The Queen's Gambit Accepted E-Book

Chris Ward

• Herausgeber: Batsford
• Kategorie: Lebensstil
• Sprache: Englisch
• Veröffentlichungsjahr: 2021

Seitenzahl: 323

Beschreibung

The Queen's Gambit Accepted (QGA) is Black's most straightforward answer to the popular Queen's Gambit. By accepting the gambit pawn, Black refuses to submit to the strategic bear-hug which is so common in the Queen's Gambit Declines. The QGA is fundamentally sound, but soon leads to sharp play where both sides have plenty of chances to play for a win.

Leseprobe

1 d4 d5 2 c4 dxc4

To my brother Jeremy, for his continued support

The Queen’s Gambit Accepted

BATSFORD CHESS OPENING GUIDES

Other titles in this series include:

0 7134 8462 4

Benko Gambit

Andrew Kinsman and Byron Jacobs

0 7134 8456 X

Budapest Gambit

Bogdan Lalic

0 7134 8461 6

French Tarrasch

John Emms

0 7134 8451 9

King’s Gambit

Neil McDonald

0 7134 8472 1

Modern Defence

Jon Speelman and Neil McDonald

0 7134 8466 7

Scotch Game

Peter Wells

0 7134 8471 3

Spanish Exchange

Andrew Kinsman

For further details for Batsford chess titles, please write to Batsford Chess Books, 583 Fulham Road, London SW6 5BY.

Batsford Chess Opening Guides

The Queen’s Gambit Accepted

Chris Ward

CONTENTS

1 d4 d5 2 c4 dxc4

Bibliography

Introduction

Part One: 3 e4

1    Central Variation with 3...♘c6

2    Central Variation with 3...♘f6

3    Central Variation with 3...c5

4    Central Variation with 3...e5

Part Two: 3 ♘f3 ♘f6 4 e3

5    Classical Variation without an early ♕e2

6    Classical Variation with an early ♕e2

7    Fourth Move Alternatives for Black

Part Three: Other Variations

8    3 ♘f3 ♘f6 4 ♘c3

9    3 ♕a4 + and 4 ♘f3 ♘f6 5 ♕a4 +

10    Third Move Alternatives for Black after 3 ♘f3

11    3 ♘c3 and 3 e3

Index of Complete Games

BIBLIOGRAPHY

Books

Queen’s Gambit Accepted, Eduard Gufeld (Batsford 1986)

Encyclopaedia of Chess Openings vol.D, (Sahovski Informator 1987)

New Ideas in the Queen’s Gambit Accepted, Glenn Flear (Batsford 1994)

The Queen’s Gambit Accepted, Iakov Neishtadt (Cadogan 1997)

Periodicals

Informator

New in Chess Yearbook

Chess Base Megabase CD-ROM

Chess Monthly

British Chess Magazine

INTRODUCTION

1 d4 d5 2 c4 dxc4

In my role as a coach of junior players, the Queen’s Gambit Accepted (QGA) has in the past caused me to reconsider some basic teaching principles. I recall demonstrating the Queen’s Gambit and then on 2...dxc4, have members of the audience chuckle. ‘As if Black would do that’ was the audience view, with the children obviously preferring 2...c6 or 2...e6, keeping the centre intact. Yes, the general opinion was definitely that 3 e4 would be ‘winning’, which brings to mind my own story. As a nine-year-old just starting out, I was happy playing the natural 3 e4 until I was told by an experienced player that it was a mistake. Hence I switched to the more solid 3 ♘f3 and 4 e3, only to discover some seven or eight years later that I had actually been misinformed! Of course at the end of the day it all boils down to taste and I had no hesitation in returning to my first love of 3 e4 (rather bitterly though in view of those wasted years!).

In the 1990s the QGA (whose advocates include Anand and Short) is considered a perfectly acceptable defence. Black does indeed concede the centre, but the intention is to strike back when the time is right. For his part, White may eschew the immediate 3 e4 in favour of the slower plan of retaining this move as a long-term aim (i.e. after he is satisfactorily developed or feels that he is ready).

Black has two main breaks in the QGA, which we deal with in turn:

...c7-c5

After this move Black’s pieces are granted more freedom to manoeuvre. His queen can leave its home square and the c-file could later become useful for a rook or two. White no longer has a free hand in the centre as e3-e4 is dissuaded by the pressure on his d4-pawn. Note that the above diagram is characteristic of the ‘Classical variation’ (Chapters 5 and 6).

...e7-e5

Black generally wants to play either ...c7-c5 or...e7-e5 in the QGA to buy some space of his own. In the above diagram, Black has just played 4...e5! and what better way could there be of preventing White from playing e4-e5 himself? In situations such as this, White gains nothing from the queen trade 5 dxe5 ♕xd1+. Usually Black is fine whether both sides retain their doubled pawns or a c4 for e5 (or e4) swap is initiated. If White advances with 5 d5, then he is reducing the scope of his light-squared bishop, whilst allowing Black access to the c5-g1 diagonal. Also 5...b5!? becomes an option because the c3-knight plays an integral part in holding White’s centre together.

I pride myself on having taken an objective approach throughout this book. I can safely say that I would be prepared to play the QGA with either colour, but first I should take to task the ‘gambit’ part. At grandmaster level, opening gambits are very rarely seen as they tend to have a bad reputation (especially in these days when analysis is aided by materialistic computer programs). Acceptance of the Queen’s ‘Gambit’ would imply that you have snatched a lukewarm to hot pawn and could suggest that you have a little suffering to do before your extra material hopefully sees you through. In reality that’s just not the way things are. The Queen’s Gambit isn’t really a gambit at all, as there is no satisfactory way for Black to keep his pawn. The only way this would be possible would be try something with...b7-b5, but in practice White could undermine this with a2-a4 and b2-b3.

While we’re on the topic, there is an idea that Black must steer clear of. After 3 e3 b5? 4 a4 c6 5 axb5 cxb5

Instead of following up with the usual undermining 6 b3, White has the devastating 6 ♕f3! Black must always be wary of this diagonal and for that matter the damage that the sneaky white queen can cause elsewhere, e.g. 3 e3 ♘f6 4 ♗xc4 ♗g4? 5 ♕b3

Simultaneously attacking f7 and b7.

If White demonstrates that he is not in a hurry to recapture on c4, preferring to produce and then defend a big centre, then there is a playable manoeuvre that can be successful in cementing the ‘gambit’ pawn. After 3 ♘c3 a6!? 4 a4?! (as we shall see in Chapter 8, 4 e4 is the correct response), Black can use the advance of the a-pawns to his advantage with 4...♘c6! 5 e3 ♘a5.

The c4-pawn will remain a thorn for a while as it is surprisingly difficult to regain it (e.g. 6 ♘f3, intending 7 ♘e5xc4, is foiled by 6...♗g4). In such instances, at worst Black will temporarily have the use of the b3-square to cause some damage.

I have mainly used modern games throughout this book to explain both new and old ideas. Obviously theory changes all the time, but whereas individual moves and sequences come and go, the basic plans always remain the same. Basically then don’t blame me if some of the analytical assessments are eventually modified(!) by future games, although I will reiterate that I’m not unhappy with them as they stand.

The arrangement of material in this book is designed to reflect current trends. Although the ‘Classical variation’ (3 ♘f3 ♘f6 4e3 e6 5♗xc4c5)

is usually regarded as the main line, I have included more games in the ‘Central variation’ (3 e4) because frankly I consider it to be a more critical test of the QGA.

CHAPTER ONE

3 e4 ♘c6

1 d4 d5 2 c4 dxc4 3 e4 ♘c6

As you will soon discover, the ‘Central variation’ (3 e4) takes up a hefty percentage of this book. My justification for this is twofold. First, 3 e4 is White’s most natural move, staking a firm claim in the centre; and second it leads to some pretty sharp lines, and is therefore the most testing response to the QGA. If the QGA is to be refuted, then 3 e4 is probably the place for White players to look.

Black has four main responses: 3...♘c6, 3...♘f6, 3...c5 and 3...e5. Since Black is currently struggling with 3...c5 and 3...e5 (see Chapters 3 and 4 respectively) it makes sense to start with the fashionable 3....♘c6. Surprisingly, White is having real trouble eliminating this slightly odd-looking move from the equation. Over the years I have reached the conclusion that this move must be respected. A variety of different attempts to bash it off the board have been tried but, amongst others, Baburin and Sadler remain its loyal servants.

Game 1

Ward-Fugslang

Copenhagen open 1992

1 d4 d5 2 c4 dxc4 3 e4 ♘c6

Having been brought up in a school of thought which holds that it is wrong to obstruct the c-pawn in Queen’s Pawn Openings, I have previously been critical of this move. Nevertheless I have been forced to reconsider my views in recent years as numerous strong players have continued to implement it to good effect. The point behind 3...♘c6 is that, although the counter thrust...c7-c5 is temporarily put on ice, the alternative strike...e7-e5 remains a dynamic option, enabling Black to put pressure on White’s centre before he has had a chance to regain his gambit pawn.

4 ♘f3

In accordance with the general rule of developing knight before bishop and, for the time being at least, keeping the centre pawns flexible. 4 ♗e3 is equally popular (see Games 5-9), whereas 4 d5 is the subject of Game 10.

4...♗g4 5 ♗xc4

For 5 d5, see Game 4.

5...e6

Both 5 ♕b3 and the good old 5 ♗xf7+ ♔xf7 6 ♘g5+ were threatened and the text seems the most natural way to parry both. Less ambitious is 5...♗xf3 6 ♕xf3 (hitting f7) 6...e6 7 d5 ♘e5 8 ♗b5+ c6 9 ♕c3! Amazingly, in our 1994 British Championship encounter, Keith Arkell just accepted a pawn loss here with 9...♗d6 10 dxc6 bxc6 11 ♗xc6+ ♘xc6 12 ♕xc6+ ♔e7 13 ♕b7+ ♕c7 14 ♕xc7+ ♗xc7. Not a strategy I could recommend! Black should have played 9...cxb5 10 ♕xe5 ♕d6 11 ♕xd6 ♗xd6 12 dxe6 fxe6. Having studied this endgame a little, I can tell you that this position isn’t as easy for White as one might think. Nevertheless, being sort of half a pawn up, he definitely has the upper hand.

6♗b5

The old continuation. White intends to meet a doubling of his f-pawns with an even weaker doubling of Black’s c-pawns. 6 d5 is considered in Game 2 and 6 ♗e3 in Game 3.

6...♗b4+?!

It almost seems incredible that I am picking on such a natural-looking move. After all doesn’t it make sense to extract this bishop before playing...♘ge7 to support the knight on c6? As this instructive game shows though, Black cannot allow White to consolidate his centre and I must recommend the immediate 6...♘ge7! Black’s aim should be to pressurise White’s d4-pawn and after 7 ♘bd2 (preparing to replace the other knight when captured) 7...a6 8 ♗e2 (a useful retreat, bearing in mind the trick 8...♘xd4?? 9 ♘xd4 ♗xe2 10 ♘xe2) 8...♘g6 (in my opinion better than the older 8...♕d6, which prepares queenside castling, but is vulnerable to a later ♘c4) 9 h3 ♗xf3 10 ♘xf3, he should continue with 10...♘h4!, as Matthew Sadler demonstrated against me at the 1994 Isle of Man open. I tried 11 d5 exd5 12 exd5 ♘xf3+ 13 ♗xf3 ♘e5 14 0-0 ♗d6 15 ♗e2, but despite having the two bishops found that my isolated pawn clogged me up a bit. In short I was planless and so I must conclude that Black has a very comfortable position.

7 ♘c3 ♘ge7 8 a3!

As 8...♗a5? falls foul of 9 b4 ♗b6 10 d5! winning a piece, this effectively forces Black to trade his bishop for a knight, whilst solidifying White’s centre at the same time.

8...♗xc3+ 9 bxc3 0-0

10 h3 ♗h5

Naturally Black is loathe to concede his second bishop which would leave White with both the bishop pair and a useful space advantage, courtesy of his nifty pawn centre.

11 g4 ♗g6 12 ♕e2 a6 13 ♗d3 ♘a5

In view of the game continuation, this looks a little slow, especially as Black seems to be caught in two minds as to whether or not he should break out with ...c7-c5 or...f7-f5. The problem is that there isn’t much else to recommend and at least this threatens 14...♘b3.

14 ♖b1 f5 15 exf5 exf5 16 ♗g5 ♘ac6

16...♕d7, hoping to utilise the pin on the e-file, is no better, e.g. 17 ♗xe7 ♖fe8 18 gxf5 ♖xe7 19 ♘e5 and White is winning.

17 d5!

The point is that 18 ♗c4 prohibits the queen from capturing this pawn.

17...fxg4 18 ♗xg6 gxf3 19 ♕e6+ ♔h8 20 dxc6 hxg6 21 ♖b4 1-0

Black must concede his queen to avoid 22 ♖h4 mate.

Game 2

Ward-Baburin

Isle of Man open 1997

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♘f3 ♗g4 5 ♗xc4 e6 6 d5

For a while, this move also looked quite good to me. By trading centre pawns, White removes a target and opens up a diagonal for his bishop.

6...exd5 7 ♗xd5 ♕f6!

I guess that this move had previously been rejected for two reasons. First, it takes away an obvious square from the g8knight, while castling queenside on the face of it may run into a ♗g5. It was perhaps not unsurprisingly then, that my opponent admitted that ‘Fritz’ was largely responsible for the development of this big opening novelty.

Both 8 ♕b3 and the simple 8 ♗xc6+ were threatened and that explains why 7...♕d7 had previously been considered best (note that 7...♗b4+ 8 ♗d2 ♗xf3 9 ♗xc6+ bxc6 10 ♕xf3 ♗xd2+ 11 ♘xd2 ♕d4 12 ♕c3 ♕xc3 13 bxc3 left White effectively half a pawn up and able to convert the endgame in Ward-Ahlander, Politiken Cup, Copenhagen 1995). After 7...♕d7 I had been successful against Matthew Sadler (British Championship, Swansea 1995) with 8 ♕b3 0-0-0 9 0-0 ♗xf3 10 ♕xf3 ♘f6 11 ♗xc6 and then more convincingly with 11 ♗g5!? ♘d4 (11...♗e7 12 ♗xc6 ♕xc6 13 ♘c3 h6 14 ♗f4 ♖d7 15 ♖ac1 ♗c5 16 b4!? is also quite difficult for Black) 12 ♕d1 ♗c5 13 ♘c3 ♘e6 (if 13...c6 then 14 ♘a4!) 14 ♗xf6 gxf6 15 ♕h5 against the Swedish IM Christer Hartman (Wrexham 1996). Of course this all seems a little irrelevant now, as this game changes the assessment of the whole variation.

8 ♕a4?!

Taking into consideration my previously impressive record with this line, perhaps I could be forgiven for a little over-enthusiasm. Okay, so clearly this move doesn’t work out and the onus is on White to find an improvement. In some analysis published in Informator 71, Baburin has suggested the sedate variation 8 ♕b3 ♗b4+ 9 ♗d2 ♗xf3 10 ♕xf3 ♘d4 (note that a queen trade allows a ♗xc6+ intermezzo, i.e. doubled isolated pawns are off the agenda!) 11 ♕d3 ♗xd2+ 12 ♘xd2 and the more crazy line (which is definitely worth investing some time on) 8 ♘c3 ♗b4 9 h3 ♗xf3 10 ♕xf3 ♘d4 11 ♕d3 c6 12 0-0! ♖d8 13 e5 ♕xe5 14 ♗xf7+ ♔xf7 15 ♕c4+ ♕e6 16 ♕xb4 ♘c2 17 ♕f4+ ♘f6 18 ♖b1, both apparently with equal chances.

8...♗xf3 9 ♗xc6+ bxc6

In terms of pawn structure alone, Black gets the worse deal from this bishop and knight trade. His queen is the more active of the two, so 9...♕xc6? makes no sense.

10gxf3 ♗c5 11 ♘c3 ♘e7

Outwardly White’s position doesn’t seem too bad. However, upon closer inspection it becomes clear that he has problems with his f-pawns as well as the safety of his king.

12 f4 0-0 13 ♕c4 ♗b6 14 ♗d2 ♖ad8 15 0-0-0 ♗xf2 16 ♘e2 ♖xd2 17 ♖xd2 ♗e3 18 ♕d3 ♗xd2+ 19 ♕xd2 ♘g6 20 ♖d1 ♕h4!

The extra queenside pawn is hardly significant. Hence Black ignores that feature and instead sets about exposing the weaknesses in White’s kingside structure. Specifically, Black attacks the isolated h-pawn whilst keeping tabs on the f-pawn.

21 f5 ♘e5 22 ♕f4 ♕e7

Having provoked the concession of the e5-square, Black’s menacing queen demonstrates its flexibility by preparing to switch to the queenside. However, 22...♕f6! (potentially eyeing up the b2-pawn) would have been more accurate as it would have prevented the counterplay that follows.

23 f6! ♕xf6 24 ♕xf6 gxf6

Previously White was just having to defend. Now, however, he has no less than six isolated pawns to target. Nevertheless Black is still favourite for the whole point.

25 ♖d4 ♖b8 26 ♖a4 ♖b7 27 ♔c2 c5 28 ♖a3 c6 29 ♖g3+ ♔f8 30 ♖h3 ♖b4 31 ♘g3 ♖c4+ 32 ♔b3 a5?!

The more cautious 32...♔g7 33 ♘f5+ ♔h8 would have been more accurate.

33 ♖xh7 a4+ 34 ♔a3 ♖b4 35 b3 ♔e7 36 ♖h8 ♖b7 37 ♘f5+ ♔d7 38 ♖a8 axb3 39 axb3 c4 40 b4 c5 41 bxc5 c3 42 ♘d4 ♘c6

43 ♘xc6??

A reckless move, played only because I was trying to be clever and, ‘knowing’ that the game was destined to be a draw, wanted to grasp the opportunity for a snazzy finish. As it transpires, both 43 ♖a4 (which I had seen) and 43 ♖a6 (which I hadn’t) were simple draws.

43...c2

This was the move I was expecting, but I’d also seen the failure of 43...♖b1 44 ♘b8+! ♔e6? (44...♔e7 45 ♘c6+ will be a draw by repetition) 45 ♖a6+ ♔e5 46 ♘c6+ ♔xe4 47 ♘b4.

44 ♘e5+ fxe5 45 c6+ ♔xc6!

A ridiculously simple but strong move from Alex, using good endgame judgement rather than needlessly entering complex fantasy variations. I had needlessly spent time on 45...♔d6 46 cxb7 c1♕+ 47 ♔a2, when despite plenty of available checks Black can’t win without the access to the d5-square that is required to mop up the b7-pawn, and 45...♔c7 46 ♖c8+! ♔xc8 47 cxb7+ ♔xb7 48 ♔b2 when only White could be better in the king and pawn ending (but actually it’s a draw).

46 ♖c8+ ♔d6 47 ♖xc2 f5!!

Excellent. With White’s king cut off, Black’s one passed pawn is far stronger than both of White’s put together.

48 exf5 e4 49 ♖b2 ♖c7 50 ♔b3 ♔e5 51 ♖c2 ♖d7 52 ♔c3 e3 53 f6 ♔e4 54 ♖g2 ♖c7+ 55 ♔b2 ♖f7 56 ♔c2 ♖xf6 57 ♖g3 ♖c6+ 58 ♔d1 ♔d3 59 h4 ♖a6 60 ♔c1 ♖c6+ 61 ♔d1 ♖a6 62 ♔c1 ♖a1+ 63 ♔b2 ♖h1 64 h5 ♖xh5 65 ♔c1 ♖h1+ 66 ♔b2 ♔d2 0-1

Game 3

Inkiov-Lautier

French Team Championship 1996

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♘f3 ♗g4 5 ♗xc4 e6 6 ♗e3 ♗xf3 7 gxf3 ♕f6

Essentially this differs from the previous two games in that Black has doubled White’s f-pawns without risking the doubling of his own c-pawns. Of course matters aren’t so simple because, after all, White gained a bishop for a knight out of the sixth move trade.

8 ♗b5 ♗b4+ 9 ♘c3 ♘ge7 10 a3?!

This only forces the bishop back to where it wants to go (namely b6 to continue pressurising the d4-pawn). Compared to Game 1, White is unable to follow up with 11 b4 and 12 d5 because his knight would be hanging.

10...♗a5

11 f4 0-0-0 12 e5

Giving away squares left, right and centre. Nevertheless, something had to be done to avoid the immediate loss of the d-pawn.

12...♕h4 13 ♕a4 a6 14 ♗xc6

Naturally this is unattractive as White concedes his good bishop, knowing that Black has another knight where that came from! Unfortunately, the retreat 14 ♗e2 b5! 15 ♕d1 ♘xd4! 16 ♗xd4 ♘f5 is clearly no better.

14...♘xc6 15 0-0-0 ♗xc3 16 bxc3 g5!?

White has all seven pawns on the same colour as his bishop and now Black uses the fact that the queens are on the same rank to open things up a bit.

17 fxg5 ♘xe5 18 ♖hg1

18...♘g4 was threatened, but now that would run into 19 d5!, when it would be White’s turn to use the pin.

18...♘c6!

The knight returns, having provided the black queen with more options.

19♕c2

White chooses to let his h-pawn go in the hope of some counterplay. The more passive 19 ♖g2 could have been met with either 19...♕e4 or 19...e5.

19...♕xh2 20 ♖h1 ♕d6 21 ♕b3 ♘a5 22 ♕b4 ♕d5!

Because his h-pawn is a little weak, Black is not interested in an endgame just yet. Instead he retains the light-squared blockade and enables a...♖d6-b6 manoeuvre.

23 ♗f4 ♘c4 24 a4 ♖d7 25 ♖h4 e5 26 ♗g3 ♖hd8 27 ♖e1 a5 28 ♕b3 ♖d6! 29 ♗xe5 ♖b6 30 ♕c2 ♘a3 31 ♕f5+ ♔b8 32 ♗xc7+

Desperate times call for desperate measures. It’s too late to try and defend, e.g. 32 ♖h3 ♕a2! or 32 ♖e2♕c4!

32...♔xc7 33 ♖e7+ ♔b8 34 ♕xd5 ♖xd5 35 ♖xh7 ♖xg5 36 ♖hxf7 ♘c4 37 ♔c2 ♖g1 38 ♖f8+ ♔a7 39 ♖ee8 ♖b2+ 40 ♔d3 ♘b6 41 ♖e5 ♖d1+ 42 ♔e4 ♖e2+ 43 ♔f3 ♖xe5 44 dxe5 ♖e1 0-1

Game 4

Karpov-Milov

Biel 1997

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♘f3 ♗g4 5d5

An aggressive alternative which aims to unsettle Black before he gets his kingside developed.

5...♘e5

6♗f4

In recent years this has risen in popularity, while the speculative 6 ♕d4 (!? or ?!) has gone out of fashion. The justification for the latter is 6...♘xf3+ 7 gxf3 ♗xf3 8 ♗xc4 ♗xh1? 9 ♗b5+ c6 10 dxc6 a6 11 c7+ with a winning advantage, but 8...e5 9 ♗b5+ c6 10 ♕xe5+ ♕e7! 11 dxc6 ♕xe5 12 cxb7+ ♕xb5 13 bxa8♕+ ♔d7 14 ♕xa7+ ♔e6 leaves White without checks and struggling after 15 ♘c3 ♗b4 (i.e. again threatening 16...♕e2 mate and the h1-rook). To further dampen White’s enthusiasm for this line, there are also other promising eighth move alternatives for Black to consider (e.g. 8...f5!?).

6...♘g6

At this stage, Black wisely rejects capturing on f3 with either piece. White’s pawns would become doubled, but Black is in real danger of seriously falling behind in development and/or becoming rather vulnerable on the light squares. It’s not difficult to imagine him being overrun in the centre.

7 ♗e3

Better than 7 ♗g3 after which the bishop is left with less scope after 7...e5.

7...♘f6 8 ♘c3 e5

This is almost universally played now. The old move 8...e6 9 ♕a4+!? ♕d7 10 ♕xd7+ ♔xd7 11 ♗xc4 leaves Black’s king a little awkwardly placed, but 10...♘xd7 takes the pressure off d5 and is well met by 11 ♘d2!

9 ♗xc4 a6 10 0-0 ♗d6 11 ♗e2

Vacating the c4-square and unpinning the knight so that it can begin a journey there (or to c5) via d2.

11...0-0 12♘d2 ♗d7

Black wants to keep his good bishop and is already considering hassling White’s bishop with...♘f4.

13♖c1

Probably the best continuation. Karpov considers 13 g3 to be overly cautious, particularly as Black should (failing anything drastically brilliant on the kingside) grasp any opportunity to get in...c7-c6, e.g. 13...b5 (threatening 14...b4 winning the e-pawn) 14 a3 c6! Similarly 13 ♘c4 might be premature because of 13...♗b4 (again indirectly hitting e4) 14 ♕c2 b5 15 ♘d2 c6!

13...♕e7 14 a3

Implementing typical Karpovian suffocation techniques. The possibility of responding to ♘c4 with...♗b4 is removed from the agenda.

14...b5?!

Controlling c4, but irrevocably weakening both the c6- and c5-squares. I’m not sure that this is a fair trade and suggest that the immediate 14...♘f4 would have been preferable.

15♘b3 ♘f4

Another problem with the...b7-b5 concession is that potential...c7-c6 break is pretty much gone for good, thus leaving all of Black’s eggs in the kingside basket. For example, after 15...c6?! 16 dxc6 ♗xc6 17 ♘a5 ♗xe4 18 ♘xe4 ♘xe4 19 ♗f3 f5 20 ♕d5+ ♔h8 21 ♗xe4 fxe4 22 ♖fd1 Black has holes and weaknesses everywhere, whereas White’s position could hardly be more solid.

16 ♗f3 ♔h8

Black figures that the g-file is going to be a good way to enter a rook or two into the proceedings and so vacates the g8-square in favour of the immediate 16...g5!?

17 ♘a2 g5 18 ♘c5 ♖g8 19 ♘b4 ♖g6 20 ♕c2

Karpov may be one of the coolest characters around (most players would be worried by the build-up of enemy troops around the white king), but here he shows that he is in no hurry to capture the pawn on a6 (or for that matter the bishop on d7).

20...g4 21 ♗e2 ♖ag8

The problem throughout for Black is that his g-pawn continually seems to get in the way.

22 ♖fd1 ♘6h5

23 g3!

Played not so much just to attack the knight, but rather to prevent Black from playing 23...g3 himself. In that case, Black would be more than happy to sacrifice a piece (or even two!).

23...♗c8 24 ♘c6

Taking the knight would be tantamount to suicide, i.e. 24 gxf4? g3 25 hxg3 ♘xg3 26 fxg3 ♖xg3+ 27 ♔f 1 ♕h4!

24...♕g5 25 ♗f1 ♖h6 26 ♕c3

Offering future lateral defence as well as setting his sights on the e5-pawn.

26...♘f6 27 ♘d3!

Again 27 gxf4?! is unnecessary. It is better to frustrate Black who has a massive strike force but a g-pawn that is annoyingly surplus to requirements!

27...♕h5 28 h4 gxh3

Now of course this pawn is getting in the way on the h-file and, in the absence of checkmate, Black’s position is rapidly coming apart at the seams.

29 ♘dxe5 ♖g7 30 ♗xf4 ♘xe4 31 ♕e3 ♕f5 32 ♗xh6 h2+ 33 ♔xh2 ♘xf2 34 ♗xg7+

Karpov has played a superb game. He never appeared fazed by his opponent’s offensive movements and with his king shaken, but not stirred, he finishes the game embarrassingly ahead on material.

34...♔xg7 35 ♖d4 1-0

Game 5

Savchenko-Zaja

Nova Gorica 1997

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♗e3

In breaking the rule ‘knights before bishops’, White at least prevents the 4♘f3 ♗g4 pin.

4...e5

4...♘f6 is seen in Games 6-9.

5 d5 ♘ce7 6 ♗xc4 ♘g6

A typical manoeuvre for the black knight, protecting the e5-pawn and eyeing up the f4-square for a future invasion.

7 ♗b5+

Considering that his fixed pawns are on light squares, this check, essentially seeking a trade of good for bad bishop, is very logical. Having said that, the immediate 7 ♕b3!? (with possible ideas of d5-d6) is also worthy of attention.

7...♗d7 8 ♕b3 b6?!

This game is a good demonstration of how White has the superior pawn structure and can expect an easy ride in these type of positions if Black is unable to generate the right kind of piece play. The text is too slow and Black must be prepared to offload the b-pawn, e.g. 8...♘f6!? 9 ♗xd7+ ♘xd7 10 ♕xb7 ♖b8 with 11...♘h4 to follow. True, it’s fairly unclear, but in contrast to the game, where he is slowly suffocated, at least Black gets his share of the action.

9 ♘e2 ♘f6 10 f3

Defending the e-pawn and stopping 10...♘g4. Later we see another reason for delaying moving a knight to c3 (the knight opts to aim for c4 instead).

10...a6 11 ♗xd7+ ♘xd7

Black maintains a grip on the c5-square, but the damage has already been done in the weakening of c6.

12♘d2 ♗c5

Black returns the compliment, offering to exchange his bad bishop. Unfortunately, these exchanges only make things less congested for when White’s rooks come into play along the c-file.

13 0-0 ♕g5 14 ♘c4 0-0 15 ♔h1 ♗xe3 16 ♕xe3 ♕xe3 17 ♘xe3

Very simple chess. For Black now it’s going to be a matter of all hands to the pump to defend the c7-pawn.

17...♘e7 18 ♖fd1

Slightly surprising as you would think that bringing a rook (and then two!) to the c-file immediately would be more logical.

18...♖a7 19 ♖ac1 ♖c8 20 ♘c4 f6 21 ♖c3 a5

Creating the illusion of guaranteeing the c5-square for a knight. In reality White can always get b2-b4 in anyway and besides a d5-d6 ‘mixer’ sometime would also weaken the b6-pawn.

22 ♖c2 b5 23 ♘e3 ♖b8 24 ♖dc1 ♖bb7 25 g4 g6 26 ♔g2 ♔f7 27 h4 a4 28 g5 f5

29 ♖c6!

Having also weakened Black’s kingside pawn structure, the white rooks now seek new life along the sixth rank.

29...fxe4 30 fxe4 ♘xc6

Effectively throwing in the towel, but who could blame him. All of Black’s pieces are defending.

31 dxc6 ♖b6 32 ♘d5 ♖ba6 33 cxd7 ♖d6 34 ♘f6 ♖d2 35 ♖c2 1-0

Game 6

Savchenko-Ibragimov

Alushta 1993

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♗e3 ♘f6 5 ♘c3

More natural than 5 f3 (Games 8 and 9), but nonetheless allowing Black’s reply.

5...♘g4!?

The less direct but certainly interesting 5...e5 6 d5 ♘a5!? is the subject of the next main game.

6 e5!?

If this really works out, then it is a very nice concept. Black is intending to capture the bishop and strike out with...e7-e5. If White then relents with d4-d5, he could easily end up with a restricted lightsquared bishop and embarrassing problems on the dark squares. 6 e5!? intercepts this plan and highlights the fact that at present Black is unable to undermine this e-pawn with...c7-c5. With that said and done, the obvious recapture is also not without punch, i.e. 6 ♗xc4 when:

a) 6...e5 7 ♕b3!? (7 ♕f3 could also be tried, but I would avoid 7 d5 for the reasons already stated) 7...♕d7 8 0-0-0!? (instead the less promising 8 ♗b5 exd4 9 0-0-0 a6 10 ♗xc6 ♕xc6 11 ♗xd4 leaves White with a development advantage, but Black with the two bishops) 8...exd4 9 ♘f3 ♗c5 (alternatively 9...♘a5 10 ♕b5 ♕xb5 11 ♗xb5+ c6 12 ♗xd4 cxb5 13 ♘xb5 is too risky) 10 ♕b5 ♗d6 11 ♗xd4 0-0 12 h3 ♘xd4 13 ♕xd7 ♗xd7 14 ♖xd4 ♘xf2 15 ♖f1 ♗e6 16 ♗xe6 fxe6 17 ♖xf2 ♗c5 18 ♖fd2 ♗xd4 19 ♘xd4 with a clearly better endgame for White in Vyzmanavin-Baburin, Gorky 1989.

b) 6...♘xe3! 7 fxe3 e5 8 ♕h5 (if 8 ♕b3 now, then Black could choose between the ambitious 8...♕h4+ 9 g3 ♕h5 and 8...♗d7) 8...g6 9 ♕f3 f6 results in a very murky position. With both 10 0-0-0 and 10 ♘d5 at his disposal, White certainly isn’t lacking in activity. However, he may live to regret the absence of a dark-squared bishop.

6...♘xe3

Natural, but this time probably not best, as it seems to give White more options. I don’t really believe in plans for a kingside fianchetto here and would prefer 6...♗f5!? 7 ♗xc4 e6 8 ♘f3 ♘xe3 9 fxe3 ♗e7. White’s pawn on e5 spearheads a nice centre, but Black can console himself with the advantage of the two bishops. As an...f7-f6 break isn’t a very attractive proposition, Black should soon consider moving his knight in order to challenge with...c7-c5.

7 fxe3 ♘a5?!

7...e6?! looks too passive, so Black should probably try 7...♗f5. However, White would then have additional options to those discussed above.

8♕a4+!

Initiating a fairly forced sequence which, in view of the fact that...♘a5 is a common theme, Black players should familiarise themselves with.

8...c6 9 b4

9...cxb3 10 axb3 e6

The only real alternative, 10...b6 11 b4 ♘b7 12 ♕xc6+ ♕d7 13 ♕e4, leaves Black horribly uncoordinated and probably having to turn to 13...♔d8 as a solution to 14 ♗b5.

11 b4

11 ♕xa5 ♕xa5 12 ♖xa5 ♗b4 13 ♖c5! would also leave Black with a grim position with just a rook and pawn for two pieces, a poor bishop and ♘e4-d6 staring him in the face.

11...♘c4 12 ♗xc4 b5 13 ♘xb5 cxb5 14 ♗xb5 ♗d7 15 ♘e2 ♗xb5 16 ♕xb5+ ♕d7 17 ♕xd7+ ♔xd7 18 0-0 f6 19 exf6 gxf6 20 ♖xf6 ♗xb4 21 ♖a6 ♖he8 22 ♘f4

White has already handled this endgame nicely. As well as the dominating rooks, Black must also keep tabs on the knight which as well as 23 ♘xe6, harbours other menacing prospects via 23 ♘d3.

22...♗d6 23 d5 ♗c5

23...♗xf4 would have been dealt with by 24 dxe6+ (and if necessary 25 ♖f7+) before recapturing.

24 dxe6+ ♔c7 25 ♖f7+ ♖e7 26 ♘d5+ ♔b7 27 ♖xa7+ ♖xa7 28 ♘xe7 ♗xe3+ 29 ♔f1 ♗c5 30 ♘f5+ 1-0

Game 7

Illescas-Sadler

Linares Zonal 1995

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♗e3 ♘f6 5 ♘c3 e5 6 d5 ♘a5!?

see following diagram

7♘f3

Here 7 ♕a4+?! c6 8 b4? (8 dxc6 ♘xc6 9 ♗xc4 ♗b4 10 ♘f3 0-0 11 0-0 ♗xc3 12 bxc3 ♘xe4 13 ♗xf7+ ♖xf7 14 ♕xe4 ♗f5 was equal in Vyzmanavin-Sadler, Paris rapidplay 1994) would be unsuccessful because of 8...cxb3 9 axb3 b5 10 ♕xa5 ♕xa5 11 ♖xa5 ♗b4.

7...♗d6

It looks more sensible to defend the e-pawn like this than to play 7...♘g4 8 ♗g5!, although that possibility shouldn’t be completely ruled out.

8♕a4+

Absolutely critical as the calmer 8 ♘d2 runs into 8...♘g4!?

8...♗d7!

A very tricky move which heralds an idea essential to the survival of this variation. Note that 8...c6 9 dxc6 ♘xc6 10 ♗xc4 0-0 11 ♖d1 is just better for White, as Black has trouble finding a place for his queen.

9 ♕xa5 a6!!

A fantastic idea. The queen is trapped, although that’s certainly not the end of the story.

10♘b1??

The obvious threat was 10...b6, but there were two better attempts at queen retrieval than this:

a) At first 10 ♘a4!? (preparing 12 ♘xb6 if necessary) looks quite good because after 10...♘xe4 11 ♗xc4 b5 12 ♗d3 ♘f6 (12...♕e7 is met by 13 0-0, as the c7-pawn provides another escape route for the queen), White can save his knight and prevent 13...e4 with 13 ♘c3. However, Black should insert 10...♕e7! Then 11 a3 ♘xe4 12 ♗xc4 b5 13 ♗d3 ♘f6 14 ♘c3 e4 15 ♘xb5 (perhaps 15 ♘xe4 ♘xe4 15 0-0 0-0 16 ♖fe1 would be more shrewd, but Black is fine) 15...exf3 16 ♘xc7+ ♗xc7 17 ♕xc7 fxg2 18 ♖g1 0-0 would be very double edged, whilst 11 0-0-0 ♗b4 12 ♕xc7 ♗xa4 13 ♕xc7+ ♔xe7 is simply better for Black.

b) 10 b4!? has been recommended as the refutation of Black’s play. As you will see though, I’m not so sure. 10...cxb3? 11 axb3 b6 12 ♕a2 ♗b4 13 ♗d2 ♗xc3 14 ♗xc3 ♘xe4 15 ♗xe5 is winning for White, and so Black should investigate 10...b6! Then after 11 ♕a3, he should continue his queen hunt with 11...a5. In my opinion both 12 ♕b2 axb4 13 ♘d1 b5 and 12 ♕c1 axb4 13 ♘e2 ♘xe4 14 ♕xc4 f5 15 ♘g3 ♘f6 16 ♗g5 h6 17 ♗xf6 ♕xf6 leave Black in the more comfortable position despite only having two pawns for the piece.

10...♘xe4 11 ♔d1

11 b4 would have been a better practical decision, but it compares unfavourably to the notes to 10 b4.

11...c3 0-1

An amazing position to resign in especially as computer programs still suggest that White is a little better! Of course, that’s rubbish though no doubt Illescas was still pretty shell-shocked and besides 12 b4 b6 13 ♕a3 a5 14 ♗d3 axb4 15 ♕c1 ♘c5 (or 15...♗a4+) is pretty grim.

Game 8

Christiansen-Benjamin

USA Ch., Chandler 1997

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♗e3 ♘f6 5 f3 e5 6 d5

Other than preventing...♘g4 ideas, the advantage of 5 f3 over 5 ♘c3 is that 6...♘a5? is now unplayable because of 7 ♕a4+c6 8♗d2.

6...♘e7

6...♘d4 is seen in the next game.

7 ♗xc4 a6

Played instead of 7...♘g6 and thus sidestepping 8 ♗b5+ ♘d7 (it’s quite common for the knight to blockade such a check because Black naturally wants to preserve his good bishop) 9 ♘e2 a6 10 ♗d3 ♗d6 11 ♕d2 ♕h4+ 12 g3 ♕e7 13 ♘bc3 ♘f6 14 h4! h6 15 0-0-0 ♗d7 16 ♔b1 ♗c5 17 ♖c1 ♗xe3 18 ♕xe3, which is better for White because of his space advantage and pressure along the c-file. It is somewhat ironic how often Black misses his ‘bad’ dark-squared bishop, which solves a lot of problems by sitting on d6.

8 ♘e2 ♘g6 9 0-0 ♗d6 10 ♕d2 ♕e7 11 ♗b3 ♗d7 12 ♘bc3 0-0 13 ♘g3

This assessment of this type of position is critical to the whole 3...♘c6 variation. The prevailing view is that White has the slightly better chances, although Black could hardly be more solid. I suppose that Black has the only real weakness (the c7-pawn), with Black’s bad bishop guarding it and White’s bad bishop trying to ensure that it doesn’t advance and get swapped off.

13...♖fd8

I must confess, though, to being rather baffled by Black’s play in this game and particularly with his unnatural reluctance to hop his knight into the tempting f4-square. If the knight were then captured by White’s bishop, Black would have the e5-square at his disposal and could perhaps initiate some action against the enemy king. White has an obvious plan of attacking the c7-pawn and it certainly seems to me that any kingside play by Black here comes too little too late.

14♔h1 ♗c5

In the middlegame, a plan of trading ‘bad’ for ‘good’ bishops is often to be recommended. Having seen several games of this kind, though, I would have to question the validity of such a policy here, at least while all the rooks remain.

15 ♖fe1 ♗xe3 16 ♕xe3 ♘e8

The d6-square certainly looks ripe for occupation by a black knight and indeed it would perform some useful functions. As always the problem for Black would be how to avoid losing his c-pawn.

17 ♘f5 ♕f8

17...♗xf5? loses the e-pawn as well as donating the e4-square to the other white knight, but 17...♕f6 18 ♖ac1 ♘f4 looks more to the point.

18 ♖ac1 ♘d6 19 ♘xd6 ♕xd6 20 ♖ed1 ♖dc8 21 ♖d2 b5?!

This prevents 22 ♘a4, but White easily finds another way to get his knight to the promising c5-square.

22 ♘d1 a5 23 ♘f2 h5 24 ♘d3 ♖ab8 25 a3 h4 26 g3 hxg3 27 hxg3 ♔f8 28 ♔g2 ♔e7 29 ♘c5 ♗h3+

Presumably now after 30 ♔xh3?, Black would have to go 30...♘f4+! and indeed that looks very strong. However, that all seems irrelevant as White can satisfactorily ignore this sacrifice and continue with his queenside play.

30 ♔f2 ♕f6 31 ♘d3

Safe and simple. White vacates his knight from the c-file in anticipation of wreaking havoc with his rooks. The immediate 31 ♖dc2 also looked good, while the dinky 31 ♘b7 (with 31...♖xc7 32 ♖c6 in mind) would also have been a nice move to play.

31...♕b6 32 ♖dc2 ♗d7 33 ♖xc7!

No doubt not the bolt out of the blue that a casual observer might imagine. The pressure on the c7-pawn was always likely to become unbearable.

33...♕xe3+ 34 ♔xe3 ♔d6 35 ♖xc8 ♖xc8 36 ♖h1

The presence of fixed pawns on the same colour as his bishop render this endgame not exactly straightforward for White. Nevertheless, Larry Christiansen now shows that he is more than up to the task.

36...♖h8 37 ♖xh8 ♘xh8 38 f4 f6 39 ♗d1 ♘g6 40 b4 axb4 41 axb4 ♘f8 42 ♗e2 ♗e8 43 ♘f2 ♔e7 44 fxe5 fxe5 45 ♘d1 1-0

Game 9

Thorsteins-Greenfeld

Reykjavik 1996

1 d4 d5 2 c4 dxc4 3 e4 ♘c6 4 ♗e3 ♘f6 5 f3 e5 6 d5 ♘d4!?

Certainly bold and in many respects a very logical response. Black is in effect chastising White for having eschewed a natural developing move (5 ♘c3), whilst simultaneously creating the odd darksquared hole. Black challenges his opponent to part with his good bishop, albeit at the cost of a pawn.

7 ♗xd4 exd4 8 ♕xd4

The problem for Black here is that White might just be able to paper over the cracks in the long run by switching the ‘troublemaker’ pawns (particularly the fand e-pawns) to dark squares and thus achieve world domination with his centre.

8...b4+ 9 ♘c3 0-0 10 0-0-0