �f7 6S f6 -*.c4 69 -*.h5+ �gS 70 -*.g4 -*.b5 71 -*.e6+ �f7 72 �f5 -*.eS 73 �g6 b5 74 -*.xd5) 63 -*.e2 �d6 (if 63 . . . �xe3?, then 64 -*.xb5+ �xb5 65 f6, and the pawn goes on to queen. Other variations are also unacceptable : 63 . . . �d6 64 -*.xb5 + �xb5 65 �f4 �b4 66 g5 �c3 67 g6 �e8 68 �e5 �b2 69 g7 �xg7 70 f6 �xcl 7 1 fg a2 72 gS(11t') a 1 (11t') 73 11t'g1 + �b2 74 11t'xa1 + �xa 1 7 5 �xd5, o r 6 3 . . . �d6 64 -*.xb5 + �xb5 6 5 �a2! �d6 66 �f4 �c7 67 g5 �e8 68 g6 �f6 69 �g5 �e7 70 �b4 b5 7 1 �c6 + ! �d6 7 2 �xf6 a 2 73 g 7 a1 (11t') 74 gS(11t') �xc6 75 11t'e6+) 64 �f4 -*.a4 65 g5 �b2 66 g6 �d3 + 67 �xd3! ed 68 ..txd3 �e7 69 g7 �f7 70 f6 �g8 (70 . . . a2 7 1 -*.h7 a 1 (11t') 72 g8(11t')+ �xf6 73 11t'f8+ �e6 73 -*.f5 mate) 71 �g5 a2 72 �h6 a1 (11t'), and a familiar mate follows : 73 -*.h7+ �f7 74 g8(11t')+ �xf6 75 11t'f8+ etc.

Now let us see what happened in the game. 62 �g3 �c4

62 . . . �d7? loses in a way which we have seen before : 63 �f4 -*.c4 64 g5 a2 65 �xa2 ..txa2 66 g6 �e7 (66 . . . �e8 67 -*.a4+)

1 70 Fianchetto System

67 g7 �f7 68 f6 .tc4 69 .tg4 �g8 70 .te6+ t;)f7 7 1 �f5 .tb5 72 �g6 .te8 73 .txd5 b5 74 .ta2 b4 75 .tb3.

63 .te2 But not 63 �f4? t;)b2! threatening . . . t;)d3 + .

63 .te8! The strongest defence. 63 . . . li)xe3 is inadequate : 64 .txb5 t;)c2

65 g5 t;)xd4 66 f6 �d6 67 g6 �e6 68 g7 �f7 69 .te8 + �g8 70 �f4, and wins.

64 �f4 Podgayets revealed a striking variation culminating in a

problem-like mate : 64 g5 t;)xe3 65 �f4 t;)c2 66 �e5 t;)b4 67 �f6 �d6 68 g6 t;)c6 69 ..tb5 e3 70 g7 t;)e7 7 1 ..txe8! e2 72 li)xe2 a2 73 t;)c3! t;)g8+ 74 �f7 t;)h6+ 75 �f8 a l (V) 76 t;)b5 mate (109) .

109 B

It would indeed be pleasant to finish the game like this, but unfortunately Black has a defence : 64 . . . �d6 65 ..txc4 de 66 t;)a2 ..ta4 67 t;)c3 .tb3 68 t;)b5 + �e7 69 li)xa3 c3 70 �f4 b5! 71 �xe4 b4 72 �d3 ba 73 �xc3 .td5 with a draw.

64 t;)b2 In my adjournment analysis I considered 64 . . . b5 to be more

precise : 65 t;)a2 t;)b2 (but not 65 . . . �d6 66 g5 t;)b2 67 g6 b4 68 �g5! b3 69 t;)cl .ta4 70 g7 a2 71 g8(¥) a l (V), as White obtains an irresistible attack with 72 Vb8 + �d7 73 f6! Vxc l 74 ..tg4+ �c6 75 Vc8+) 66 f6 t;)d3 + 67 �f5 b4 68 .td l . But the move played is not yet the cause of a catastrophe, either.

65 �g5 The sole winning attempt. It is true that after 65 t;)a2, the

following long variation gives White good prospects : 65 . . . �d6

Fianchetto System 1 71

66 �gS �e7 (66 . . . lilc4 is a mistake in view of 67 ..txc4 de 68 �f6 ..ta4 69 �g7 ..tb3 70 f6 c3 7 1 lilxc3 a2 72 lilxa2 ..txa2 73 gS) 67 �h6 �f6 68 lilb4 ..tf7 (inadequate alternatives are 68 . . . ll:ld3 69 lilxdS+ �f7 70 gS a2 7 1 g6+ �f8 72 g7+ �f7 73 ..thS+ �g8 74 llle7 mate, and 68 . . . ..tc6 69 ..tfl ! lildl 70 gS + �f7 -70 . . . �xf5 7 1 ..th3 mate, is pretty - 7 1 g6+ etc.) 69 ..tfl ! lild l (it is too late for counterplay with 69 . . . lild3 70 gS+ �e7 7 1 lilc2 a 2 7 2 �g7) 7 0 g5+ �e7 7 1 �g7 lilxe3 7 2 f6+ �e6 73 ..th3+ lilfS+ 74 ..txfS + �xf5 75 �xf7 e3 76 �g7! e2 77 ll:lc2 a2 78 f7 e 1 (1i') 79 lilxe l a 1 (1i') 80 f8(1i')+. However, after the correct 65 . . . b5! , nothing can be found for White.

65 lild3 65 . . . lilc4 comes too late : 66 �f6 lilxe3 67 �e7.

66 lilb3!? al 67 ll:lal bS 68 ..tdl b4 (IJO)

Bringing the king a little closer to the pawns does not work : 68 . . . �d6? 69 ..tb3 lilcl 70 �h6 b4 7 1 f6 �e6 72 �g7 lilxb3 73 lilxb3 ..ta4 74 lilcS + ! �d6 75 f7 a 1 (1i') 76 f8(1i')+ �c6 77 1i'c8 + �d6 78 'ifd8+ �c6 79 1i'd7+ �b6 80 1i'b7+ �aS 8 1 1i'a6 mate.

1 10 w

69 -*.b3 lilcl After 69 . . . �d6 70 �f6, the g-pawn cannot be stopped.

70 ..txdS �d6 71 ..tc4 ..tbS

Not 41 . . . �e7 42 f6+ �f8 43 �h6 ..tf7 44 dS b3 75 d6!, or 7 1 . . . ..ta4 72 f6 b3 73 �h6! b2 74 f7 b1 (1i') 75 f8 ('if)+ �d7 76 1Wf5 + �c7 77 1i'a5 + , but here 74 . . . ba (11') leads to a completely unclear position - G. Flear.

An interesting possibility is 71 . . . b3 72 ..txb3 lilxb3 73 lilxb3


�d5 74 �a1 �c4 75 .tf6 �c3 76 �e7 �b2 77 �xeS �xa1 78 f6 �b2 79 f7 a 1 (W) 80 f8(W) Wa4+ 81 �f7 �b3+ . We had reached this position in our analysis, and Zaitsev suggested 82 .tf6 Wxe3 83 �e5 .tc2 84 Wf4 .td3 85 d5 with chances of success. But in the post-mortem, Timman pointed out the more effective 82 d5! Wxe3 (82 . . . Wxd5+ 83 'Oftg6 with a technically won ending) 83 Wb4+ ..ttc2 84 g5 Wf3 + 85 �g7 e3 86 Wc4+ �b2 87 Wb5 + �c1 88 d6 e2 89 Wc5+ �b2 90 Wb6+ �c1 91 d7 e l (W) 92 d8(W) Wec3 + 93 Wdf6 etc. According to computer analyses, this ending is a win for White.

72 ..tg8 �e7 73 �h6

After 73 f6+ �f8 74 ..te6 b3 (74 . . . ..ta4 leads to the same result as in the game) 75 .i.xb3 �xb3 76 �xb3 ..tc4 77 �at �f7 78 �f5, White wins. However, Black has the saving move 74 . . . .i.d7!, and if 75 ..txd7 (it is more sensible to settle for a draw) 75 . . . b3 76 �f5 b2 77 g5 �e2! 78 g6 �g3 + ! 79 �g5 �h5! 80 �xh5 ba(W) 8 1 g7+ �f7 82 .i.e8 + (82 ..te6+ �xf6 83 g8(W) 'ifh 1 +) 82 . . . ..ttxf6, it is Black who unexpectedly wins.

73 �f8? It is only now that Timman commits the decisive error. He

could have drawn with 73 . . . b3! 74 .i.xb3 �xb3 75 �xb3 .i.c4 76 �al �f6!, as pointed out by his second, the Hungarian grandmaster Sax.

74 ..te6! .i.d7 75 gS b3 76 g6

1-0 Black resigned rather than be mated : 76 . . . ..txe6 77 fe b2 78

g7+ rJ;e7 (78 . . . �g8 79 e7 �f7 80 g8(W)+ �xe7 8 1 Wg5+ rJ;f7 82 Wd5+ rJ;f6 83 'ifc6+) 79 g8(W) ba(W) 80 Wf7+ �d6 8 1 Wd7 mate.

After this victory, it became clear that another duel for the world crown between Kasparov and myself was inevitable.

Game No. 24 Wojtkiewicz-Karpov

Haninge 1990 I have played the White side of the Griinfeld so often, especially

the g2-g3 system, that when the symmetrical variation arose in


the present game, it was with some interest that I handled it for the other colour.

1 lt:lf3 lt:lf6 2 c4 g6 3 g3 .i.g7 4 .i.gl 0-0 5 d4 c6

I have rarely played either the King's Indian or the Griinfeld with Black. But given the choice between them, I prefer the latter.

1 1 1 w

6 lt:lc3 d5 7 cd cd 8 lt:le5 e6 9 0-0 lt:lfd7

10 f4 lt:lc6 1 1 .i.e3 lt:lb6 (1 11)

Kasparov had played this knight move against me in the first and third match game in Seville, and now I decided to use it myself. An alternative of equal value is 1 1 . . . f6 (see Game No. 23). In the present game, you might say that I combined the two ideas.

Before proceeding with the main game, in which White played 1 2 b3, let us look at the two encounters from Seville, where I preferred 1 2 .i.f2. The first match game went as follows :

12 j_f2 .i.d7 13 e4 l0e7 14 l0xd7 In the game Drasko-Nikolic, Vrnjacka Banja 1987, where it

appears this position occurred for the first time, White opted for 14 ed lt:lbxd5 1 5 lt:lxd5 lt:lxd5 1 6 Wb3 .i.c6 1 7 : acl 1Wa5 1 8 .:l. c5 'it'a6, and a draw was agreed.


Another possibilty is 14 a4! de 15 aS .!Dbd5 16 .!Dxe4 .:t b8 17 'iVb3 ..te8 18 .l:tfcl .!Dc6 1 9 'iVa3 .!Dcb4. Nikolic-Hulak, Zagreb 1987, continued 20 .:t c4? .!Da6 2 1 .!Dd6 .!Dac7 22 .:t ac l .!DbS 23 .!DxbS ..txbS 24 : cS ..te8 2S b4 b6, and White achieved nothing; but Gutman's suggestion 20 .!Dc3! gives him the advantage. At move 1 7, an alternative is 17 . . . ..tc6 18 .!DeS .!Dc7 19 ..txc6! with the initiative; this is stronger than 18 .:t fcl a6 ( 1 8 . . . .!Dc7! at once is more precise) 19 .:t c4 ( 1 9 .!DeS! is correct) 19 . . . .!Dc7 20 .!Dc3 ..txg2 21 �xg2 .!bedS, with equality; Andersson-Hulak, Wijk aan Zee 1 987.

14 . . . 'iVxd7 15 e5 .:l.fc8 16 :t el ..t/8 17 ..tj3 .:l. c7 An accurate move which equalises. I would have answered the

incautious 1 7 . . . a6 with 1 8 ..te2 .:tc7 1 9 g4 ! ..th6 20 'iVd2, and if 20 . . . .!Dc4, then 2 1 ..txc4 : xc4 22 .!De4 ! . But now this idea doesn't work : 18 ..te2 : ac8 19 g4 ..th6 20 'iVd2 .!Dc4 21 ..txc4 .:t xc4, and both black rooks are on the c-file.

18 b3 : ac8 19 'iVd2 lbc6 20 'iVb2 a6 21 ..te2 'iVe7! 22 l1Jb1 liJb4 23 lbc3 lbc6 24 l:lJb1 lbb4 25 .:l. c5 t:tJd7 26 : xc7 : xc7 27 lbc3 l:lJc6 28 l:lJb1 t:tJb4 29 l:lJc3 lbc6 30 lbb1 i-J.

In the third game, Kasparov varied with 12 . . . CfJe7! (a refinement on 12 . . . ..td7 of the first game) 13 a4 aS 14 'iVb3 (Black would have had more problems after 14 e4 de l S ..txe4 .!DbdS 16 'iVb3. Greenfeld-Birnboim, Tel-Aviv 1988, continued 16 . . . f6 17 .!Dc4 �h8 1 8 .:t fe l CfJb4 19 .:t ad 1 .!bedS 20 .!De3 : a6 2 1 ..txdS ed 22 .!DexdS ..tg4 23 .:t d2 .:t e6 24 .!Dxb4 .:t xe l + 2S ..txel ab 26 'iVxb4 : e8 27 ..tf2, and White gained the advantage) 14 . . . ..td7 1 S .:t fc 1 ..tc6 1 6 .!DbS .!Dbc8! 1 7 e 3 .!Dd6 1 8 .!Dxd6 'iVxd6 1 9 .tel .:t fb8 20 ..tfl f6 2 1 t:tJf3 'iVd7 22 'iVc2 .!DfS 23 ..td2 .!Dd6 24 b3 .:t c8 2S 'iVdl h6 26 .te l gS 27 : a2 'iVe8 28 : ac2 .U8 29 ..td3 g4 !-! .

1 2 b3 This move was played for the first time in Portisch-Korchnoi,

Reggio Emilia 1 987/8. It seems to me that if White is seeking the initiative, it is worth trying 1 2 .!Dxc6 be 1 3 ..tf2.

12 ..td7 13 'iVd2 f6

In Portisch-Korchnoi, Black first played 1 3 . . . .:te8 1 4 .:t fcl , and only then 14 . . . f6. After l S .!Dd3 .:te7 16 �h1 ..te8, the players agreed a draw. Evidently on that day they were just not in the mood for a fight; the result has nothing to do with the


position. In the present game, I carry out a regrouping in the centre as Black, without wasting time on rook manoeuvres.

JJ2 B

14 IL!d3 ILleS 15 .l:t acl IL!d6 16 -*.f2 (1 12)

The position is characterised by so-called dynamic equilibrium. The pawn chains are closed, and the forces cannot clash for some time yet. Some precise and none-too-obvious manipulation of the pieces is called for, and of course such operations are not to everyone's taste. But I personally have always liked this kind of game, based as it is on a wealth of subtle points. I had previously encountered such situations with White, but was also ready to try them with Black, especially since the configuration of pieces is nearly symmetrical.

Here, by the way, is an illustration of the subtleties inherent in the position. Black just needs to play one careless move, 16 . . . ..te8?, and White immediately seizes the initiative : 1 7 ILleS! -*.f7 ( 1 7 . . . "W/e7 18 IL!xe6!) 18 e4 de 19 ILJ3xe4 IL!xe4 20 -*.xe4 "W/e7 21 IL!xb7!

16 17 IL!e5 18 �h1 19 g4!?

f5 "W/e7 .l:t fc8

1 9 IL!a4 or 19 a3 would have preserved equality, but White is trying for more.

19 IL!e4 20 IL!xe4 fe 21 IL!xd7


The knight cannot hold out on eS for long, but perhaps it was worth exchanging it for the black one : 2 1 .l:. c3 �xeS 22 de .l:. xc3 23 'ii'xc3 .l:. c8 24 'ii'd2 b6 2S .tg3, intending f4-fS .

2 1 .tg3!? at once is also interesting : 2 1 . . . �xeS 22 de .l:. xc l 23 .l:. xc l .l:. c8 24 .l:. xc8+ .txc8 2S 'ii'c3 (White's aim i s to carry out f4-fS, but the situation does not yet call for it : 2S fS? ef 26 gf gf 27 'ii'xdS + ..te6, or 2S 'ii'd4 b6 26 fS ef 27 gf gf 28 'ii'xdS+ J.e6 29 'ii'a8+ �f7, with a good game for Black in either case) 2S . . . 'ii'd7 (but not 2S . . . 'ii'd8 26 .th4!, or 2S . . . .td7 26 'ii'c7) 26 e3 b6 27 a4 .tb7, and again the chances are about equal.

l /3 B

21 'ii'xd7 22 a3 (1 13)

One thoughtless move, made on general grounds - and already Black has a chance to seize the initiative. 22 e3 was correct, avoiding weakening the queenside.

22 aS! 23 .th3 'ii'e7

Forcing the rook to abandon the c-file. l4 .l:.al �d8

Black could have prevented the f4-fS break by playing 24 . . . .th6 2S gS .tg7. After 26 .l:.fcl , followed by e2-e3 and .th3-fl , it would be hard for either side to play for a win. Instead, I attempt to bring my knight over to the kingside, where the decisive events will shortly take place.

25 fS gS The sharp 2S . . . �f7!? is also interesting.

26 .tg3 �c6

Fianchetto System 1' 77

At this point I convinced myself that after 26 . . . lllf7 27 fe'! 1Wxe6 28 : f5 h6 29 : an llld6 30 ..te5! lllxf5 3 1 gf 1Wd7 32 ..txg7 1Wxg7 33 f6 the initiative is with White, so I decided to bring the knight back again.

27 e3 .:t d8 28 fe 1Wxe6 29 .:. rs h6 30 1We2

30 .:. afl looks more consistent, yet White cannot successfully exploit his domination of the f-file : 30 . . . llle7 3 1 ..tc7 .l: dc8! (but not 30 . . . .l: e8 3 1 ..te5 lllxf5 32 gf 1Wa6 33 ..txg7 �xg7 34 f6+, with an attack for the exchange), and now 32 ..te5 is not so effective : 32 . . . lllxf5 33 gf 1Wa6! 34 ..txg7 (34 f5 .:t f8!) 34 . . . �xg7 35 f6+ �h8! 36 1Wf2 .:t c7 37 ..td7 .l: d6 38 f7 .:t f8 39 1Wf5 �g7 40 1We5 + .:t f6 4 1 ..tb5 1Wd6, and White's attack is spent. On the other hand after 32 ..txa5 Black takes the exchange, and his chances must be preferable.

30 31 .:.n 32 : act

.:t xc6 .:t bl

33 34

llle7 .:tdc8 .:tc6 1Wxc6

Rather a passive move. It was worth considering 34 .:t f2!?, and if 34 • . . 'ii'cl + 35 ..tfl 1Wxa3, then after 36 1Wb5 1Wb4 (36 . . . a4? 37 1Wxb7 ab 38 ..td6) 37 1Wd7 White has a dangerous initiative for the pawn. Instead, 34 . . • a4! leads to unclear play : 35 ba (better than 35 b4 1Wc1 + 36 'ii'fl 1Wxa3 37 ..td6 lllg6 38 1Wb5 'ii'c l + 39 ..tfl 1Wc6 with advantage to Black, or 36 ..tfl 1Wxa3 37 1Wb5 1Wxe3 38 1Wxb7 .:t f8 etc.) 35 . . . .:t xa4 36 1Wb2.

34 ..tf8 35 ..tn lllc8 36 1Wb2 1Wd7 37 .tel ..td6 38 .tel

The advantage of the bishop pair is illusory - it was better to eliminate the bishop on d6. After 38 :n (more precise than 38 ..txd6 lllxd6) 38 . . . ..txg3 39 hg, it would be difficult to breach White's fortress.

38 llle7 39 .:. cl : f8!


40 �g2 Not 40 J.xa5?, on account of 40 . . . l: f2. But 40 �gl , followed

by J.e2-d l , was more tenacious. 40 ll:lg6 41 b4 a4 42 .-c2 .-f7!

The threats on the f-file are becoming formidable. 43 .-dl

The black a-pawn is immune : 43 .-xa4 ll:lf4+ ! 44 ef .-xf4 45 J.g3 .-xc l 46 J.xd6 .-el l 47 J.xf8 .-xe2+ 48 �gl e3 49 h3 .-f2+ 50 �h l e2, and wins.

43 �h8! Preparing to open the g-file.

44 l:c3 (114) 44 �gl would likewise be answered by 44 . . . h5! , intending to

take the g-pawn; after 45 gh .-f5 ! 46 hg (46 �g2 ll:lh4+ 47 J.xh4 gh is crushing) 46 . . . .-h3 47 J.g3 J.xg3 48 hg .-xg3 + 49 �hl l: f2 50 l: c8 + �g7, i t i s all over.

Nor does 44 l: c2 help : 44 . . . h5! 45 gh ll:le7 (45 . . . ll:lh4+ is premature because of 46 J.xh4 gh 47 J.g4) 46 h3 (46 l: c3 g4 ! and . . . .-xh5) 46 . . . ll:lf5 47 l: c3 ll:lh4+ 48 ..txh4 gh 49 .-el (49 .-gl l: g8 + 50 .tg4 .-r3 mate) 49 . . . .tg3 50 .-n .-f2+, and Black wins.

1 14 B

44 45 gh 46 J.xh4

hS! tt:Jh4+!

I f 46 �gl (46 �hl .-fl +), then 46 . . . .-fs, followed by . . . .-f5-h3, is decisive.

46 gh


47 •et h3+ ! 48 �hl

Other continuations are less stubborn : 48 �g1 •g7+ , or 48 �xh3 : g8 49 ..tg4 : xg4 50 �xg4 •B + 51 �g5 ..te7+ 52 �h6 •f6 mate.

48 49 •x£2 50 : cl 51 ��· 52 �hl 53 ��·

55 ..td3 is met by 53 . . . : f3.

•£2 : xfl : xhl+ : gl+ : fl

53 ..tg3 54 �hl �h7 55 ��· �h6 56 �hl : hl+ 57 ��· : £2 58 �hl ..td6 59 ��· : gl+ 60 �fl : g8 61 :ci : g3

6 1 . . . ..th2 was immediately decisive; so was 6 1 . . . ..txb4 62 ab h2 63 �f2 : g 1 64 : xg 1 hg(•)+ 65 �xgl a3.

The game was now adjourned. Realising that Black's threats were unanswerable, White resigned without playing on.

0-1

Game No. 25 Vaganian-Kasparov

Barcelona 1989 1 10£3 10£6 2 c4 g6 3 g3 ..tg7 4 ..tg2 0-0 5 10c3 d5 6 cd 10xd5

The reader will recall that in the World Championship matches of 1 986-7, Kasparov preferred the symmetrical system, in which Black prepares . . . d7-d5 with . . . c7-c6 and recaptures on d5 with

the pawn. In a game we played later (and which I shall examine further on), he chose . . . <'i)xd5. In the last two or three years, Kasparov has played a number of quite interesting games with this line, and the one I have selected here as the 'principal' game is among the most fascinating of all. In the notes, as usual, I shall also mention other important examples from recent practice.

1 15 w

7 0-0 <'i)c6 8 d4 <'i)b6 (1 15)

Here White has the choice between the slow e2-e3, as in the actual game, and the more committal d4-d5 and e2-e4. Let us first consider the latter.

9 d5 Another possibility, seen more rarely, is 9 ..tf4. 9 . . . li)a5 10 e4 The modest 10 'ii'c2 quickly leads to equality; here are two

examples : Tukmakov-Halifman, Simferopol 1 988, went 10 . . . c6 1 1 de

<'i)xc6 12 .l:l. d1 ..tf5 1 3 e4 ..td7 14 -*.f4 .l:l. c8 1 5 'ii'e2 'ii'e8 1 6 h3 -*.e6 17 <'i)d5 ..txd5 18 ed <'i)b4 19 <'i)e1 'ii'd7 20 'ii'd2 <'i)a6 !-! .

Plachetka-Smejkal, Trnava 1 989, varied with 12 . . . 'ii'c7 1 3 <'i)b5 'ii'b8 14 ..tf4 e 5 1 5 ..te3 ..tf5 1 6 'ii' c5 .l:l. c8 1 7 <'i)xa7 <'i)a4 1 8 <'i)xc6 be 1 9 'ii'e7 t-t.

In Nikolic-Kasparov, Skelleftea 1 989, White played instead 10 ..tf4, and there followed 1 0 . . . c6 1 1 de 'ii'xd 1 1 2 .l:l. axd 1 <'i)xc6 1 3 <'i)b5 ..tg4! 14 b3 e5! 1 5 ..te3 e4 ! 1 6 <'i)fd4 lDxd4 1 7 <'i)xd4 <'i)d5 1 8 h3 <'i)xe3 19 fe ..td7. Black's position is the more pleasant, though the game was eventually drawn.


10 . . . c6 11 i.g5 An inadequate alternative is 1 1 .J:I. e 1 tL!ac4 ! 1 2 1We2 i.g4 l 3 de

( 1 3 ..tg5 h6) l 3 . . . be 14 h3 ..t xf3 1 5 i.xf3; Yurtayev-Belov, Podolsk 1989. At this point, 1 5 . . . 1Wd6 (instead of 1 5 . . . aS), with the idea of 16 . . . 1Wc5 or 16 . . . 1W b4, would have given Black a good game.

11 . . . h6 Black failed to justify 1 1 . . . i.g4 in Hansen-Kasparov, Thessa

loniki 01 1 988 : 1 2 h3 ..txf3 l 3 1W xf3 cd 14 tL!xd5 tL!xdS 1 5 .J:I. ad 1 1W c7 1 6 ed ..txb2 ( 1 6 . . . tL!c4 1 7 1We2 .J:I. fe8 1 8 .J:I. c 1 b 5 1 9 b3 "iie5 20 "ii xe5 tL!xe5 21 d6, and Black is in a bad way) 17 .J:I. fe 1 .J:I. fe8 18 l: xe7! .J:I. xe7 1 9 d6 .J:I. e 1 + 20 .J:I. xe 1 "ii xd6 21 .J:I. e7 .J:I. f8 22 "iid5, and although Black did manage to save himself, his position does not inspire much confidence.

12 i./4 cd 13 ed 13 tL!xd5 is strongly answered by 13 . • • tL!ac4! . In Pastircak

Hort, Czechoslovakia 1 984, White continued with the straightforward 14 tLJc7?, and ended up in a difficult position after 14 . . . .1:1. b8 1 5 "iixd8 .J:I. xd8 1 6 .J:I. ad 1 ( 1 6 tL!a6 e5!) 1 6 . . . ..tg4 1 7 tL!d5 e5 18 tL!xb6 .J:I. xd 1 19 .J:I. xd 1 tL!xb2 ! 20 .J:I. d2 ef. Instead 14 1W b3 might appear dangerous for Black, but with 1 4 . . . e5! 1 5 tL! xb6 "ii xb6! 16 ..txh6 ( 1 6 "ii xc4 ef) 1 6 . . . i.xh6 17 1W xc4 1W xb2 1 8 .J:I. ab 1 1Wa3 1 9 tL!xe5 i.e6, he obtained the bishop pair and plenty of chances for the pawn in Shpilker-Krasenkov, Moscow 1 987.

13 . . . tL!ac4 14 1We2 g5 If 1 4 . . . tL!xb2, then 1 5 tL!e5 ! . 15 i.cl e6! 16 de Better 1 6 .1:1. d 1 , although 1 6 . . . ed 1 'l tL! xd5 tL!xd5 1 8 "ii xc4 tLJ b6!

1 9 "ii b3 Wf6 gives Black slightly the better game (Popovic). 16 . . . ..txe6 1 7 .J:I. d1 1Wc8 18 tL!d4 A sounder line is 1 8 h4 ..tg4 1 9 hg. 18 . . . i.g4! 19 i./3 ..th3! 20 tL!d5 tL!xd5 21 ..txd5 .J:I. d8 22

tL!c6 be 23 ..txc4 W/5 24 ..td3 We5 25 1Wxe5 ..txe5 Black has a clear endgame advantage, and went on to win in

Hjartarson-Popovic, Belgrade 1 989. 9 e3 e5

9 . . . .J:I. e8 is also seen. Kasparov, for example, played it twice against Portisch in 1 988. Both games continued 10 d5 tL!a5 1 1 tL!d4 i.d7. I n Saloniki, Portisch played 1 2 "iie2, which led to equality after 1 2 . . . c6 1 3 de tL!xc6 1 4 tL!xc6 i. xc6 1 5 ..txc6 be


1 6 l:l. d 1 'ii'c8 1 7 .i.d2 'ii'e6 1 8 .i.e1 aS. In Reykjavik, he innovated with 12 b4?!, but after 12 . . . �ac4 1 3 a4 aS 14 bS 'ii'c8 1 S l:l. e 1 .i.h3 16 .i.h1 'ii'g4 17 �ce2 l:l. ad8 1 8 'ii'b3 'ii'd7 1 9 �f4 .i.g4 20 l:l. a2 'ii'd6, Black had some initiative in the complexities. All the same, this game too ended in a draw.

In Ljubojevic-Kasparov, Barcelona 1 989, White introduced a different novelty, though he moved the same pawn : 12 b3. However, after five more moves - 1 2 . . . cS 1 3 de �xc6 14 �xc6 .i.xc6 1 S .i.xc6 'ii'xd 1 1 6 l:txd 1 be 1 7 .i.d2 a S - a draw, once again, was agreed.

Finally, another game by Kasparov went 12 e4 cS 1 3 de lt:)xc6 14 ltlxc6 .i.xc6 1 S 'ii'b3 'ii'd3 16 .i.e3 : ac8 17 l:l. fd 1 'ii'a6 18 .i.d4 'ii'aS, with equality; Hjartarson-Kasparov, Reykjavik 1989.

10 dS �as Practice has shown that retreating with 10 . . . �e7 is too passive.

1 1 e4 c6 12 .i.gS

At this point we shall digress to examine the game KarpovKasparov, Amsterdam 1988, in which the same position arose, only with the rooks on the e-file instead of the f-file. The order of the opening moves was 1 d4 �f6 2 c4 g6 3 �f3 .i.g7 4 g3 dS S cd �xdS 6 .i.g2 �b6 7 �c3 �c6 8 e3 0-0 9 0-0 l:l. e8 10 : e l e S 1 1 d S �aS 1 2 e4 c6 1 3 .i.gS (1 16) .

116 B

Play continued 1 3 . . . f6 14 .i.e3 �ac4 ! 1 S de �xe3 (Black is willing to give up a small amount of material rather than accept defects in his pawn structure. At the same time he activates his forces) 16 'ii'xd8 : xd8 17 cb .i.xb7 (if 1 7 . . . �xg2, then 1 8 ba('ii' ) �xe1 19 'ii'xa7 destroys all Black's illusions) 1 8 : xe3 .i.h6 19


:ee l lt)c4 20 : ad 1 �f8 2 1 h4 : ac8 22 .th3 (White appears to have seized the initiative, but this is purely temporary) 22 . . . : xd l 2 3 : xd 1 lt)xb2 24 : d7 : xc3 2 5 : xb7 (it was worth testing Black by first playing 25 : d8 + �e7 26 : d7+ �f8; the point is that 26 . . . we8 is risky, since the king needs to guard g7 as a retreat square for the bishop) 25 . . . lt)c4 (the opposite bishop ending after 25 . . . : xf3 26 : xb2 is unpleasant for Black) 26 lt)h2 (not 26 : c7?, in view of 26 . . . : c 1 + 27 .tfl lt)d2 28 : xc 1 lt)xf3+ 29 �g2 lt)xh4+ 30 gh .txc 1 , and now it is White who has to hold an opposite bishop ending a pawn down. On the other hand, a draw results from 26 .te6 lt)d6) 26 . . . lt)d6 27 : xh7 (or 27 : xa7 : xg3 + ! 28 fg .te3 + with equality) 27 . . . .tg7 28 h5 gh 29 .C. xh5 ll c l + 30 �g2 : c2 31 .te6 lt)xe4 32 lt)g4 : d2 33 .tb3 (the last chance was 33 �f3 lt)g5+ 34 �e3 : b2 35 .tb3, but 33 . . . lt)c5 34 .tc4 maintains the balance) 33 . . . a5 34 :f5 lt)d6 35 ll h5 lt)e4 36 : f5 lt)d6 37 :h5 lt)e4 t-t .

12 f6 13 .te3 cd 14 ed

After 14 .txb6 'it'xb6 1 5 lt)xd5 'it'd8 1 6 b4 lt)c6, the chances are approximately equal .

14 .tg4 Black's plans involve the advance . . . f6-f5, which may be carried

out with the bishop on either g4 or c8. Mikhalchishin-Gavrikov, Budapest 1 989, went 14 . . . : f7 1 5

.tc5 f5 1 6 'it'cl ! ( 1 6 lt)d2 e4 17 lt)b3 li)ac4 1 8 'it'e2 lt)e5 1 9 : ad 1 lt)bc4 2 0 lt)d4 b6 2 1 lt)c6 'it'e8 led t o equality i n Gligoric-Gavrikov, Moscow 1989) 16 . . . e4 1 7 lt)g5 : c7 1 8 b4 lt)bc4 (other continuations are worse, for example 1 8 . . . lt)xd5 19 lt)xd5 'it'xd5 20 .C. d 1 , with a won position) 19 ll d 1 .txc3 20 'it'xc3 'it'xg5 2 1 .te3! (peace was quickly concluded in Mikhalchishin-Ftacnik, Palma de Mallorca 1 989 : 2 1 .td4 .td7 22 ba 'it'e7! 23 .tb6 : cc8 24 .tc7 lt)d6 25 .txd6 'it' xd6 26 'it'd4 t-1. On the other hand, 2 1 'it'd4? b 6 22 d6 : xc5 23 be .te6 i s good for Black) 2 1 . . . 'it'd8! (better than 21 . . . lbxe3 22 'it'xc7 lbxd 1 23 : xd 1 ) 22 ba lbxe3 23 'it'xe3 'it'd6 24 f3 ef 25 .txf3 .td7 26 : ac t (better 26 : ab l . Now the game is level) 26 . . . : xc 1 27 : xc1 b6 28 'it'f4 'it'xf4 29 gf : c8 30 : xeS+ .txc8 3 1 ab ab 32 �f2 �f7 33 �e3 we7 34 �d4 �d6 t-t .

1 5 .tc5


Practice has also seen 1 S h3 .txf3 1 6 .txf3 fS 1 7 .te2, with complex play.

1 1 7 B

15 l U7 16 b3 fS 17 .tb4 (11 7)

A well-known position. Black normally used to play 1 7 . . . 'l:lc8, whereupon White has the strong move 18 'lrel . This occurred, incidentally, in Vaganian-Thorsteins, Copenhagen 1 988. The idea is that Black no longer has the important tempo-gaining move . . . 1rd8-b6 (it would be met by 'l:lc3-a4), his knight on aS is indirectly attacked, and the position is very difficult for him. After 1 8 . . . .txf3 19 .txf3 'l:ld6 20 .te2 b6 (better 20 . . . e4) 2 1 .ta6, White gained the advantage and won.

But Kasparov has a remarkable surprise in store for his opponent.

17 'l:lac4! An unexpected piece sacrifice, although admittedly the material

is quickly regained. Kasparov had, of course, prepared this novelty in his pre-game analysis, knowing that Vaganian regularly plays this variation.

18 be e4 White's pawn-couple looks strong, but Black has ample

resources. A notable point is that as long as the black knight was on aS, the white bishop on b4 was splendidly placed, both attacking and defending; but once the knight has sacrificed itself, the bishop is constantly exposed to the threat of . . . a7-aS. Vaganian himself commented on this paradox of the instant transformation of this bishop.


19 J:cl I f 1 9 cS, Black has the very strong counter-blow 1 9 . . . a S !

19 1i'd7?! The situation is extremely tense, and Black fails to choose the

correct path. Evidently Kasparov underestimated his opponent's retort. After the game, he demonstrated the following variations : 19 • • • .!Oxc4 20 .!Oxe4 fe 2 1 J: xc4 .txf3 22 1i'd2 11' d7, with equality; or 19 . . . ef 20 .txf3 .txf3 2 1 11' xf32 .!Oxc4. The latter, however, was played in Ionov-Urban, Katowice 1 99 1 , and after 22 lObS .!OeS 23 1i'b3 aS 24 .td6 a4 25 "ife3 f4 26 gf .!Og4 27 11'e6 11'h4 28 f5 1i'gS 29 fg .!O e S + 30 � h l .!Oxg6 3 1 J: c7 White neutralised his opponent's initiative and went on to win.

20 .!Ob1! ef 21 .txf3 .txf3 22 11'xf3 f4 23 g4 11'a4 24 1i'b3

This time it is White who plays inaccurately. After 24 .tc3! .txc3 2S .!O xc3 11' xc4 26 J: fd 1 J: d8 27 d6, he could feel fairly confident.

24 1i'd7 The players are committing errors alternately. Black should

have played 24 . . . f3, with a good game for the pawn; White's king is badly placed, and his pieces are rather passive.

/ 18 w

25 f3 h5 ( 118)

26 h3 J: e8 27 1i'd3


It was not worth chasing after a second pawn. After 27 .i.c5! 'IJJ/c7 28 .i.f2 .l::t e3 29 'IJJ/d l , White would have set up a solid protective wall while keeping a material advantage. Now the mutual turmoil subsides, and the game quickly heads towards a peaceful result.

27 • • • bg l8 bg .l::t e3! 29 'IJJ/xg6 .l::t f6 30 'IJJ/gS 'IJJ/f7 31 .tc3 .l::t g6 32 'IJJ/d8+ .l::t e8 33 'IJJ/h4 .i.xc3 34 .l::t xc3 'IJJ/g7 35 .!::t el 'IJJ/d4+ 36 ��2 .l::t xel 37 'IJJ/xel �xc4 38 'IJJ/e8+ �g7 39 'IJJ/e7+ �b6 40 'IJJ/h4+ �g7 41 'IJJ/e7+ �b6 42 'IJJ/fB+ �b7 43 'IJJ/f7+ �b6 t-t

Index of Variations

1 d4 lL!f6 2 c4 g6

Seville Variation 3 lL!c3 4 cd 5 e4 6 be 7 .tc4 8 lL!e2 9 .te3

10 0-0 lO . . . ed 1 1

d5 lL!xd5 lL!xc3 .tg7 c5 lL!c6 o-o .tg4

I I f3 lL!a5 12 .txf7+ .l:txf7 13 fg l:t xfl + 14 �xfl 'ii'd6

1 4 . . . ed 1 5 ed 'ife8 15 1 5 . . . 'ii'd7 34 1 5 . . . 'ii'b6 35 1 5 . . . e5 1 6 de 37

1 6 .:te l 37 16 d5 37

1 4 . . . 'Wd7 1 5 de 13 1 5 g5 'ii'e6 13 1 5 . . . .:t d8 14 15 h3 14

1 4 . . . 1Ve8 15 15 e5

1 5 'ifa4 28 1 5 �g1 .:t d8 29

188 1 ndex of Variations

1 5 . . . 'it'e6 16 'it'd3 cd 1 7 cd J:l. d8 1 8 g5 29 1 8 h3 30

16 . . . 'it'c4 17 'it'd2 30 1 7 'it'xc4 30

15 'it'd5 1 5 . . . 'it'e6 15

16 ..tf2 1 6 ltlf4 16

16 J:l. £8 16 . . . J:l. d8 1 7 ltlf4 19

17 1i'e 1 19 17 'it' a4 ltlc6 20 1 7 . . . ltlc4 1 8 ltlf4 20

1 8 g5 20 1 1 . . . : r8 21 17 . . . b6 18 "ifc2 J:l. c8 21

18 . . . 'it' c4 22 1 8 . . . : rs 19 ..tg1 22

1 7 'it'c2 J:l. c8 25 1 7 . . . 'it'c4 1 8 g5 25

1 8 1i'e4 25 1 8 'it'b2 25

17 g5 1 7 �g1 16

17 'it'f7 1 7 . . . 'it'e4 17

18 'it'el h6 1 8 . . . 'it'f5 17 1 8 . . . ltlc4 17

19 gh 18

Exchange Variation with 7 ..tc4 - other systems 3 ltlc3 d5 4 cd ltlxd5 5 e4 ltlxc3 6 be ..tg7 7 ..tc4 c5 8 ltle2 ltlc6 9 ..te3 0-0

10 o-o

Index of Variations 189

1 0 .l:[ b 1 46

1 0 : c 1 ..td7 46

10 . . . cd 1 1 cd 'ifa5 + 1 2 �fl .l:[ d8 47 1 2 . . . 'ifa3 47 1 2 . . . ..td7 1 3 h4 h5 48

1 1 f3 12 ..td3 13 cd 14 :cl 15 'ifa4 16 d5 17 'ii' b4

1 7 'ifa3 b5 42 17

1 7 . . . b5 42 1 7 . . . b6 42

18 : fdl 18 de 43 1 8 d6 43

1 8 �c3 43

1 8 �d4 44

18

�aS cd ..te6 ..txa2 ..te6 ..td7

e6

ed 44

Modern Exchange Variation 3 �c3 d5 4 cd �xd5 5 e4 �xc3 6 be ..tg7

6 . . . c5 83

7 �f3

7 ..te3 c5 8 'ii'd2 cd 84 8 . . . 'ifa5 90 8 . . . 0-0 9 d5 91

9 : c 1 91

9 �f3 cd 93

1 3 . . . e5 48 1 3 . . . : ac8 48 1 3 . . . : res 50

9 . . . 'if aS 1 0 l:t c 1 93 9 . . . ..tg4 1 0 �g5 104

190 1 ndex of Variations

7 c5 8 : bt 0-0 9 ..te2 cd

9 . . . ..tg4 53 9 . . . 'ilfa5 10 l:t b5 53

1 0 0-0 'ilfxc3 53 10 . . . 'ilfxa2 1 1 ..tg5 53

9 . . . b6 10 ..te3 57 10 0-0 ..tb7 1 1 W'd3 57

9 . . . tt:Jc6 10 ..te3 61 10 d5 61

10 cd 'ilfa5+ 1 1 ..td2

1 1 'i!f d2 67 1 1 W'xa2 12 0-0 ..tg4

1 2 . . . lZJc6 68 1 2 . . . b6 1 3 'ilf c l ..tb7 68

1 3 . . . 'ilfe6 1 4 ..tc4 69 1 2 . . . tt:Jd7 1 3 n a t 71

1 3 ..tb4 71 1 2 . . . W'e6 1 3 W' c2 77

13 ..tg5 79

Russian System

3 tt:Jc3 d5 4 tt:Jf3 ..tg7 5 'ilfb3 de 6 'ilfxc4 0-0 7 e4 lZJa6

7 . . . ..tg4 8 tt:Jg5 108 8 tt:Je5 108 8 ..te3 tt:Jfd7 9 : d l 109

8 ..te2 8 W'b3 1 15 8 ..tg5 115 8 ..tf4 115

8 c5 9 d5

9 W' b3 lZJb6 1 1 1

9 e5 1 16 9 de 1 16

9 e6 10 0-0

1 0 .tg5 1 1 6 10 ed 1 1 ed .tf5

1 1 . . . b5 1 1 7 1 1 . . . b6 1 1 7 1 1 . . . .l:t e8 1 2 .l:t d 1 1 18

1 2 .tf4 b6 121 12 . . . .tf5 - see below

12 .tf4 1 2 l: d 1 l: e8 1 18 1 2 .tg5 132 1 2 .te3 132

12 : es 1 2 . . . 'i!Vb6 126 1 2 . . . tt:Jd7 129

13 l:t adl tt:Je4 1 3 . . . 1Vb6 121 13 . . . tt:Jd7 121

14 .td3 14 tt:Jb5 121

14 .txc3 15 be 125

4 .tf4 System 3 tt:Jc3 d5 4 .tf4

4 tt:Jf3 .tg7 5 .tf4 e5 6 de 135

4 .tg7 5 e3 c5 6 de

6 .txb8 135 6 'i!Va5 7 : ct

7 cd 136 7 'i!Va4 136

Index of Variations 191

19 2 1 ndex of Variations

7 tt:lf3 0-0 8 J:t c l de 9 .txc4 �xc5 10 .tb3 �a5 1 1 0-Q tt:lc6 -see below

7 de 7 . . . tt:le4 144

8 .txc4 0-0 9 tt:lf3

9 tt:le2 147 9 �xc5

10 .tb3 10 tt:lb5 148

10 tt:lc6 10 . . . �a5 1 1 0-0 tt:la6 148

1 1 0-:0 �a5 12 h3 138

Fianc�etto System 3 tt:lf3 .tg7

4 g3 c6 4 . . . d5 5 cd tt:lxd5 6 tbc3 0-0 7 0-0 tt:lc6 8 d4 tt:lb6 9 d5 180

5 .tg2 d5 6 cd

6 tt:lc3 0-0 7 �b3 152 6 cd 7 ltlc3 0-0 8 tt:le5 e6

8 . . . .tf5 153 9 0 - 0

9 .tg5 154 9 tt:lfd7

9 . . . �c6 160 10 f4

10 tt:lf3 160 10

10 . . . f6 160 10 . . . tt:lxe5 161

1 1 .te3 1 1 . . . f6 164

tt:lc6

tt:lb6 1 73

9 e3 ll e8 181 9 . . . e5 182

BFAfiNC HIE (,R0NFH D

''Si nce his surprise choice of the GrOnfeld Defence in the 1 986 World Championship match, Gary Kasparov has regularly adopted the GrOnfeld in tournaments and matches around the world. Against Karpov, however, this opening has proved an unfortunate choice a,nd Karpov has won more World Championship games against Kasparov in this opening than in any other. Highlights of this book i nclude:

• Original winning ideas against the GrOnfeld • Unique i nsights i nto Karpov's World Championship

preparation • Detailed coverage of key fashionable variations ,, • Fascinating tips from a former World Champion

No ambitious player can afford to take on or play the GrOnfeld Defence without this book!

Anatoly Karpov is one of the greatest players of all time. During his time as World Champion between 1 975 and 1 985, his tournament record was unprecedented as he outclassed his contemporaries time and again. His fierce rivalry with Gary Kasparov has now endured for five matches for the world title and numerous tournaments around the world.

a me in Action

:>en Game in

rUI U t;UI I l plete l iSt Of theSe and other Botsford chess books please write to:

B.T. Batsford Ltd 4 Fitzhardinge Street London Wl H OAH

-'=-PUBLISHER'sJ� \ net prlca ;

,l

ISBN 0 -7 1

1 11 1 9 7 8 0 7 1 3 "4 0 4 0 � '1 "

beating the griinfeld - actualités du club - 3eme open de...

Documents