''Fischer's Weapon''

�·

• •
�·

~ ana ion
aq Revised
2nd Edition

�. Andrew Soltis

Chess Digest, Inc.

 Copyright© 1995
Andrew Soltis

All rights reserved under Pan American and International Copy­

right conventions.

ISBN: 0-87568-197-2

1his is a revised and expanded 2nd edition of the 1 st edition pub­

lished in 1992. Added are 23 important theorical revisions and one
game.
No part of this publication may be reproduced, stored in a re­
trieval system, or transmitted in any form, or by any means: elec­
tronic, electrostatic, magnetic tapes, mechanical photocopying, re­
cording, or otherwise, without prior and current permission from the
publisher.

Author: Andrew Soltis

Editor: Ken Smith
Computer Typesetting: E laine Smith
Cover: Elaine Smith
Proofreader: Hugh Myers, Sid Pickard and Bob McLean
Final Preparation & Diagrams: Ken S mith

Publisher: Chess Digest, Inc.®, 1601 Tantor, (P.O. Box 59029) Dal­
las, Texas 75229

Send the publisher $2.00 for the New Chess Guide that catalogs
every chess bookfor general sale in the United States. You are given
publishers, page counts, notation, and critical reviews. Also
included is a free Chess Improvement Course for Beginners up
through Master level players.
Winnini: With The Ruy Lopez Exchani:e Variation 3

TABLE OF CONTENTS
Page
CHAPTER ONE 5
Introduction 5
Fischer-Spassky, Sveti
Stefan 1992
A Little Strategy, A Li ttle History 10
The Terrible Pin 13
Endgames, Endgames, Endgames 15

CHAPTER TWO 25
1 e4, e5 2 Nf3, Nc6 3 Bb5, a6
4 B xc6, dxc6 5 0-0
The Pin 5 ... B g4

CHAPTER THREE 37
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 B xc6, dxc6 5 0-0 , B g4
The Main Line Pin 6 h3, h5!

CHAPTER FOUR 49
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 B xc6, dxc6 5 0- 0
The Endgame 5 .. .f6 and 7 . . .c 5

CHAPTER FIVE 72
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 Bxc6, dxc6 5 0-0, f6
Other 6 d4, exd4 lines

CHAPTER SIX 79
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 B xc6, dxc6 5 0-0
Modern Main Line 5 .. .f6 6 d4, B g4
4 Winning With The Ruy LQPez Excbange Variation

CHAPTER SEVEN 86
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 Bxc6, dxc6 5 0-0
Bronstein's 5... Qd6

CHAPTER EIGHT 94
1 e4, e5 2 Nf6, Nc6 3 Bb5, a6
4 Bxc6, dxc6 5 0-0
The Irregulars
5...Qe7?! 95
5... Qf6? 98
5 ...Bd6 !? 102
5 ... Ne7 105
5...Be7? 109

CHAPTER NINE 111

How Fischer Played It
Chapter One: Introduction 5

CHAPTER ONE: Introduction

When the first edition of this book appeared in 1992, it

evoked a good deal of interest - particularly after a certain day that
September. What prompted the interest that day was a curious inci­
dent off the coast of what was then Yugoslavia:

FISCHER-SPASS KY
Ninth Match Game
Sveti Stefan 1992
1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0, f6 6 d4,
exd4 7 Nxd4, cS 8 Nb3, Qxdl 9 Rxdl, Bg4 10 f3, Be6 1 1 Nc3, Bd6
12 Be3, b6 13 a4, 0-0-0 14 aS, Kb7 15 eS, Be7 16 Rxd8, Bxd8 17
Ne4

17••• Kc6??
18 axb6 cxb6
19 NbxcS! Bc8
20 Nxa6 fxeS
21 Nb4 Resigns!

The ringside annotators immediately focused on White's theo­

retical novelty at move 17 and Black's answering blunder. But 17
-
6 Winnini: With The Ruy Lopez Exchani:e Variation

Ne4 was hardly new: it had been played successfully twelve years
before - and at a Yugoslav international tournament, no less. Clearly,
Bobby Fischer was well aware of the analysis and Boris Spassky
wasn't. And that's precisely the value of good opening preparation.

I have a special fondness for the Exchange Variation of the

Ruy Lopez because it was the first opening I felt I knew. There really
wasn't that much to know about it when, as a high school Class B
player, I began giving up my bishop on the fourth move. Few opening
books of the day revoted more than a column of analysis to the Ex­
change Variation - and fewer gave more than a footnote to my favorite
line in it, 5 0-0.

What attracted me to it was a mixture of reasons:

(a) Nobody else seemed to know anything more about this
opening than I did.

(b) You could learn to play a major opening - the Lopez -

without having to memorize reams of main-line theory, and

(c) It placed an onus on Black.

This last reason was odd, since in most opening systems it is

White who attacks and Black who can sit back and wait for his oppo­
nent to prove that he has something. But after 1 e4, eS 2 Nf3, Nc6 3
BbS, a6 4 Bxc6, dxc6 White assumes a small but perceptible advan­
tage in pawn structure. All things being equal - which, admittedly,
they rarely are - White should win the endgame.

Later, some computers reputedly tested this all-things-being­

equal theory. In hundreds of simulations the computers found that in
the basic Exchange Lopez pawn structure - White pawns at a2, b2,
c2, e4, f2, g2 and h2 versus Black pawns at a6, b7, c6, c7, f7, g7 and
h7 White should, in fact, win.
-
Chapter One: Introduction 7

At the time I began playing the Exchange Lopez, long before

the computer era in chess, this oversimplified view about its endgame
potential comforted me because it meant Black had to prove some­
thing, not White. Black had to show that he had compensation for his
crippled queenside pawn maj ority. If he doesn't, if he simply ex­
changes pieces, White can create a passed kingside pawn that is not
balanced by anything on the queenside. I found this was a theory that
could be put into practice.

SOL TIS-L. RASCHEN

Marshall C.C. Fall "A" Tournament 1964
1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 S 0-0!, f6 6 d4,
exd4 7 Nxd4, cS 8 Nb3, Qxdl 9 Rxdl, Bd7 10 Bf4, 0-0-0 1 1 Nc3,
Be6 12 Rxd8ch, Kxd8 13 Rdlch, Kc8 14 Kfl, b6 15 Nd5, BxdS 16
RxdS, Ne7 17 Rdl, Ng6 18 Bg3, Be7 19 Ke2, Rd8 20 Rxd8ch,
Kxd8 21 Ke3, Bd6 22 f4, Kd7 23 Nd2, bS 24 Nf3, Ne7 25 eS,
Nf5ch 26 Ke4, Nxg3ch 27 hxg3, fxe5 28 NxeSch, BxeS? 29 Kxe5

29 c6 30 g4, Ke7 31 gS, as 32 fS, a4 33 a3, b4 34 c4, b3

•••

3S f6ch, gxf6ch 36 gxf6ch, Kd7 37 f7! Resigns.

White sweeps the queenside pawns away and wins the race
after 37 ... Ke7 38 f8(Q)ch, Kxf8 39 Kd6, Kf7 40 Kxc6, Kf6 41
KxcS, Kg5 42 Kb4, Kg4 43 cS, etc.
8 Winning With The Ruy Lopez Exchange variation

At the same time, I was reading books that ckscribed the Ex­
change Variation as a terrible opening. White very rarely won when
he played 4 Bxc6 in those books. In fact, he usually lost. What both­
ered me is that often his loss had nothing to do with the way he played
the opening.

Here is a typical example. It was played in the same legen­

dary tournament in which Emanuel Lasker used the White side of the
opening to defeat Jose Capablanca and overtake him in the race for
first prize.

ALEKHINE-EM. LASKER
St. Petersburg 1914
1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 Nc3, f6 6 d4,
exd4 7 Qxd4, Qxd4 8 Nxd4, Bd6 9 Be3, Ne7 10 0-0-0, 0-0 1 1
Nb3!, Ng6 1 2 BcS, Bf4ch! 1 3 Kbl, Re8 1 4 Rhel, b6 1 5 Be3, BeS
16 Bd4, Nh4 17 Rgl, Be6 18 f4!, Bd6 19 Bf2, Ng6 20 fS !, Bxb3 21
axb3, Nf8 22 Bxb6, Bxh2 23 Rhl, cxb6 24 Rxh2, bS 25 Rel, Nd7
26 Ndl, a5 27 Rh3, b4 28 Nf2, NcS 29 R(3)e3, a4! 30 bxa4, Nxa4
31 eS!, fxeS 32 RxeS, Reb8 33 Ne4, b3 34 Re2, Nb6!, 35 cxb3, NdS
36 g4, h6 37 gS, hxgS 38 NxgS, Nf6 39 Re7, Rxb3 40 Rg2!, NdS.

Both sides have played resourcefully and it appears White has

the upper hand. Actually, he should play 41 Rel, after which a draw
is almost assured.
Chapter One: Introduction 9

41 Rd7?? Rd3!

But now White must lose the Exchange because 42 ... Rdlch
43 Kc2, Ne3ch is threatened, in addition to 42...Nc3ch. And since 42
Ket allows 42...Ralch and 43 ... Nb4 mate, White had to play 42
RxdS, RxdS. But his moderate drawing chances began to evaporate
soon after 43 Ne6, Kf7 44 Rxg7ch, Kf6 and he resigned 45 moves
later.

For the rest of his life Alekhine stated with authority that the
Exchange Variation favored Black. (Lasker knew better. His victory
over the then-almost-unbeatable Capablanca came a few rounds
later.)

But, it seemed to me, a major cause of White's problems in

this and similar games won by Black was his choice of fifth move.
Why reveJop a knight on c3 that doesn't go anywhere, except, as in
the above game, to dl? And why, with 7 Qxd4, should White rush
into an endgame? If Black wants an ending, I thought, Jet him do the
forcing of it.

And that's when I began to appreciate 5 0-0. By indirectly

safeguarding the e-pawn, it threatened Black's pawn (6 Nxe5, Qd4 7
Nj3, Qxe4? ? 8 Rel). And the endgames that result from 5 0-0 are
much better than the 5 Nc3 ones. They can even end quickly:
10 Winning With The Ruy Lopez Exchange Variation

BARENDREGT-SLIWA
Marianske Lazne 1961
1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0, f6 6 d4,
exd4 7 Nxd4!, cS 8 Nb3, Qxdl 9 Rxdl, Bd7 10 Bf4, 0-0-0 1 1 Nc3,
Re8? 12 NdS!, Bc6 13 Nxc7, Rxe4

14 f3!, Re2 (14 . .. Rxf4 15 Ne6!) 15 Na8! (threatens 16 Nb6 mate!),

bS 16 Nb6ch, Kb7 17 Nd7, Bxd7 18 Rxd7ch, Kc6 19 Rc7ch, Kb6
20 Rc8, Nh6 21 Rb8ch Resigns (e.g. 21 ... Kc6 22 Rdl intending 23
Na5 mate)

A Little Strategy, A Little History

The Exchange Variation is not particularly complex from a

strategic point of view. But to start out, it's worth comparing it with
the main line of the Lopez.

With the moves 1 e4, eS 2 Nf3, Nc6 3 BbS White makes an

indirect threat to the e-pawn. He knows that even if he had a free
move he could not win a pawn directly (3... "pass" 4 Bxc6, dxc6 5
Nxe5, Qd4!). But he realizes that at some point he will have protected
his own e-pawn and at that point the threat of Bxc6 followed by NxeS
will become very real.
Chapter One: Introduction 11

In the early days of the Ruy Lopez, Black tried a variety of

third moves, but had some difficulty dealing with this indirect threat
to his center pawn. As soon as White defended his own pawn with d2-
d3 or Nc3, they began to flounder. Even some of the best players in
the world would meet 1 e4, eS 2 Nf3, Nc6 3 BbS, Nf6 4 d3, for
example, with 4 Bd6?. That move does meet the 5 Bxc6 threat - but
.••

delays his development horribly.

Enter Paul Morphy. In his celebrated match with Adolph

Anderssen in 1 858, Morphy popularized the move 3 a6 as a means
..•

of dealing with the threat to the e-pawn. The move had been tried in
the previous decade by players who didn't understand it and who met
4 Ba4 with 4...bS, thereby driving the bishop from one good diagonal
to a better one - weakening the Black position in the process and get­
ting nothing in return.

But Morphy had a better idea. After Anderssen responded to

his 3 ...a6 with 4 Ba4, Morphy continued 4 Nf6! S d3, BcS (the
•••

pawn is still safe because 6 Bxc6, dxc6 7 Nxe5, Qd4 threatens mate
as well as the knight). Anderssen appreciated this last point and re­
plied 6 c3, reviving the threat of Bxc6.

But Morphy illustrated the usefulness of his third move by

now inserting 6 bSI. He then castles quickly and can even play . d7-
... . .
12 Winning With The Ruy Lopez Exchange Variation

d5, after which Morphy had solved his most serious opening problem.
(After two unsatisfactory Lopezes, Anderssen switched to 1 a3!? for
the rest of their match.) And as a result of the American's success, the
"Morphy Defense" became the basis of the main Lopez lines, which to
this day are characterized by 1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Ba4,
Nf6 5 0-0, Be7 and if 6 Rel (or 6 d3, 6 Nc3, etc), then 6... b5!.

The Exchange Variation alters the sequence of events slightly.

White exchanges on c6 before he defends his e-pawn. And, since most
Black defenders now use 3 ..a6 instead of developing moves such as
.

3 ...Nf6 or 3 .. Bc5, White can accomplish this with a gain of time.

.

After 3 ...a6 4 Bxc6, dxc6 5 0-0 White is well ahead in development.

In fact, Black has no pieces off his first rank.

How then should Black defend the e-pawn? If he plays 5..J6,

White should be eager to force an exchange of center pawns (6 d4,
exd4) that leaves Black's e6 hole as a potential target (see Chapters 4-
5) . If he plays 5 ... Qd6 or 5 ... Bd6 or 5 ... Qf6 (Chapters 7-8), his
pieces block the development of other pieces.
Chapter One: Introduction 13

The Terrible Pin

But, said the theoreticians, there had to be something wrong

with 5 0-0. They found the flaw in the form of 5 Bg4, establishing
..•

an annoying pin.

If White can break the pin, the position should favor him.
This is particularly true after 5...Bg4 6 h3, Bxf3?!, since the ex­
change of the light-squared bishop reprives Black of his primary
claim to compensation for his double-pawns. After 6...Bxf3 we are
heading once again into a very favorable endgame for White:

BERNSTEIN-BENKNER
Amsterdam Olympiad 1954
1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Bxc6, dxc6 5 0-0, Bg4 6
h3, Bxf3?! 7 Qxf3, Qf6? 8 Qxf6!, Nxf6 9 d3, Bc5 10 Nd2!, Nd7 1 1
Nc4, f6 1 2 g3, 0-0-0 1 3 Kg2, Nb6 1 4 Nxb6ch, Bxb6 1 5 f4, Rhe8 16
f5, c5 17 b3!, Ba5 18 Be3, Bc3 19 Rbl !, b6 (19.. . Bd4 ? 20 Bd2!
traps the Bishop by 21 c3!) 20 g4, Kb7 (20. . . h6 21 h4, Rh8 22 Rhl,
Kd7! was the recommended defense with the king coming to the
kingside rescue) 21 g5, Kc6 22 h4, Rg8 23 Kf3, Rdf8 24 Ke2, b5 25
h5, b4.
14 Winnini: With The Ruy Lopez Exchaui:e variation

26 h6!, gxh6 27 gxf6, Rxf6 28 Rgl, Rxgl 29 Rxgl, Bd4? (29 . . Rj7 .

was necessary) 30 Rg7, Bxe3 31 Kxe3, aS 32 Rxh7 Resigns.

A bit trickier than this is the realization that after S 0-0, Bg4
6 h3 Black cannot gambit a pawn with 6...BhS, since the weakened
White kingside following 7 g4! is not really that weak. I appreciated
this after seeing the first nine moves of Fuchs-Udovcic, Berlin 1962,
which went 7 Bg6 8 Nxe5, Qh4 9 Qf3 and "leaves Black insuffi­
•••

cient compensation for a pawn" (Leonard Barden in The Ruy Lopez,

1963). What I didn't know until later is that Black went on to win that
game. I had better luck.

SOLTIS-A. STERN
Marshall C.C.Championship Preliminaries 1964
1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Bxc6, dxc6 S 0-0, Bg4 6
h3, BhS? 7 g4!, Bg6 8 NxeS, Qh4 9 Qf3, Nf6 10 d3, BcS 1 1 Kg2,
Bd4 12 Nxg6, hxg6 13 Qg3, Qh7 14 Nd2, 0-0-0 lS Nf3, BcS 16 a3,
Ne8 17 eS, f6 18 Be3, Be7 19 Rhl , Qg8 20 Qf4, gS 21 Qe4, Rd5 22
c4, Rd7 23 Rael, g6 24 exf6, Bxf6 2S BxgS, Bxb2 26 cS, Nf6?

27 Qb4, Rxd3 28 Qxb2, Rxf3 29 Bxf6!, Qd5 30 ReS!, Rxf2ch 31

Kxf2, Qxhl 32 Bxh8!, Qh2ch 33 Kel , Qglch (33 Qxb2 34 Re8ch)
...
Chapter One: Introduction 15

34 Kd2, Qg2ch 35 Re2, QdSch 36 Qd4, Qa2ch 37 Kel, Qblch 38

Qdl Resigns.

The real test of 5...Bg4 6 h3 was 6...hS!, another idea of the

imaginative, turn of the century Russian Semyon Alapin. At move six,
the advance of the h-pawn constitutes a piece sacrifice which is only
temporary if accepted at move seven - but may be permanent and
unsound if Black leaves the bishop at g4 at move eight, nine, ten and
so on. The original idea of the Dutch master Johan Barendregt was to
meet 6 hS with "7 d4!", which threatens to keep the piece with 8
•••

hxg4, hxg4 9 NgS.

Barendregt also hoped to obtain sufficient compensation if

Black met 7 d4 with 7 Bxf3 8 Qxf3, exd4 9 Rdl or 8 ... Qxd4 9
••.

Nc3. But it was soon discovered - circa 1 964 - that in the latter line
Black need only retreat his queen to e6 via 9 Qc4! to sharply reduce
•••

White's compensation to almost nothing. At the time, it seemed that

was the end of the 5 0-0 revival. But it was only the beginning (See
Chapter Three).

Endgames, Endgames, Endgames

You had better like to play the endgame if you're going to

handle either side of the Exchange Variation. The queens go off the
board very early in quite a few of the key variations. Sometimes it is
White's decision to trade queens (See Chapter Six) and sometimes it is
Black's (Chapter Four).
More often than you'll see in other openings books, we1l be
examining the full score of some 40-, 50- and even 60-move games. A
number of illustrative examples are included in the pages that follow.
And because so many of the key variations we will consider
lead to forcing play in an endgame, you will find that our analysis will
conclude not in a "plus-over-equals" -- but in a "and White wins".
To give you a feeling of what Black is up against in the Ex­
change Ruy Lopez, we'll begin with a typical recent game:
16 Winning With The Ruy LQPez Exchange Yariatjon

ROMERO-VAN der STERREN

Wijk aan Zee 19'Jl

1 e4 es
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6
5 0-0 ffi

The most popular defense by Black. It accounts for more than

half of all Exchange Variation games these days and is considered in
Chapters Four-Six.

6 d4 exd4

This sentences Black to a somewhat passive endgame. The

more active 6...Bg4! is examined in Chapter Six.

7 Nxd4 cS

The only really consistent plan, although Chapter Five con­

siders m iddlegame alternative strategies.

8 Nb3 Qxdl
9 Rxdl Bg4

Whether this is a finesse that improves Black's defensive

chances - or just a waste of time - was debated through much of the
past 15 years. It is still seen often.

10 f3 Be6
1 1 Bf4
Chapter One: Introduction 17

We'll consider the attack on the forward c-pawn (J 1 Be3, b6

12 a4) in Chapter Four. The text allows - if not virtually forces -
Black to swap down to a bishops-of-opposite-color ending.

1 1... c4
12 Nd4 0-0-0

The point of his last move. Now 13 ... cS is threatened.

13 Nc3 Bes
14 Ne2!

White makes sure he will recapture on d4 in the coming ex­

change of minor pieces with a knight, thereby preventing Black from
playing . ..Ne7-c6.

14... Ne7
15 Be3

15... Bxd4

White's knights are very strong (15. . . Bfl 1 6 Kj2, Ng6 1 7

Nf5!) and Black decides he has better drawing chances without them.
However, he is already thinking about drawing. Black's chief problem
18 Winning With The Ruy Lopez Exchange Variation

in this and many similar Exchange Variation endgames is that he

lacks counterplay.

This opening is not like the typical 1 e4 Open Game, but re­
ally has some of the flavor of a 1 d4 game. In the King Pawn games it
is often White attacking somewhere - usually in the center or kingside
- and Black matching him stroke for stroke. If the attack ends, the po­
sition is frequently dead even. (Black does not need counterplay to
survive, just good defense.)

But in Queen Pawn openings, White often establishes an early

positional advantage, such as deriving from a pawn majority on the
queenside or an edge in space there. Black, in such positions, is forced
to complicate matters elsewhere, such as with a kingside attack.

Here we have a similar struggle. White's positional advantage

was staked out with his fourth move. All things being equal, he will
win with his kingside majority. He will (a) bring his king towards the
center, such as to f2, then (b) trade two pairs of rooks along the d-file,
and then (c) create a passed kingside pawn. Black usually finds it very
difficult to create his own passed pawn. Therefore he needs some
other form of counterplay.

16 Nxd4 Bf7
17 Kf2 Rd7

This seems to be playing into White's hands by offering to

trade all four rooks. But with only one open file, such a series of ex­
changes is inevitable. If Black avoids trades (e.g. 1 7. . . Rde8? !) he will
simply be surrendering a great open line to the enemy.

Also, we must consider the alternative strategy , the mobiliza­

tion of his own majority. The way to do that is b7-b5 and
.•• •••c7-c5,
in one move order or the other. But it was Jose Capablanca who
pointed out that advancing Black's queenside pawns in the Lopez is a
double-edged sword. The more they advance the greater is their ag­
gressive power - and the lesser is their defensive ability. Once Black
Chapter One: Introduction 19

starts pushing pawns, he will have to start defending pawns (e.g.

17... c5 18 Ne2, b6 with play as in the game).

18 Ne2 Rhd8
19Rxd7 Rxd7
20 g4!

Signaling his intention to advance his kingside pawns (h2-h4

and g4-g5 with or without support fromf3-f4 and Kf3).

20... b6

Hoping to anchor a knight on d4 after 21. cS and 22 Nc6.

•• ••.

21 Ng3 c5
22 c3!

A key defensive move. Black will not be able to create a

passed pawn now through pawn-only-moves. White will avoid ex­
changing on b4 himself. And if Black ever gets his pawns to b4 and
a4, where they threaten a3!, White will anticipate that by pushing
•••

his own a-pawn to the third.

That mini-position - White pawns at a3, b2 and c3 versus

Black pawns at a4, b4, c4 and c5 - is an important one to understand
because Black will not be able to promote using only pawn moves.
White, however, will most likely be able to create a kingside passer
using just pawn moves. And on that basis, White has won hundreds of
Exchange Lopezes.

22... Nc6
23 f4

This keeps the knight out of e5 and thereby out of d3.

(Black's rook can always go to d3 but, lacking a target there, it doesn't
have the impact of a knight on d3.)
20 Winning With The Ruy Lopez Exchange Variation

23 •.• Be6?!

Black needs counterplay, but this (which hopes for 24 f5?) is

the wrong way to get it. He should begin thinking about extreme
measures such as 23... g5! with the idea of 24 fxg5, Ne5 25 h3,
Nd3ch. This game is typical of how a Black defender places too much
faith in passive defense.

24 h3 g6
25 Ke2 Rd3
26 Rdl! Rxdl?

Playing into White's hands. Black should get his king into
play with 26 Kd7, since 27 Rxd3, cxd3ch 28 Kxd3, Bxa2 makes
..•

Black's queenside pawns healthy again.

27 Kxdl Kd7
28 Ke2 Ke7

Black's king lacks a point of penetration and White can meet

28 Kd6 with either the slow buildup (29 K/3, 30 h4 and 3 1 g5) or
•••

the immediate 29 f5, Bishop moves 30 Bf4ch followed by h3-h4 and

Ke3.

29 f5 Bf7
30 Bf4

White must keep the knight off e5 since otherwise he is vul­

nerable to a Nd3 raid. The key to winning such endgames as White
•••

is to make sure your queenside is secure before trying to promote on

the kingside. White can generally avoid pawn-queening races because
time is on his side.

30... g5
31 Bc7 b5
Chapter One: Introduction 21

32 Nfl!

Careful. At first it seems White has an alternative winning

plan in 32 Ke3 followed by a sacrificial 33 eS! that will allow White
to occupy e4 powerfully with the knight or king. In such positions
there are usually several good winning plans.

But here 32 Ke3? would be an error because of 32... Kd7!

embarrassing the bishop. After 33 Bb6, Kd6 Black gets to occupy eS
while White doesn't get the use of e4. And Black need not fear 34
eSch, KxeS!.

32... Kd7
33 Bg3 NeS?

This natural-looking advance was available for the last few

moves, but it is actually a form of positional surrender.

34 BxeS!

Of course. White gets a protected, passed pawn and targets at

e5 and gS for his knight. Black's only (dim) hope is that somehow
White will not be able to penetrate further with his king.

34... fxeS
22 Winning With The Ruy Lopez Exchange Variation

35 Ne3 Kd6
36 Kf2 Kc6

Black feints in the direction of Kb6-a5 and b5-b4 but he

••. •••

must have known that White would be winning on the kingside with
f5-f6 and Nf5-h6 long before he could have made inroads on the other
wing.

There was, however, nothing better. On 36 Ke7 37 Kf3,

•.•

Kf6 White can play 38 NdSch!, Bxd5 39 exd5 with a won king and
pawn ending. (But he must remember the rule about the queenside
mini-position, so that after 39... b4 40 Ke4, a5 4 1 d6, a4 he plays not
42 Kd5??, a3! but42 a3!).

37 h4 h6

On 37...gxh4 White continues most simply with Kg2-h3xh4.

38 f6

Now the h-pawn is doomed. Black finds a means of keeping

the game going for nearly ten moves.

38... Kd7
39 Nf5 gxh4
40 Nxh6 Bg6
41 Kf3 Ke6
42 g5 h3

Or 42...Bh5ch 43 Kg2, b4 44 Kh3, Bf3 45 r7, Ke7 46 g6,

Kf8 47 Kxh4, Bxe4 48 Kg5, Kg7 49 Nf5ch, Bxf5 50 Kxf5 and wins.

43 Nf5 Bh5ch
44 Kg3 h2
45 Kxh2 Bf3
Chapter One: Introduction 23

46 Kg3! Bhl
Taking the e-pawn allows 47 f7, Kxf7 48 Nd6ch, while
46 BhS allows another knight check, on g7.
•.•

47 Kf2! Black Resigns

The Bishop actually gets trapped now by 48 Ng3, since

47 ... Bxe4 again loses to 48 f7.

Our analysis of Black's defenses to the Exchange Variation

will be considered in this order:

CHAPTER TWO:
The Pin (5. . . Bg4) - Introduction

CHAPTER THREE:
The Main Line Pin (6 . . . h5)

CHAPTER FOUR:
The Endgame (5. .f6 and 7. . . c5)
.

CHAPTER FIVE:
Other 5 .f6/6 . exd4 subvariations
.. ..
24 Winnin2 With The Ruy Lopez Exchani:e Variation

CHAPTER SIX:
Modem Main Line (5. . .f6/6. . . Bg4)

CHAPTER SEVEN:
Bronstein's 5 .. Qd6
.

CHAPTER EIGHT:
The Irregulars (5. . . Qe7?!, 5. . . Qf6?, 5. . . Bd6!?, 5. . . Ne7, 5. . . Be7?)
and finally an analysis of the variation's pioneer --

CHAPTER NINE:
How Fischer Played It.
Chapter Two: The Pin (5 Bg4l
... 25

CHAPTER TWO: The Pin <5 ... Bg4) Introduction

We'll start our analysis with what used to be called the "book
refutation" of 5 0-0.

1 e4 es
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6

The other recapture, 4 bxc6?!, has been discredited for so

•••

long, it only appears in the games of players unfamiliar with "book"

openings. Black's pawn structure appears to be healthier when sepa­
rated into three pawn isl ands than into two. And he can still regain his
pawn after 5 Nxe5 with S ... QgS!.

But after 5 d4! Black lacks the counterplay he usually gets

from the half-open d-file in 4 dxc6 lines. Play would then continue
.••

S exd4 6 Qxd4 and now 6...d6 7 0-0, Nf6 8 Nc3, Be6 9 BgS, Be7
•••

10 Qa4 with good central and queenside pressure (10... Bdl 11 Radl,
0-0 12 e5!, Nd5 13 Bxel, Qxel 14 Nxd5, cxd5 15 Qa3! or 13... Nxel
14 Qh4).

Black does better with 6...cS! 7 Qd3, Ne7 to control dark

squares. But in comparison with a "book" line of the Steinitz Defense
Deferred (1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Ba4, d6 5 Bxc6ch, bxc6 6
d4, exd4 7 Qxd4, c5 8 Qd3) here White has an extra tempo and he
should be able to put it to good use with play such as 8 Nc3, Ng6 9
Be3, Be7 10 0-0-0!, d6 1 1 NdS.

Finally, there is an ancient trap to be avoided 4 ...bxc6 5 d4,

-

exd4 6 Qxd4, Qf6 and now 7 Qd3! followed by 0-0, and Nd4-f5
assures an edge. But not 7 e5, Qg6 8 0-0, Bb7 9 e6?, which seems to
win after 9 fxe6 10 Ne5, but actually loses to 10 Qxg2cht 1 1
••• ••.

Kxg2, cSch!.
26 Winnin2 Ruy Lopez Exchange Variation

S 0-0 Bg4

6 h3

White should "put the question" to the bishop before Black

can add pressure to the f3-knight by way of .... Qf6. In our main line,
we11 see that Black gets that queen move in anyway. The difference is
the addition of the pawn moves h3 by White and hS by Black gives
•••

White the extra option of capturing the bishop at the right moment.

In a way, 6 h3 is an echo of Black's third "question-putting"

move, since 3 ...a6 is played under the assumption that after 4 Ba4
Black gains an extra option (a l ater . . b5).
.

We now have three possibilities but only the third is signifi­

cant: (a) 6...BhS? and (b) 6...Bxf3 analyzed in this chapter, while (c)
6... hS! occupies the next chapter.
Chapter Two: The Pin (5 Bg4)
... 27

A
(1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc65 0-0, Bg4 6 h3)

6... BhS?

A dubious gambit.

7 g4!

There is no solid reason to give Black a chance to defend the

e-pawn with 7 Bd6 -- except in the mistaken hope that Black will
•••

repeat the errors of the 1975 Correspondence game Trommsdorf­

Duchardt, which went 7 d3, BcS? 8 g4!, Bg6 9 NxeS, Qh4 10 Kg2,
0-0-0 1 1 Nc3!? (Just 1 1 Nd2 and 1 2 Ndf3 is safe enough), Ne7 12 f4,
hS?? after which 13 gS ! forced resignation because 14 Nf3 will win
the trapped queen.
7... Bg6

Unlike comparable positions in the delayed Exchange Vari­

ations (e.g. 1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Ba4, Nf6 5 0-0, Be7 6
Bxc6, dxc6 7 d3, Bg4 ? 8 h3, Bh5 9 g4, Nxg4! ?) Black cannot even
sacrifice a piece for two pawns here.
8 Nxe5
28 Winning Ruy Lopez Exchange Variation

This wins a pawn since 8 Bxe4? would allow 9 Rel with a

.•.

deadly battery lined up against the king.

8... Qh4

A natural move but perhaps not the best since White can oust
the queen almost immediately. An Indonesian player named Sampouw
used 8 Bd6 twice at Melbourne 1975. One game gave him
.••

compensation after 9 Nf3, h5 10 g5, Be7 1 1 d3, h4 12 Bf4, f6, but in

the other ( vs. Jamieson) Black had scant compensation following 9
Nxg6!, hxg6 10 Qf3, BeS ( 10... Qh4 11 e5!) 1 1 c3, c5 12 Na3, Qh4
13 Nc4, f6 14 b4!, cxb4 15 d4!.

The latest finesse for Black is 8. .f6 and then 9 Nxg6, hxg6
.

10 Qf3, Qd6 11 Nc3, 0-0-0. This gambit was tried in Benjamin-van

der Wiel, Cannes 1992, but worked out poorly after White returned
the pawn via 12 d4!, Qxd4 13 Rdl, Qc4 14 Rxd8ch followed by Bf4
andQe3.
9 Qf3

This covers the kingside quite nicely and will allow him to
reorganize his position with a subsequent Kg2 and Rhl . Typical play
would be 9 f6 10 Nxg6, hxg6 1 1 Kg2 and now 1 1 0-0-0 12 d3,
.•• ..•

Bd6 13 Rhl !, Ne7 14 Nc3!, g5 15 Ne2 or 13 g5 14 Nc3, Ne7 15

•..

Be3 (Pantaleev-Kristov, Bulgaria 1965).

Chapter Two: The Pin (5 Bg4)
. . 29

B
(1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 S 0-0, Bg4 6 h3)

6... Bxf3
7 Qxf3

Black reduces the pressure on eS but at the cost of his light­

squared bishop, the piece that represented his compensation for the
doubled c-pawn. If he proceeds to castle kingside and handle the
middlegame routinely, he insures a slight but clear inferiority. For
example:

(1) 7 BcS 8 d3, Ne7 9 Nd2, 0-0 10 Nc4, Ng6 11 Qg4! in­
•..

tending to push his f-pawn after the key g2-g3/Kg2 maneuver

(11...Re8 12 g3, Nf8 13 Kh2, Qd7 14 f4, f6 15 Qxd7 and 16 Be3 with
a neat edge, Malevinsky-Borisenko, U.S.S.R. 1972).

(2) 7 Bd6 8 d3, Ne7 9 Nd2, 0-0 10 Nc4, f6 1 1 Qg4, Rti

•••

wid now 12 f4, exf4 13 Bxf4, Bxf4 14 Qxf4, Ng6 was a "grandmaster
draw", in Trapl-Blatny, Bedin 1975 (although White is better). But
12 g3 is a superior preparation for the f-pawn's advance, as in (1).
White wants to preserve pawn control of eS (j41...exf4/gxf4!) in these
lines.
30 Winning Ruy Lopez Exchange Variation

(3) 7 Ne7 8 d3, Ng6 9 Be3 (or 9 Nd2 and JO Nc4 as in the
.••

previous two subvariations), Be7 10 Nd2, cS 1 1 Nc4, 0-0 12 a4, a5

and now either the usual Qg4/g3/Kg2 plan or the immediate 13 Qg3
and 14 f4 should favor White, even if it means giving up a bishop for
a knight on f4.
A sample game was Tukamakov-Kosenko, U.S.S.R. 197S,
which ended remarkably quickly: 13 Qg3, Re8 14 f4, exf4 lS Bxf4,
Nxf4 16 Qxf4, Rf8 17 Rf3, Ra6 18 Qd2, Qd4ch 19 Khl, Rfa8 (an
odd maneuver -- reaching for f6 by way of a6) 20 Raft, Rf6 21 Rxf6,
Bxf6 22 RfS! (intending to trap her majesty with Rd5), Be7 23 Qf4,
Bd6 24 Qf3, g6 2S RdS, Qg7 26 e5, Be7 27 Rd7, Re8 28 Rxc7,
Qh6 29 Nd6 Resigns. Note that 2S RdS is stronger than 2S Rxf7.

(4) 7 ...Nf6 8 d3, Be7 9 Nd2. Now the Black e-pawn is highly
vulnerable to Nc4 and Qg3. If Black tries 9 ... bS, then 10 Qg3, Qd6
11 Nf3! wins a pawn safely (I 1 . . . Nd7 12 Qxg7).

(5) And as mentioned in Chapter One, the endgame that re­

sults from7 ...Qf6 8 Qxf6, Nxf6 just plays into White's hands. This
was known as far back as the great London tournament of 1883 when
Simon Winawer beat George MacKenzie with a more radical treat­
ment than we usually see: 9 d3, Bd6 10 Nc3, h6 1 1 f4, 0-0-0 12 fxe5,
Bxe5 13 Be3, b6 and now 14 d4!?, Bxd4 lS Bxd4, Rxd4 6 e5, Nd7
17 Rxf7, NxeS 18 Rxg7, Rd7 19 Rxd7, Kxd7 20 Rf1 and White
won in 68 moves.
A modem -- and more thematic -- illustration was Fuchs­
Blatny, Varna 1962 in which Black sought to stop the advance of the
enemy f-pawn: 9 d3, Nd7 10 Nd2, BcS 11 Nc4, f6 12 g3! , 0-0-0 13
Kg2! (this preparation for 14 f4 should by now be familiar to you),
gS 14 Be3, Rhg8 lS Rhl !, hS (intending to push the g-pawn to g4 to
avoid the opening the file after 1 6 h4) 16 Raft, Rg6 17 h4, g4 18
BxcS and 19 f4 with an obvious edge.

7 ... Qd7
Chapter Two: The Pin (5,,,Bg4) 31

This is an attempt to complicate matters by castling queen­

side, perhaps with a kingside pawn storm via ... f6/... g5/... h5 and
eventually the line-opening ... g4. The pawn storm, however, can
usually be thwarted by an offer to exchange queens (Qg4 or Qj5). Or,
if White prefers, he can play for his own mate with a general queen­
side advance by pushing pawns to a4, b4, and eventually, b5.
Note that the Black queen needs to control some of the lost
light-colored squares. On 7 ... Qd6? we have a position similar to
Bronstein's 5...Qd6 (see Chapter Seven) but with the important dif­
ference that in the latter, Black keeps his second bishop on the board.
In another early game, Winawer-Englisch, London 1883, Black
played 7 ... Qd6 8 d3, f6 9 Nd2, 0-0-0 10 Nc4 and then discovered he
had nothing better than 10 ... Qe6. But after 1 1 Qg3 (preparing 12 /4),
g5 12 a4, b6 13 Be3, Ne7 White decided against the attractive end­
games available with a timely Qg4 and instead turned to the queenside
-- 14 f3, Ng6 15 Qel, a5 16 Qc3, Bb4 17 Qb3.
I'm going to give the rest of this game because it well illus­
trates many of the Exchange Variation themes that are as true in the
1990s as in the 1880s. It's also a very fine game: 17 ... Qe7 18 g3, h5
19 Kg2!, h4! ? 20 g4 (closing the kingside and banking on the opening
of the other wing), Rhe8 21 Kh2, N:f8 22 c3, Bc5 23 Radl, Ne6 24
Qc2, Bxe3 25 Nxe3, Qc5 26 Nf5, Rd7 27 Rd2, Red8 28 Rfdl, Nf4
29 d4, Qc4 30 d5!, cxd5 31 exd5, Kb8 (3 1 . .. Rxd5?? allows a fork
on el) 32 Qe4, Qxe4 33 fxe4.

With a passed d-pawn, White needs a breakthrough but the

33 ... Kb7
closed natured of the wings forces him to become creative:
34 Rf2, c6 35 c4, c5! 36 Ne3, Re8 37 Ral !, Rc8 38 Ra3, Re8 39
Rb3, Rc8 40 Nc2, Ra8 41 Kgl, Re8 42 Kfl, Ra8 43 Rff3 , Re8 44
Ne3, Rc8 45 Nf5, Re8 46 Rb5!, Ka7.
32 Wjnnin2 Ruy Lopez Exchan2e Variation

47 b4!, axb4 48 aS, Rb8 (48... bxa5 49 Rxa5ch, Kb6 50 Rb5ch and
Rxc5) 49 Rb3, Rc7 50 Rbl !, Rd7 51 Ral, Rbb7 52 axb6ch, Kb8
53 Ra6, Rd8 54 Rxc5, Nxh3 55 RcaS, Kc8 56 c5!, b3 57 c6, b2 and
Black resigns.

B ack to the main line:

8 d3

'This position has generally been considered favorable for

White -- except by Viktor Korchnoi, who has periodically placed his
faith in the Black setup. He has played both 8 ... Bd6 (and . . . Ne7-g6)
and 8...Ne7 here, usually transposing into the same position.
Chapter Two: The Pin (5. Bg4) 33

Black can transpose into the main line with 8...0-0-0 9 Nd2,
f6. However, Alexei Suetin believes 8 ... 0-0-0 to be premature because
of 9 Bd2! with a speedier attack: 9 ...f6 10 b4, g5 1 1 a4, hS 12 Qe3
(evading the line-opening 12 . . . g4 by threatening to enter Black's back
door with 13 Qa7!), Kb8 13 Na3! with a strong attack after the b­
pawn reaches the fifth rank.

8... f6

There is less promise to a plan of 8...Ne7 followed by

kingside castling with .fS because the resulting hole on e4 grants
..

White a wonderful outpost. For example, 8...Ne7 9 Be3, Ng6 10

Nd2, Bd6 11 Nc4, 0-0 12 Rfdl -- to push the d-pawn -- fS?! 13 exfS,
RxfS 14 Qg4, Qf7 15 Na5!, Nf4 16 Nxb7!, Be7 17 Qf3 with a large
advantage (1 7. . .Nxd3 1 8 Qxc6, Nxb2 1 9 Nd6! or 1 7. . . Nd5 1 8 Qe4,
Rb8 1 9 Na5, Rxb2 20 Nxc6, Bd6 21 Nd8! -- Ghizdavu-Liljedahl,
Nice Olympiad 1974).
Note that White can also transfer his QN to g3 via e2 in these
lines: 8... Ne7 9 Nc3, Ng6 10 Be3, Bd6 1 1 Ne2, 0-0 12 Ng3, Nf4 and
now 13 Bxf4 is OK, but better is 13 d4!, e.g. 13 ... Ne6 14 dxe5 fol­
lowed by NfS and Radl (Bernstein-Steiner, Groningen 1946).
When Black plays for .. f7-f5 it is important for White's
.

queen to gain access to the light-colored squares. This was highlighted

in Nunn-Korchnoi, Wijk aan Zee 1985 with a slightly different
move order: 8... Bd6 9 Nd2, Ne7 10 Nc4, 0-0 1 1 Be3, fS?! 12 exfS,
RxfS 13 Qe2, Ng6 14 Nd2, Raf8 15 Ne4, and now 15...Nf4 16 Bxf4
just reaches the kind of good-knight-versus-bad-bishop that tradition­
ally favors White.
Korchnoi, well aware of such niceties, played 15 ... Be7 and
after 16 Qg4!, Qe8 17 Ng3, R(5)f7 18 Rael, c5 19 Qc4! and White
made slow progress (19... Qc6 20 Qe4, Qb6 21 b3, Kh8 22 Qd5 in­
tending Ne4).

9 Nd2 0-0-0
10 Nc4
34 Winning Ruy Lopez Exchan2e Variation

This nicely-placed knight can work against the enemy king

position from a5 or defend his own king (from /5 via e3). White just
has to choose between the attacking plan or the endgame. Depending
on what Black does now, one may be superior to the other. For ex­
ample, a kingside pawn storm by Black will probably be best met by a
queen exchange (Qf5 or Qg4). But a slower procedure by Black
would allow White a free hand on the queenside.

10... h5
Black goes for the mate (. . . g5-g4) while stopping Qg4.
Among other scenarios for him are:

(1) 10...Bd6 invites 1 1 Bd2 and 12 b4. But the best example
of 10 ...Bd6 was Bohm-Kinnmark, Eksjo 1974 which saw White opt
for the ending with 1 1 Qg4, Qxg4?! 12 hxg4, Ne7 13 Be3, c5 14 a4,
Nc6 15 g3!, Rd7 (15... Nd4 16 Bxd4! only helps White) 16 Kg2,
Rhd8 17 g5, Nb4 18 Rael with a slight edge. Both h7 and f6 are
targets.

(2) 10...Bc5 and now the queenside attack will gain a tempo
when White gets a pawn to b4. In Watanabe-Ballmann, World
Junior 1988 Black prospered when White mishandled the attack: 1 1
Bd2, h5 12 a4, g5 1 3 Ne3, Kb8 14 b4, Bd4 1 5 Rahl, Rf8 1 6 b5!?,
cxb5 17 axb5, axb5 and now 18 Rb3, intending 19 Rfbl, c6 20 c4 or
Chapter Two: The Pin (5 Bg4)
.. 35

19 c3, Bxe3 20 Qxe3 is dangerous. But White erred with 18 Nf5?,

Nh6 19 Nxh6, Rxh6 20 Bxg5, Rg6 and although he had regained his
pawn, the game was drawn in 46 moves.

(3) 10 Ne7? is questionable because it makes the advance of

•••

the g-pawn difficult ( 1 1 Be3, g5? 12 Qxf6) and because the alterna­
tive scheme of 1 1 cS and 12 Nc6 is easily stopped by 1 1 Be3 or 1 1
.•. .••

b4, with play similar to Fischer-Kramer in Chapter Nine.

On 1 1 Be3, Qe6 12 a4 Black, having lost a tempo, can still
try to push the g-pawn now that the f-pawn is protected. However,
12 gS? 13 Qg4!, Qxg4 14 hxg4 is a most favorable version of the
.••

endgames we've examined in this chapter. In Hug-Teschner, Berlin

1971 Black was lost soon after 1 4...hS 15 gxh5, Rxh5 16 g3! (by
now a familiar theme), Bg7 17 Kg2, Kd7 18 Rhl, Rxhl 19 Rxhl,
Rh8 20 Rxh8, Bxh8 21 Bc5 and White has all the advantages in the
minor piece ending, particularly because of the light-square holes on
the kingside, e.g. 21 ... Nc8 22 Ne3, Ke6 23 Kf3, Bg7 24 Kg4, Nd6
25 Kh5, Kf7 26 Bxd6!, cxd6 27 Nc4 and 28 Na5.

1 1 Bd2

White could also slide into a slightly favorable endgame with

11 Ne3 and 12 QfS. But accepting the challenge offers greater
chances for a larger advantage.

1 1 ... Kb8

Not the only move here. It secures some safety if and when
lhc a-file is opened. We are following Bronstein-Nei, Tallinn 1971
which continued:

12 b4 g5
13 a4
36 Winning Ruy Lo_pez Exchange Variation

And here White obtained a slight edge with 13 g4 14 Qg3!

.••

(to keep closed the more dangerous file, the h-file). Then 14...gxh3 15
Qxh3, Qxh3 16 gxh3 would leave White with the greater kingside
changes because he can open matters with his f-pawn and is closer for
exploitation with his king.

In the Bronstein game Black developed with 14 ... Bh6 15

Bxh6, Nxh6 but then 16 h4 kept matters closed (16.. j5? hangs the e­
pawn). There followed 16 Qd4 1 7 Rahl, fS!? 18 QxeS, fxe4 and
•••

now 19 Qxd4, Rxd4 20 bS!, exd3 21 cxd3, axbS 22 axbS, cS (better

22 ... cxb5 23 Rxb5, Rxd3 although 24 Rjbl, b6 25 Rxh5, threatening
26 Nxb6, still favors White) 23 b6, NfS 24 RbS.
Chapter Tbree: The Main Line Pin (6 h5)
... 37

CHAPTER THREE: The Main Line Pin (6 h5l ...

Historically, it was the strengthening of this variation for

White that made the Exchange Lopez a popular opening. Today, the
line remains crucial -- but rarely played. If it weren't for the improve­
ments discovered for White in the 1960s, this 6 ... hS would retain the
bad name it achieved in the earlier part of the century. But with those
improvements, it is probably too dangerous for Black to play regu­
larly.

1 e4 e5
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0

In 1948 Reuben Fine described this move as "weak because

Black can safely reply 5 Bg4!". He gave 6 h3, h5 7 d3, Qf6 8 hxg4,
•••

hxg4 9 Ng5, Qh6 10 Nh3, Qh4 1 1 Kh2, gxh3 12 g3, Qh7 and, al­
though White has managed to close the dreaded h-file, the advantage
lies with Black. As we will show, White should not play 8 hxg4.

5... Bg4
6 h3 h5!
38 Winning Witb The Ruy Lopez Exchange Variation

7 d3

Th.is is the key line that Johan Barendregt of the Netherlands

championed virtually alone in the late 1950s and early 1960s -- only
to become disillusioned after a crushing loss at the hands of Rudolf
Teschner at the European Team Championship, Hamburg 1965.
Barendregt believed that was the end of 5 0-0 -- but the postmortems
that followed provided major improvements, and Bardendregt-Tesch­
ner actually turned out to be the beginning of this opening's history.

With 7 d3 White enables himself to meet the pressuring of the

pinned knight (7... Qf6) with reinforcements (8 Nbd2). But before we
examine the main line, it's worth mentioning an alternative procedure
for White.

He can vary with 7 c3 with the idea of accepting the doubled

pawns in the endgame of 7 Qf6 8 d4, exd4 9 cxd4, Bxf3 10 Qxf3!,
•••

Qxf3 1 1 gxf3, 0-0-0 12 Be3 (l 2. . .f5 13 Nc3, Nf6). If Black accepts

the gambit with 10...Qxd4, then 1 1 Rdl, Qc4 12 Bf4 offers plentiful
compensation.

Back in the late 1960s it was believed that 7 Qd3 "forced a

•••

draw" because of the piquant idea of 8 hxg4, hxg4 9 NxeS, Bd6! 10

Nxd3, Bh2ch with a perpetual check. Later it was found that 8 Rel!
significantly improves White's chances because again the endgame is
good for him respite the backwardness of his d-pawn: 8 ... Bxf3 9
Qxf3, Qxf3 10 gxf3, 0-0-0 1 1 Kfl !, Be7 12 Ke2, BgS 13 Na3, Ne7
14 Rgl, Bh6 15 Nc4!, f6 16 h4, cS 17 d3 (De Wit-Van der Sterren,
Amsterdam 1985).

7 •.. Qf6

It makes a little sense for Black to bug out now with 7...Bxf3
8 Qxf3, since the addition of ...h7 -h5 and d2-d3 makes the position
close to ones we saw in the last chapter. However, after 8 ... Qd7
Chapter Three: The Main Line Pin (6 h5)
... 39

White can vary from the previous prescription (9 Nd2, 0-0-0 JO Nc4,
f6 1 1 Bd2) and attack the e-pawn instead with 9 Qg3, f6 10 f4! . Then
1O ...exf4 1 1 Qg6ch offers a nice version of the endgame following
1 1 Qf7, e.g. 12 Qxf7ch, Kxf7 13 Bxf4, Bc5ch 14 Khl, Bd6 15
•••

Nd2, Bxf4 16 Rxf4, Ne7 17 Raft, Rad8 and now instead 18 b4? (an
over-finesse which might have been punished by 18. . . a5 19 bxa5, Ra8
in Kapeyush-Zhivodov, U.S.S.R. 1977) White should simply follow
with 18 es.

On the other hand, if Black continues to offer the bishop on

g4 but does so without 7 ...Qf6, he runs into some trouble. For exam­
ple, on 7 Bc5 White strengthens his refenses with 8 Be3!. The only
•••

serious alternative to the main line is 7 Bd6, which deserves serious

•..

consideration since 8 hxg4, hxg4 9 Ng5, Nh6 gives Black some com­
pensation of a long-term nature (e.g. 1 0 d4, Qe 7 1 1 d5, Bc5 12 g3, 0-
0-0 as in a 1983 Soviet game).

After 7 Bd6 White should follow this simple, solid prescrip­

•••

tion, which is useful in most of the main lines of the Pin Variation: He
should create a flight square for his king at fl, bring his QN to c4 and
eventually threaten to take the bishop on g4.

Black's bishop will then be faced with a choice of a retreat, an

unfavorable exchange on f3 or a dubious sacrifice. For example,
7 Bd6 8 Rel, Qf6 9 Nbd2, Ne7 10 Nc4 and 10 0-0-0 1 1 hxg4!,
••• •••

hxg4 12 Ng5! , with its threat of 13 Qxg4ch is good for White -

whereas 12 Nh2?, Rxh2! 13 Qxg4ch, Kb8 14 Kxh2?, Qxf2 15
Uh6!?, Rh8! is not (Skyrinya-Gara, Riga 1984). Therefore, Black
should back out gracefully with 10 Bxf3 1 1 Qxt'3!, Qxt'3 12 gxf3
•..

with the usual slightly inferior endgame (12... 0-0-0 13 f4; 1 2. . . Ng6 13
K.fI) to come.

White may even take the Bishop immediately at move eight: 8

hxg4, hxg4 9 Ng5 and then 9 ... Nh6 10 Rel (to create a flight
square), Qe7 1 1 Be3.
40 Winning With The Ruy Lopez Exchange Variation

Black can regain his piece with 1 1 ...f6 12 Qd2, fxgS 13

BxgS, Qe8, but White keeps an edge from superior pawn structure.
His king is secure after 14 Bh4 and Bg3/Kgl (Sanchez-Uriarte, San
Sebastian 1993).

8 Nbd2

Once again the queen trade (8 Be3, Bxf3 9 Qxf3) is tempting

but not as ambitious as White can be. He hopes, eventually, to be able
to threaten to take on g4, thereby forcing Black into something un­
pleasant such as 8 ... Be6 or 8...Bxf3.

8... Ne7

The power of the queen trade is revealed by the overly ag­

gressive 8...gS? and the reply 9 Nc4, Bxf3 (else Bxg5) 10 Qxf3, Qxf3
1 1 gxf3 after which the weakness of the kingside pawns is underlined
by 11 ...f6 12 h4!, gxh4 13 f4 or 12 ...g4 13 Kg2, both favoring White.

The move 8...Ne7 fits in with Black's scheme of shifting

pieces on the kingside (... Ne7-g6-f4) while anticipating the further
attack on his e-pawn. Black also has, however, two perfectly reason­
able developing moves with his bishop:
Chapter Three: The Main Line Pin (6 b5)
. 41

(1) 8 Bc5 has the drawback of leaving the e-pawn hanging so that on
•••

9 Nc4, the threat of Nxe5 leads to 9 Bxf3 10 Qxf3, Qxf3 1 1 gxf3, f6

•••

12 f4!, exf4 13 Bxf4, 0-0-0 14 Be3 and White has an excellent end­
game because of the attack against g7 following 14 ...Bxe3 15 Nxe3,
Ne7 16 f4, Rdf8 17 f5! or 16 f5 17 e5, Nd5 18 Rael, Nxe3 19
•••

Rxe3, h4 20 Kf2 and 21 Rgl (Pokojowczyk-Sliwa, Poland 1977).

(2) 8 Bd6 -- protecting the e-pawn and typically transposing back

•.•

into our main line below after 9 Rel, Ne7 10 d4, Ng6.

Note that in many of these lines White feels it necessary to

play Nc4, even if this means trading queens on f3. However, 8 b5,•••

would stop the knight move. A 1976 correspondence game showed

White getting an edge with 9 Nb3, Bd6 10 a4, Ne7 1 1 axb5, cxb5 12
Be3, but he should probably just develop as in other lines with 9 Nb3,
Bd6 10 Be3 (or 9 Rel, Bc5 JO Nb3 and 11 Be3), thereby taking aim
at the newly created hole at c5.

9 Rel

This unassuming move actually serves several defensive pur­

poses. It gives the White King an important escape square at fl (and
also at e2) in case the h-file becomes too hot. The same fl-square can
also be used for one Knight's defense of its brother after a subsequent
hxg4/...hxg4/Nh2 and Ndfl!. There is also the possibility - which
occurs in our main line - of advancing the d-pawn with the tactical
help of the Rook at el.

In general, forcing the Queen trade on f3 is favorable when

White can quickly undouble his f-pawns. If White wants an alterna­
tive to 9 Rel, then 9 Nc4 is a forcing candidate, since 9 Ng6? 10
•••

hxg4, hxg4 1 1 Bg5!, Qe6 12 Nh2 would kill the attack. Some
grandmaster games - including one candidates match game, have seen
9 Nc4, Bxf3 10 Qxf3, Qxf3 1 1 gxf3, Ng6! after which White's pawns
remain doubled and Black stands reasonably well (12 Be3, Be7 13
42 Winning With The Ruy Lopez Exchange variation

Khl, Bf6 14 a4, 0-0-0 15 a5, Nh4 16 Nd2, Ng6 and 1 7... Nf4, Hort­
Spassky, Reykjavik 1977).

9... Ng6
As mentioned earlier, White's plan is to develop refensively
until he can either capture on g4 or force Black to preserve material
equality by moving that bishop. If Black has second thoughts about
the bishop at move nine, he quickly gets the worst of it:

(1) 9...Be6 10 Nfl, g6 (and not JO.. . Ng6?? because of the Q­

trapping 1 1 Bg5!) 1 1 b4! with impending pressure on the e-pawn
from Bb2, e.g. 1 1 c5 12 bxc5, Nc6 13 Be3, Qg7? 14 Rbl, 0-0-0
•••

and now 15 Qcl!, Kb8 16 Qa3, Ka7 17 Rb2! and White soon had a
winning attack in Filipowicz-Trajkovic, Belgrade 1976;

(2) 9 Bxf3 10 Nxf3, 0-0-0 - a comparison with the previous

•••

chapter shows White is better on all fronts 1 1 Be3, Rd7 12 Qd2,

-

Ng6 13 b4!, Bd6 14 a4 and in Tatai-Jean, Monte Carlo 1967 Black

tried to offset the impact of the pawn storm with 14...Nf4, only to
discover after 15 Bxf4, Qxf4 16 Qxf4, exf4 17 b5! that it had not lost
its punch. 17...cxb5 18 axb5, Bb4 19 bxa6!, bxa6 (19. . . Bxel 20 a7!)
20 Rxa6!.
10 d4
Chapter Three: The Main Line Pin C6 h5)... 43

This blocks the keyf2-c5 diagonal and therefore threatens to

capture on g4. Compare with 10 hxg4?, hxg4 1 1 Nh2? and now the
thematic 1 1 Rxh2! 12 Kxh2, Qh4ch 13 Kgl, Bc5! wins for Black.
•••

10 d4, it becomes riskier for Black to leave his

But now, after
Bishop hanging ong4. The natural 10 ... 0-0-0 allows 1 1 hxg4!, hxg4
12 Nh2, Rxh2 after which White kills the attack with 13 Qxg4ch!,
Kb8 14 Kxh2, e.g. 14...Qxf2 15 Rfl, Qxd4 16 Nb3, Qc4 17 Bg5,
Rc8 18 Be3!, Be7 19 Ba7ch, Kxa7 20 Qxc8, Qxe4 21 Rael resigns
(Schneider-Wademark, Eksjo 1978).

10... Bd6

Once more 10 ... Bxf3 is a concession that Black has mis­

played the opening (11 Nxf3, Bd6 12 c3 and now 12 ... N/4 13 Bxf4,
Qxf4 14 Qcl favors White slightly, while castling on either wing runs
into Bg5 or Ng5).

If the6 h5 variation is going to be resuscitated in the future,

•••

the most likely route will be by way of 10 Nf4!?. After 1 1 hxg4,

•••

hxg4 12 Nh2 Black's best is 12 Nxg2! after which 13 Kxg2, Rx­

•••

h2ch! wins for Black - but 13 dxe5!, Qh4 14 Ndfl gives White an
even game. This last variation was pointed out, incidentally, by a
computer.

But 12 g3! is a solid defense that can lead to a serious ad­

vantage for White after 12 ... Qh6? 13 Nh4, g5 14 Qxg4, gxh4 15
gxf4! or 12...Bb4 (to eliminate a defender with . . . Bxd2) 13 gxf4!,
Qh6 14 Kfl ! and the king walks safely away, e.g. 14 Qxf4 15 •••

Nxe5!, Rhlch 16 Ke2, Rh2 17 Nd3, Qh6 18 Nfl! resigns (Adorjan­

Horner, London 1975).

This means that after 10...Nf4 1 1 hxg4, hxg4 12 g3, Black

must regain the piece with 12 ...gxf3 13 Qxf3, Ne6 (so that the N
reaches d4 after a Q-trade) 14 dxe5, Qh6! 15 Nb3, Qh2ch 16 Kfl,
BcS! with considerable complications. In Povah-Corden, Birming-
44 Winning With The Ruy Lgpez Exchange Variatign

ham 1977 White emerged victorious after 17 Be3 (not 1 7 Nxc5, Nd4!
18 Qg2, Qhlch), Bxe3 18 Qxe3, 0-0-0? (18... Qh3ch!) 19 Qf3! but
the practical chances may be more balanced than it appears.
White may improve with 17 Qg2, QhS 18 NxcS and 1hen
18 NxcS 19 Be3, QeS 20 c3 with a small edge. This is better 1han
.••

20 Qf3, 0-0-0 21 QfS, QxfS! 22 exfS, Na4 which turned to Black's

favor in Malyutin-Laguminia, Forli 19'Jl .

1 1 hxg4!

At last White can make this capture.

1 1... hxg4
12 Nh2

White must not open the h2-d6 diagonal voluntarily (12

dxe5 ?, Bxe5! 13 Nxe5 ?, Qh4! or 13 g3, 0-0-0 14 Nxe5, Nxe5 threat­
ening 15 ... Rhlch! 16 Kxhl, Qxj2 and mate along the h-file).

12... Rxh2!

At first it seems that 12exd4 wins because of 13 Ndfl,

•••

Bxh2ch 14 Nxh2, Qh4 or 13 g3, Rxh2!. However, the subsidiary

benefits of White's 9th and 1 0th moves are revealed by 1 2 exd4? 13
•••

eS!! after which White obtains a clear edge in all lines :

(a) 13 ...BxeS 14 Qxg4, Rxh2 15 Nf3!, Rh8 16 NxeS or,
(b) 13 ...BxeS 14 Qxg4, Rh4 15 Qf3!, 0-0-0 16 g3, Rh7 17
Ng4, or,
(c) 13 ... NxeS 14 Ne4, Qh4 1 5 Nxd6ch, cxd6 16 Bf4, fS 17
Bg3 !, QhS 18 Qxd4, 0-0-0 19 RxeS! - Grodzensky-Amzaev, Cor­
respondence 1975.
Chapter Three: The Main Line Pin (6 b5)
.. 45

13 Qxg4!

This is what White missed in the semifinal game

(Barendregt-Teschner, European Team Championship, Hamburg
1965). Instead, he played 13 Kxh2? and after 13 Qxf2 14 Re2,
•••

exd4ch 15 e5, Bxe5ch 16 Rxe5ch, Nxe5 17 Khl (J 7 Qfl or 1 7 Ne4

allow 1 7... Nf3ch! and wins), 0-0-0 and resigned in view of 18 Nfl,
Rh8ch 19 Nh2, g3! .

As the foremost champion - at that time - of 5 0-0, Baren­

dregt said afterwards, "I've lost not a game but a variation. " The
postmortem discovery of 13 Qxg4 changed a lot of minds.

13... Rh4

The endgame forced by 1 3 Qh4 appears balanced and in his

•••

authoritative Russian-language book ( 1982) on the Ruy Lopez, Yefim

Geller gave 14 Qxh4, Rxh4 15 Nf3, Rh8 16 Be3, f6 17 g3, 0-0-0 18
Radl, Rhe8 as roughly equal.

White may obtain more out of the position by way of a timely

exchange on e5, thereby creating the 3-vs.-2 pawn majority on the
kingside that is more "promotable" than Black's 4-vs-3 queenside
majority. In Geller's line, for example, 18 Kg2, Rhe8 19 dxe5 is an
46 Winning With The Ruy Lcwez Exchange Variation

improvement: 19 ...NxeS 20 NxeS, RxeS 21 Bd4, Re6 22 Bc3 fol­

lowed by exchanges of rooks and perhaps bishops.

It is important to appreciate how White wins such a position

because it is the underlying premise of the Exchange Variation - all
things being equal, White's majority beats Black's majority. (After 5
0-0, all things were not equal: Black has two bishops to White's one.)

The position after 22 Bc3 can be won with play such as in

this amateur game, Valkesalmi-Backe, Hallsberg 1980-1 : 22...gS 23
Radl, Rde8 24 f3, Be5 2S Rd3, Bxc3 26 Rxc3, fS 27 Rce3, f4 28
R(3)e2, Rh6 29 Rhl, Rxhl 30 Kxhl, Rh8ch 31 Rh2, Rd8 32 Kg2,
Rdl 33 gxf4, gxf4 34 Kh3, Ral 3S a3, Rbl 36 c3, Rgl 37 Rg2,
Rxg2 38 Kxg2, Kd7 39 Kh3, Ke6 40 Kg4, KeS 41 c4, cS 42 a4!,
b6 43 b3!, c6 44 KgS, bS 4S Kg4! (Black cannot create a passed
pawn), 4S bxc4 46 bxc4, Kd4 47 KfS! (so that Black's c-pawn will
..•

not queen with check), 47 ... Kxc4 48 eS, Kb3 49 e6, c4 SO e7, c3 Sl
e8(Q), c2 S2 Qb8ch, Ka2 S3 Qxf4, Kb2 S4 QeSch. Black Resigns.

It is important to realize that memorizing specific moves in

such an endgame is not as important as the basic principles - paralyz­
ing Black's majority, creating your own passed pawn, neutralizing the
enemy rooks. Aside from the line given above, there is lS RhS and
•••

the procedure 16 c3 and 17 dxe5. In a few games 16 c3, f6 was seen,

with one example running 17 dxeS, NxeS 18 NxeS, BxeS 19 Be3,
Rd8 20 Radl, Rxdl 21 Rxdl, Ke7 22 Kfl, Rh4 23 f3, Rhlch 24
Ke2, Rh2 2S f4!, Bd6 26 Kf3 and White was winning (Kyhle-K.
Karlsson, Rilton Cup 1980-1).

14 QfS!

Barendregt was one of the first to show the strength of this

forcing move, after which an endgame can scarcely be avoided.

14•.. Ne7
Chapter Three: The Main Line Pin (6,,,h 5) 47

Black can preserve the integrity of his pawn structure with

14 ... Rf4 15 Qxf6, Rxf6 but 16 Nf3 (preventing castling or piece re­
treats because of Bg5 or dxe5), 16 Re6 17 dxeS, Nxe5 18 Nxe5,
•••

BxeS 19 c3, 0-0-0 20 Be3 is another version of the favorable ending

mentioned in the note to 13...Rh4.

Typical play would then be 20 f6 21 f3, Re7 22 Radl, Bd6

••.

23 Kf2, c5? 24 c4!, b6 25 b3, Rh8 26 Rhl, Ree8 27 Rxh8, Rxh8 28

f4, Re8 29 Kf3, Kd7 30 Rd5, Ke6 3 1 f5ch, Ke7 32 Bf4 and in
Nurmi-Henao, World Junior 1975 White won soon after 32 g5 33 •..

fxg6, Rg8 34 Bxd6ch, cxd6 35 Rh5, b5 (35. . . Rxg6 36 Rh7ch, Ke6

37 Rb7) 36 Rh6.

15 Qxf6 gxf6

The doubled pawns are temporary since White will eventually

play dxeS. It will then be just as hard for White to create a passed
pawn when Black has his at e5 and f7 as when they are on f7 and g7.
Two examples of thematic play by White:
(1) Slow kingside expansion 16 dxeS, fxeS 17 g3, Rg4 18
-

Nfl (or 18 Kg2, 19 Nf3 and 20 Rhl), 0-0-0 19 f3!?, Rg6 20 Be3, f5
21 Bf2, Rf8 22 Kg2, Rgf6 23 Nd2, Ng6 24 exf5, Rxf5 25 Re4, Rg5
26 Be3, Nf4ch 27 Bxf4, exf4 28 g4, Rd5 29 Re2, BeS 30 Ne4!?,
Hxb2 31 Rhl , Bd4 32 Rh7, Be3 33 g5, Rdl 34 Kh2, Rd7 35 Rh6
48 Winnin2 With The Ruy LQPez Exchan2e Variation

and despite his pawn minus, White eventually cracked through in 8 1

moves thanks to his passed pawn and rooks in O'Donnell-G. Garcia,
Saint John Open 1988.

(2) Play for exchanges - 16 Nf3, RhS 17 Be3, 0-0-0 18 g3,

Rdh8 19 dxeS! (19 Kg2, f5! equalizes), fxeS 20 Kg2, Kd7 21 Rhl,
Rxhl 22 Rxhl, Rxhl 23 Kxhl, cS 24 Kg2, Nc6 2S c3, bS 26 Nd2,
c4 27 f4, exf4? (Black should temporize with . . . Nd8-e6-c5) 28 gxf4,
f6 29 Kg3, Ne7 30 Kg4, Ke6 31 Nf3, cS 32 fSch, Kd7 33 Bf4, Nc6
34 Bxd6, Kxd6 3S Kf4! (35 Kh5, Ne5! 36 Nxe5, Kxe5 37 Kg6, b4!
complicates), 3S b4 36 eSch!, fxeS 37 Ke4, a5 38 Nd2, a4 39
•••

Nxc4ch, Ke7 40 a3!, bxa3 41 Nxa3!?, Kf6 42 Nc4, Ne7 43 Ne3.

Black Resigns - Adorjan-Perecz, Hungarian Championship 1972.
Chapter Four: The Endgame (5 f6 and 7 c5)
... ... 49

CHAPTER FOUR: The Endgame <s r6 and 7 csl ... ...

While there are many different endgames that may arise in the
Exchange Variation, there is one that has become a major sub varia­
tion and the focus of the attention of the world's best players. Al­
though improvements seemed to be appearing regularly during the late
1 970s and early 1 980s, they have not overthrown the basic conclusion
that White stands very well.

1 e4 es
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6
s 0-0 f6

Here 6 d4 is the natural continuation, but 6 h3 has been tried

recently to avoid the main line of 6 d4, Bg4. After 6 h3 Black can try
to punish the novelty with 6 ... gS!? 7 d4, g4, but after 8 hxg4, Bxg4 9
c3, Bd6 10 Nbd2, Qd7 1 1 Nc4 White has a promising game (Gurie­
li-Cuevas, Debreczen 1992).

6 d4 exd4

For alternatives, see Chapter 6.

7 Nxd4

It is important to make Black pay a positional price (i.e. the

weakening of control of d5 and the vulnerability of the cS pawn) for
the trade of queens. The complacent 7 Qxd4 allows Black too easy a
game ( 7... Qxd4, 8. Bd7 and 9 0-0-0).
. . ...
50 Winning With The Ruy Lopez Exchange Variation

7... cs

Black has had varying success with 7 ... Bd6 and 7 ...Ne7 over
the years (See Chapter 5). But I'm sure you saw instantly what was
wrong with 7 ...Bc5?? Play enough blitz games and every so often
you'll win a bishop this way (8 Qh5ch!).

8 Nb3

The knight is more useful on this square than e2. From b3 it

observes the c-pawn (stopping . Ne7, for example) and provides the
. .

marauding move Na5.

8... Qxdl

Rarely seen but not easily refuted is 8...Be6, passing the ba­
ton to White (but not 8. . . Bd6? because of 9 Nxc5, Bxc5 10 Qh5ch).
White does best to remain in a middlegame where the hole at e6 might
be exploited and Rdl will gain control of the d-file. For example,
8...Be6 9 Qf3 and 9...Bd6? 10 Nxc5!, Qc8 11 Nxe6, Qxe6 12 Bf4
actually occurred in a master game, Filipovic-Cvijic, Banja Vrucica
1987.

9 Rxdl
Chapter Four: The Endgame (5 .f6 and 7 c5)
.. ... 51

Here we have a parting of the ways for Black. He can cede

the d-file to White with (a) 9 Be6, or block it with either of his bish­
...

ops (b) 9 Bd6 and (c) 9

... ••• Bd7. He can also insert the finesse move
(d) 9 Bg4 i n order to provoke 1 0 f3 before the bishop retreats to d7.
..•

(a)
9... Be6

It is important for Black to begin clearing his first rank be­

cause Bf4 is coming up fast. Fischer pointed out the inadequacy of
9 b6? -- 10 Bf4 forces 10 Ra7 after which 1 1 Nc3, Ne7 12 a4!
••• •.•

begins to tie Black up.

10 Bf4
52 Winning With The Ruy Lo.pez Exchange Variation

If Whi te waits, Black will develop his other bishop at d6 and

then meet 11 NaS with 1 1 ...0-0-0!.

10••• c4!

Black has no real choice since 10 Rc8 1 1 Nc3 and 12 NdS

.•

leaves him passive and underdeveloped.

1 1 Nd4 0-0-0

This temporarily exploits the ct-file pin and threatens 12 c5.

•.•

Black hopes to be able to advance his queenside pawns with tempo

and use the presence of his king there to shepherd a candidate for
queenhood.

12 Nc3 Bfi
13 Nd5!

This forces a rapid and favorable liquidation because of the

direct attack on c7.

The older move. 13 Nf5 also works well following

13 Rxdlch 14 Rxdl, Ne7 after which 15 Ne3, Ng6 16 Bg3, h5 17
••.

h4! retains serious chances for advantage. (But not 1 7 Ncd5 ?, Bc5!
when Black was better, e.g. 18 Bxc7, Bxe3 1 9 fxe3, Bxd5 or 18 Nxc4,
b5! 1 9 Nxc7, bxc4 20 Nxa6, Bel Pinter-Smejkal, Budapest 1975).
--

For example, 17 b5 18 Ncd5, Bd6 and now instead of 19

..•

Bxd6, cxd6 20 Nf5 which was good enough for an edge in Hodges­
-

Frtiz, 1993 Postal Master Paul Hodges recommends 19 Nb6ch!,

-

and if 19...Kb7 then 20 Bxd6.

Another version of this which well illustrates the dangers to

Black of both allowing and avoiding exchanges is (after
13 Nf5)
13 Re8 (preserving rooks) 14 Rd2!, Ne7 15 Nxe7ch, Bxe7 1 6
••.
Chapter Four: The Endgame (5 ,f6 and 7 c5)
.. .. 53

Radl. Now if Black is not careful he will have no play in the center
because of enemy rook pressure. So, normal play is then 16 Rd8 17
•••

NdS, BxdS 18 RxdS, bS 19 Rxd8ch -- with pleasure !

-- 19...Rxd8,
Kxd8ch, Kxd8 21 Be3, Kd7 22 f4, aS 23 Kf2, a4.
This bishop endgame has been played more than once at the
master level and it is particularly instructive. It can be won by White
even if Black creates a passed queenside pawn. Wrong is 24 g4?
because of 24 gS! and now since 2S fxgS, fxgS or 2S f5 allows a
•••

strong Kd6-eS, the best try is probably 2S eS, gxf4 26 Bxf4 but
•••

after 26 Ke6 27 exf6, Bxf6 Black stands better (Spiridonov­

..•

Spassky, Sochi 1973).

But correct (Grodzensky-Koshil, Correspondence 198S) is
24 Kf3!, Bd6 2S g4 and now even 2S b4 26 h4, h6 27 b3, cxb3 28
•••

cxb3, axb3 29 axb3, cS makes it difficult for Black to use his now­
healthy queenside. The game went 30 Ke2!, Kc6 3 1 Kd3, KbS 32
Bel, Bc7 33 hS, Bd6 34 gS, hxgS 3S eS!, Be7? (35.. .fxe5 36 fxg5,
Be7! 3 7 g6, Bf6 just manages to hold) 36 exf6, gxf6 37 h6!, Bf8 38
fxgS, fxgS 39 BxgS, c4ch (else it's zugzwang already) 40 bxc4, Ka4
41 h7, Bg7 42 Bh6!, Bh8 and now:

43 Bf8!, b3 (zugzwang: 43... Kb3 44 Bc5, Ka4 45 Bd4) 44

Kd2!, Bb2 (zugzwang again: 44... B/8 allows Kcl -bl and eventually
llb2) 4S Bd6, KaS 46 Ba3! Black Resigns (because of 46... Bh8 47
54 Winning With The Ruy Lopez Exchange Variation

Kcl and 48 Bb2). Even though this 13 NfS gives White the best game,
we prefer the main line 13 NdS!

Returning to 13 NdS...

13... BxdS

Clearly the knight cannot be allowed to remain indefinitely on

such a fine square (13... Bd6 14 Bxd6, Rxd6 15 Nf5).

14 exdS RxdS
15 Ne6! Rxdlch
16 Rxdl Bd6

With c7, f8, g7 and, most of all, d8, under attack, Black has
little choice but to give back his extra pawn.

17 Bxd6 cxd6
18 Rxd6

Quite a position for the 1 8th move of a game, yet there is a

logic to the play on both sides. White has regained his pawn with stiff
interest. In Kholmov-Todorecevic, Timisoara 1977. Black played
18 ... Ne7 and after 19 Nxg7 decided to resign rather than play a pawn
Chapter Four: The Endgame (5,,,f6 and 7 c5)
... 55

down -- with three weak pawns islands remaining. (There will be five
pawns)

If the 9 Be6 line is to be revived, it needs a major infusion of

•••

new thinking.

(b)
(After 1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6
S 0-0, f6 6 d4, exd4 7 Nxd4, cS 8 Nb3, Qxdl 9 Rxdl)

9 ••• Bd6

This was the recommended move on the books before

Fischer's revival of S 0-0 at the 1966 Olympiad. A year before that
Svetozar Gligoric had shown how useful the bishop move was in a
game that went 10 Nc3, Be6 1 1 Be3, b6 12 a4, a5!? and then 13
NbS, 0-0-0 14 Nxd6ch?, cxd6 lS Bf4, dS! with rough equality.

10 NaS!

It is important to prevent the smooth coordination of Black's

queenside pieces. The knight not only discourages the QB from
leaving the first rank, but also threatens 11 Nc4, Be7 12 Bf4.
56 Winning With The Ruy Lopez Exchange Variation

10... b5

A cute example of a pin beating a threatened discovery was

Hort-Zhelyandinov, Havana 1967: 10 Bg4? 11 f3, 0-0-0? and now
•••

14 fxg4, Bxh2ch leads to obscure complications. But 14 eS!! wins a

piece (14...fxe5 15 fxg4; 14. .. Bxe5 15 Rxd8ch and 16fxg4).

Fo11owing the game Bagirov-Keres, Moscow 1967 a lot of

attention was devoted to10 Nh6, preparing 1 1 Nf7 and the over­
••• •••

protection of d6. On 11 Bxh6, gxh6 12 Nc4, Be7 13 Nc3, Kti 14

Nd5, Be6 15 Nxe7, Kxe7 16 Ne3 White's position looked very
promising, but Keres managed to draw.
Improvements can probably be found after 16 Ne3, but if
White continues more thematically a move earlier with 15 NaS!,
Black doesn't seem to have anything better than 15...BxdS 16 Rxd5,
Rd8 17 Radl or 15 ... b6? 16 Nc6 -- a concession that Black has the
worst of it.

1 1 c4!

White stops the c-pawn from advancing to a more secure

square and begins to size it up for attack via Be3 and Nd2-b3. The b­
pawn can also be attacked, by Nc3.

Anticipating the latter with 1 1 ... b4 merely enhances the for­

mer: 12 Nd2, Kti 13 Ndb3. And if 13 f5 14 exf5, Nf6, seeking
•..

complications, then 15 Be3, Ng4 16 Bxc5, Bxh2ch 17 Kfl, BeS 18

f3 with advantage for White in Peshina-Romanishin, U.S.S.R. 1968.

1 1. .. Ne7

Black can wait for a move by White's bishop and then play
. .Nh6-f7, in the belief that White won't waste a tempo to play Bxh6
.

once he's already developed it on e3 or f4. For example, 1 1 Be6 12 •.•

Be3. Nh6 13 h3. 0-0-0 14 Nd2, Nf7 seems to coordinate his forces
Chapter Four: The Endgame (5 ,f6 and 7,,,c5)
.. 57

nicely. The problem is that his king position is severely aerated and
after 15 cxb5, axb5 16 Nc6 (and Na7ch) he loses a pawn.

12 Be3

12... f5

This was Lajos Portisch's choice in the influential game

against Fischer at Havana 1966 (see Chapter 9). There are three
alternatives that suggest themselves:

(1) 12...RbS takes aim at b2 but doesn't offset the pressure

against c5, e.g. 13 Nd2, Be6 14 Rael, Kf7 15 Ndb3 and now
15 ...RhdS 16 cxb5, axb5 17 Nxc5, as in Lukin-Kaminsky, U.S.S.R.
1970, is very dangerous.

(2) 12 Kf7 13 Nc3, Rb8 is similar but walks into a well­

•••

timed 14 eS! . Then 14 fxeS 15 Ne4! favors White and 14 ...BxeS 15

.•.

8xc5, Bxc3 (else Nc6 or Nd5) 16 bxc3, Ng6 17 Nc6 is similar to

Fischer-Portisch. We are then following Dvoretsky-lvanov, Moscow
1972 which went 17 ...RaS 18 cxb5, axb5 19 Na7!, Be6 20 Nxb5,
Ra5 21 a4 and White keeps the extra pawn.
58 Winnini: Witb The Ruy LQPez Exchange Variation

(3) 12 ...Ng6!? may be best since it enables Black some

counterplay by way of 13 Nc3, NeS and the attack on c4 as well as on
h2 via Ng4. Then in Reeve-Hamilton, Canadian Zonal 1986,
••.

White forced matters with 14 f4, Nxc4 lS Nxc4, bxc4 16 eS, fxeS 17
fxeS, BxeS 18 BxcS but Black managed to draw following 18 Kf7 ..•

19 Rflch, Bf6 20 NdS, Rd8 etc. Among the improvements for White
is 14 NdS!.

13 Nc3

And here again it is White who may improve with the imme­
diate 13 Nb3!, bxc4 14 NxcS followed by winning the c4-pawn with
his undeveloped knight. For example 14... Rb8 lS Na3, f4 (l 5... Rxb2
16 Nxc4) 16 Bd4, Nc6 17 Na4!, Rb4 18 Nc3, Nxd4 19 Rxd4, Bes
20 RdS! as given by M. Thomas in Spanish Exchange Variation
1 980.

13 ... f4
14 eS!

A key tactical idea in such positions. White liquidates Black's

cS weakness but gains in terms of minor piece scope.

14... Bxe5

The reason the note to White's 13th should be considered as

an improvement on the main line is that here, instead of Portisch's
14 BxeS, Black can reduce his positional deficit by liquidating via
•••

14 fxe3! lS exd6, exf2ch 16 Kxf2, 0-0ch 17 Kgl, cxd6 18 Rxd6,

•••

BfS. Then, after 19 Rel White has only a slight edge.

lS BxcS Bxc3

White's threats included 16 Rel and 16 Bxe7 (16... Kxe7 1 7

Nc6ch).
Chapter Four: The Endgame (5,,.f6 and 7,,,c5) 59

16 bxc3 Kf7

And this is better than Portisch's 16 ...Ng6 17 Nc6, Be6?

which allowed White to begin the harvest with 18 cxbS, axbS 19
Na7!. By connecting rooks, he allows easier defense of his queenside
pawns.

17 Bxe7

'This could be delayed but the White position is attractive

enough after 17 ... Kxe7 18 Rd4 (main threat: 1 9 Reich), gS 19
Retch, Kf6 20 RdS followed by RcS. Everything is going right for
White.

Portisch's experience with 9 ... Bd6 led to a deep examination

of the more natural 9...Bd7. For nearly a decade it became the focus
of the endgame variation.
60 Winning With The Ruy Lopez Exchange Variation

(c}
(After 1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 S 0-0, f6
6 d4, exd4 7 Nxd4, cS 8 Nb3, Qxdl 9 Rxdl)

9 ... Bd7

Black's king readies for castling.

10 Bf4

White goes after the natural target at c7. Yet the only time
Bobby Fischer played this subvariation, in an exhibition match in
Greece, he chose 10 a4!? with the idea of advancing the pawn to a5
and then trying to win the c-pawn with Be3.

There is also something to be said for 10 Be3, b6 1 1 Nc3, 0-

0-0 12 a4 with an attack on the queenside with 13 a5 corning up. An
illustrative game, which may someday have historic significance, ran
12 Re8 13 aS, c4 14 axb6!, cxb3 15 b7ch, Kxb7 16 Rxd7. White
•••

has a positional edge and he ruthlessly enlarged it until after 16... Kc8
17 Rd3, bxc2 18 Rxa6, Bd6 19 NbS, Rxe4 20 Nxd6ch, cxd6 21
Rc3ch, Kd7 22 Ra7ch, Ke6 23 Ra8!, d5 24 Rxc2, Rc4 25 Rxc4,
Chapter Four: The Endgame (5 f6 and 7 c5)
... ... 61

dxc4 26 Kf1 Black was lost -- Shirov-Kamsky, Soviet Cadet

Championship 1986.

10... 0-0-0
1 1 Nc3 Be6

This may seem like a simple loss of time considering his ninth
move, but consider how quickly Black can collapse with the
alternatives:

(1) 11 ...c6?? (suicidal) 12 Na4!, b5 13 Nboch resigns,

Glass-Bauer, Biel 1977.

(2) 1 1 Bc6, which also aims at discouraging the White

•..

Knight from reaching d5 has the serious drawback of allowing 12

NaS!.

(3) 11 ...c4 only temporarily strengthens his pawn structure:

12 NaS, b5 13 Nd5 and now 13 Bg4 14 f'3, Bc5ch 15 Kf1 only
••.

temporarily staves off the problems in defending c7. After 15 Bh5

..•

16 Nc6!, Rd7 17 Nxc7! Black is lost, e.g. 17 Rxc7 18 Rd8ch, Kb7

•••

19 NaSch, Kb6 20 Rb8ch, Kxa5 21 Bxc7ch, Ka4 22 a3 and Ra8

wins. Black varied with 17 Ne7 in Ribli-Medina, Wijk aan Zee
••.

1 972, but lost quickly after 18 Nxa6, Bb6 19 Rxd7, Kxd7 20

Nab8ch!, Ke6 21 Rdl, Nc8 22 Rd5!.

(4) 11 g5 12 Bg3, h5 (for 12 ... Be6 see our main line below)
•••

is a kingside demonstration that fails to divert White's attention from

13 NdS!. Then Black must defend c7 with 13 ... Bd6 14 Bxd6, cxd6
after which15 Rd2, Be6 16 Radl is dreadful for Black (e.g. 16. . . Rhl
17 Nxc5!, dxc5 ? 18 Nb6ch or 17. .. Bxd5 18 Rxd5, Rel 19 /3, Kcl 20
e5!, fxe5 21 Ne4 with a won game, Passerotti-Kovacs, Reggio
Emilia 1978).

12 Rxd8ch
62 Winning With The Ruy Lopez Exchange Variation

There's no reason to pass up this tempo-gainer.

12... Kxd8
13 Rdlch Kc8
14 Nd5

1his continues the series of forcing moves which leaves White

with a small but clear advantage. Since the mid- 1 970s, few masters
have been willing to defend the Black position because it is so sterile.

14... BxdS
15 Rxd5 b6

At various points before this, Black could have kicked the f4-
bishop with g7-g5. White would have retreated to g3 in most cases
•••

e.g. 1 1 ... gS 12 Bg3, Be6 13 Rxd8ch, Kxd8 14 Rdlch, Kc8 15 Nd5,

BxdS 16 Rxd5. Then 16 ...Nh6 may be played, now that 17 Bxh6 is
not possible.
But Black pays a price to get his knight to h6 and then to f7.
After 16...Nh6 17 f3 (J 7 Nxc5 ?, c6!) b6 18 Nd2, Nf7 19 Nc4, Be7
20 Kf1 White has a slightly better version than in our main line be­
cause his knight can go to f5 via e3.

O'Kelly-Westerinen, Malaga 1967 went 20...Rd8 21

Rxd8ch, Bxd8 22 Ne3, Kd7 23 c4, c6 24 Ke2, Be7 25 NfS, Bd8 26
Chapter Four: The Endgame (5. ,f6 and 7 ...c5) 63

b3, Bc7 27 Bxc7, Kxc7 28 g3, bS 29 f4, Kd7 30 h4! and wins. It's
the old story: White can create a passed pawn; Black can't.

16 Nd2

Moves such as Kf1 and c3 also suggest themselves in such

positions. But with Black about to gain time with his knight, White
believes his knight will be more helpful in the next stage than his king.
Once rooks go off the board, it will be the king's turn.

16... Ne7
17 Rd3

Another very basic endgame has arisen. White has a kingside

majority and potential queenside targets. Black has some good outpost
squares for his knight (d4, e5, b4). The existence of only one open file
means a trade of rooks is likely, after which the game may be decided
by which king reaches the fourth rank first. White will most likely
mobilize his by way of Be3, f4 and Kf2.

One thematic game (Browne-Huebner, Wijk aan Zee 1975)

went 17 ... Nc6 18 c3, Nd8?! 19 Be3!, Bd6 20 f4, Ne6 21 g3 with a
growing advantage, e.g. 21. g5 22 eS!, fxeS 23 fxg5 and Ne4 or (as
••

played) 21...Rd8 22 h4, Bf8 23 Rxd8ch, Nxd8 24 Kf2, Kd7 25 g4.

Let's go back and look at the position after 18 c3 (l 7. . . Nc6 18 c3):
64 Winning With The Ruy Lopez Exchange Variation

After the game 18 gS and 18 Be7 followed by 19 Rd8

••. ••• •••

were suggested as alternatives. The latter appeared in a few later

games without much success, e.g. 18 Be7 19 Kfl, Rd8 20 Rxd8ch,
•••

Kxd8 21 Ke2, Kd7 22 Be3, Ke6 23 f4, Bd6 24 Kf3, Kf7 25 h4,
Ke6 26 g4 and White had a mobile majority and Black didn't in
Kuporosov-Kantsler, Sochi 1980.
Black could do little while White took his time to created a
passer: 26... Kf7 27 gS, Ke7 28 Kg4, Kf7 29 Nf3, Ke6 30 hS, Ne7
31 fSch, Kf7 32 Bf4, Nc6 33 Be3 , Ne7 34 Kh4, Nc6 35 Kg4, Ne7
36 h6, g6 37 gxf6, gxfSch 38 exfS, Kxf6 39 BgSch, Kf7 40 f6, Ng6
41 KfS, Bg3 42 Nd2, Bh2 43 Ne4, Nt'8 44 Bh4, Ne6 45 Bg3, Bgl
46 Nd2, Nt'8 and now 47 Bxc7 would win, but better is 47 Nf3!, Be3
48 NeSch, Kg8 49 f7ch, Kh8 50 Ng4! and the BeS mate threat wins.

(d)
(1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0, f6
6 d4, exd4 7 Nxd4, cS 8 Nb3, Qxdl 9 Rxdl)

9 ••• Bg4!?

The main point of this seemingly random sortie is to weaken

the b6-gl diagonal and create the basis for pins and other favorable
tactics.
Chapter Four: The Endgame (5 ,f6 and 7 c5)
.. ... 65

For example, 10 f3, Be6 11 Bf4, c4 12 Nd4 a position we

-

considered slightly favorable for White in (a) when the White f-pawn
was on the second rank - might now be met by 12 ... 0-0-0 13 Nc3,
Rxd4!? (considered better than 13 ...BcS 14 Be3) 14 Rxd4, BcS 15
Be3, Bxd4 16 Bxd4, Ne7 and after 17 Nc6 Black will be only
•••

slightly inferior with good drawing prospects because of the bishops

of opposite color. For example, 17 Rel, Nc6 18 Be3, Rd8 and Black
was comfortable in Santo-Roman-Miralles, French Championship
1991. In the late 1 970s this was thought to improve Black's chances
considerably. Since then, doubt has been cast on that judgment.
Similarly, 12 NaS allows 12 BcS with check - after which
••• -

13 Kfl , Bb6! 14 Nxb7?, Rb8!.

10 f3

There is no reason to reward Black for his last move by

playing something artificial like 10 Rd2 or 10 Rel .

10 .•• Be6

To exploit the diagonal, Black needs an open d-file and con­

trol of c4. After 10 Bd7? 1 1 Bf4, 0-0-0 12 Nc3 he has a position
•••

virtually identical to ones we considered favorable to White in

Subsection (c).
For example, 12 c4 13 NaS, BcSch 14 Kf1 favors White
.••

because of lines such as 14 Bb4 15 Bxc7!. In one of the earlier

•••

9 Bg4 lines White varied with 14 Khl, bS 15 NdS and still did well,
•••

e.g. 15 Bb6 16 Nxb6ch, cxb6 17 Rd6!, bxaS 18 Rb6, Be8 19

•••

Rb8ch or 15 c6 16 b4!, Ba7 17 Nc7, Bb8 18 Ne6, Bxf4 19 Nxc6!

•••

and wins, Ribli-Zuidema, Wijk aan Zee 1972. Pretty!

After 14 Kf1 Black does better with 14 bS although 15 NdS

•••

retains an edge, e.g. 15 ...Ne7 16 Bxc7, NxdS 17 RxdS, Kxc7 18

RxcSch, Kb6 19 b4! (Benjamin-A. Ivanov, U.S. Championship
1993).
66 Winning With The Ruy Lopez Exchange variation

1 1 Be3

Protecting the key diagonal and thereby avoiding the . Rxd4 . .

trick mentioned above. White could also play 1 1 Nc3 and transpose
into the lines below following 1 1...Bd6 12 Be3, b6 13 a4.
This move order (11 Nc3, Bd6 12 Be3) was Fischer's prefer­
ence when he stunned Spassky in the ninth game of their 1 992 re­
match. But 1 1 Nc3, Bd6 12 a4 is inexact because Black can dispense
with b6 and play 12 0-0-0 13 Be3, c4! (14 Na5, Ne7 15 Nd5,
••• ..•

Bxd5 16 exd5, c3 as in Malyutin-lvanchuck, Kramatorsk 1989).

Considering what surprise lies ahead of us - at move 15 - we
might consider 1 1 Nc3, Bd6 12 eS! ?
This tricky idea works fine after 12 BxeS 1 3 NxcS or
•••

12 Bxb3 13 exd6. But it leads nowhere if Black finds 12 fxeS 13

••• •..

Ne4, Bxb3! 14 axb3, Ke7 (Motwani-Agdestein, Novi Bad 1990)

since Black's king turns out to be better placed on e7 than c8.

Black could, however, then insist on the bishops of opposite

color with 1 1 Nc3, c4 12 Nd4, BcS. But after 13 Be3, Bxd4?! 14
Rxd4, Ne7 lS Bf4, Rd8 16 Radl, Rxd4 17 Rxd4, Nc6 18 Rdl, NeS
19 NdS Black's tactics have backfired and White had a considerable
edge in Kupreychik-Gusev, U.S.S.R. 1972.

11 ••. b6

Advancing the c-pawn lacks punch and the White knight

would stand well on cS or d4.

12 a4

We can appreciate White's logic by comparison with other

positions from the Endgame Variation. Once Black is committed to
defending the c-pawn on cS, White begins the flank attack with a2-a4-
a5. This makes much more sense in positions in which queenside
castling by Black is delayed, as in this case because of the open ct-file.
Chapter Four: The Endgame (5 f6 and 7,,,c5)
... 67

12... Bd6

This prepares 13 0-0-0. Because of the risks involved in our

•••

main line below, Black can consider two major alternatives:

(1) 12... aS. As usual, thi s solves one headache - the defense
of cS by creating another, the hole at bS. White would reply 13 Na3!
-

and the threat of 14 NbS requires Black to undergo some gymnastics,

e.g. 13 Bd6 14 NbS, Rd8 lS Rd2! (simple chess - it would be bad
•••

to play 15 Nxd6ch, cxd6 16 Bf4 because of 16. . . d5! as Gligoric

played in a similar position in Section b), Ne7 16 Radl, Nc8 17 Bf4,
Ke7 and now White has a choice of 18 Bxd6ch, cxd6 19 Nel l with
the knight headed for dS, or 18 eS, BxeS 19 BxeS and 20 Nxc7.
18 Bxd6ch was the scheme in de la Villa-Gild. Garcia, Sala­
manca 1989, a nicely played White victory that should be a model for
White play against the ...aS defense: 18 Bxd6ch, cxd6 19 Ncl, Rhf8
20 Ne2, fS 21 Nec3, Rd7 22 Rel , Kf7 23 exfS, BxfS 24 NdS, Be6
25 Nbc7, BxdS 26 NxdS, Re8 27 Rde2, Rxe2 28 Rxe2, Rb7 29 c4!
(stopping b5 and thereby sentencing the rook to passivity), Na7 30
...

Kt'2, Nc6 31 f4, Nb4 32 Nc3, Rd7 33 Rd2, Ke6 34 g4, hS!? (else
White wins with Kf3, f5ch, Kf4, etc) 3S gxhS, KfS 36 Kf3, Nc6 37
NhS, Nb4 38 Rg2!, dS 39 Rg6, dxc4 40 Nd6ch and wins.
68 Winning With The Ruy Lopez Exchange Variation

(2) 12 Kfi. By foregoing queenside castling, Black avoids

•.•

some of the possible problems with d-file pins. Then 13 Nc3, Bd6 14
a5, c4! lS Nd4, bS should hold the fort for Black, e.g. 16 NfS?, BxfS
17 exfS, Ne7 18 g4, hS with a clear advantage to Black, or 16 Nce2,
Ne7 17 Bf4, BeS 18 BxeS, fxeS 19 Nxe6, Kxe6 20 f4, Rhd8 21 fSch,
Kfi and even after a double-trade of rooks, White could not win in
Rosentalis-Psakhis, Klaipeda 1988.
But White improves quite a bit with 16 f4 and the idea of
16 ... Ne7 17 eS!, fxeS 18 fxe5, BxeS 19 Rflch, Bf6. Now 20 Rael,
Bd7 21 BgS, Rae8 22 Re5 with a threat of Ne4 did well in a 1986
Soviet game, Kosikov-Bezman. But White can improve further with
the immediate 20 BgS!, e.f. 20 Rad8 21 Bxf6, gxf6 22 Ne4!, Rxd4
..•

23 Rxf6ch and White won in Sherzer-P. Littlewood, Lloyds Bank

1987. No better is 20 Rhf8 21 Ne4, NdS because of 22 Bxf6 and 23
•••

Ng5ch.
A simpler way of gaining an edge is for White to exploit the
temporary vulnerability of c7 by playing 13 Bf4 (instead of 13 Nc3).
Then 13...Rc8 14 Nc3 is an improved version of the previous lines.
And the tactical refenses of 13 Bxb3 14 cxb3, Bd6 lS Bxd6, Rd8
•••

(16 Na3!, cxd6 1 7 Nc4) or 13 Bd6 14 Bxd6, Rd8 lS a5!, c4? (16
•••

axb6, cxb3 1 7 bxc7, bxc2 18 cxd8(N)ch, Ke8 1 9 Na3!, cxdl(Q)ch 20

Rxdl, Kxd8 21 Bf8ch!) both fail.

(3) 12 ...c4 is a relatively new idea that enables Black to castle

faster, with 13 Nd4, 0-0-0. The most dangerous way of handling it is
14 Nc3, Bfi lS aS, bS 16 NdxbS!?, which gives White plenty of
compensation for the sacrificed piece after 16 Rxdlch 17 Rxdl,
•••

axbS 18 NxbS and now: (a) 18 ...Be8?? 19 Na7ch! and 20 Rd8; (b)
18 ... Ne7 19 a6! (this also works against 18... Bb4), Nc6 20 a7, Kb7
21 Rd7!, and (c) 18...Be7! 19 a6, Be8 20 a7, Bc6 21 Ral, Ba8 22
Ra4 (Malishauskas-Jandemirov, Katowice 1993.)

13 as 0-0-0
14 Nc3 Kb7
Chapter Four: The Endgame C5,,,f6 and 7 c5)
... 69

Black has little choice once he castles queenside (14 . . . c4 15

Nd4 attacks the bishop).

15 eS!

One of those elegant little moves that make the Exchange

Variation so surprisingly strong. Now 15 ... fxeS? 16 axb6, cxb6 17
Ne4! creates a fatal pin on the d-file (1 7... Bcl 18 Rxd8, Bxd8 19
Nbxc5ch!). Similarly, 15...Bxb3 loses to 16 exd6, Bxc2 17 Rdcl,
Bg6 18 dxc7!.

15... Be7

In the mid- 1 970s this was considered adequate for Black

but...

16 Rxd8 Bxd8

17 Ne4!

White seems to be winning with 17 axb6, cxb6 18 Ne4, Bxb3

19 Nd6ch?! (or 19 cxb3, Bel!), Kc6 20 cxb3, Ne7 21 Rxa6 but then
21 NdS! led to equality in Adorjan-Ivkov, Skopje 1976.
...
70 Winning Witb The Ruy Lo.pez Exchange Variation

White can also try 18 BxcS, Bxb3? 19 Bf8!, but this is short­
circuited by the superior 18 Nh6, and Black will regain a pawn (19
•.•

Bd4, Nf5).

17 •.. Bxb3

After the sensational Fischer-Spassky rematch game,

annotators called17 Ne4 a novelty. Actually, it was Spassky's reply,
17 Kc6??, that was new. After 18 axb6 he was lost (1 8. . . cxb6 1 9
•••

Nbxc5!, bxc5 20 Rxa6ch o r 1 9 . Bc8 20 Nxa6).

. .

18 cxb3

Now 18.. .fxeS allows a strong 19 axb6, cxb6? 20 Nd6ch and

21 Nf7. B ecause White has delayed the exchange on b6, Black cannot
play 18 ...Bc7. and on 18 ... Be7 White invades with his rook 19 --

axb6, cxb6 20 Rdl, fxeS 21 RdS, Bf6 22 Rd7ch or 20 Kc7 21 e6!. •••

Biyiasas­
One of the few games to reach this position was
Vukovic, Bar 1980, in which Black responded 18 ...fS 19 Rdl ! , Ne7.
White then obtained a winning position soon after 20 NgS, Nc6 21
axb6, BxgS 22 BxgS, Kxb6 23 Rd7!.

By a roundabout route w e once again have a situation of

healthy White kingside pawns working against weakened Black
queenside pawns and the game ended with Black losing the queening
race that began with 23 ...Re8 24 Rxg7, RxeS 25 Rxh7, Relch 26
Kf2, Rbl 27 h4!.

In the aftermatch of the 1 992 Fischer-Spassky miniature,

18 Ne7 was suggested as an improvement. However, 19 axb6, cxb6
•.•

20 exf6, gxf6 21 Rdl leaves Black with serious kingside problems,

e.g. 21 NfS 22 Bf2, Kc6? 23 g4!
•••

Incidentally, I've been asked a number of times about where

Fischer got 17 Ne4. There are several possibilities, including: (a)
Chapter Four: The Endgame (5 f6 and 7,,.c5)
... 71

Bobby discovered it himself; (b) Fischer saw the Peter Biyiasas game
published in Shakhmatny Bulletin or elsewhere; (c) Biyiasas showed
him the game in 1 9 8 1 when Fischer was staying at Biyiasas' San
Francisco apartment, and (d) Fischer saw the first edition of the book.

I don't know the answer. Peter Biyiasas says he has no recol­

lection of analyzing the Exchange Lopez with Fischer - but he did
keep his old tournament bulletins in the apartment in places Fischer
could find them. But I suspect the real answer is (a).
72 Winning With The Ruy Lopez Exchange variation

CHAPTER FIVE: Other 6 ... exd4 Lines

(1 e4, e5 2 Nf6, Nc6 3 Bb5, a6 4 Bxc6, dxc6
5 0-0, f6 6 d4, exd4 7 Nxd4)

If Black refrains from forcing an exchange of queens (7. . . c5)

after 6 ... exd4, then he must live through a middlegame in which e6
and to some extent f5 are vulnerable squares. Although common in the
early days of the Exchange Variation revival ( 1 966- 1 980), these
variations have largely disappeared from master play.

Nevertheless, they cannot be said to be refuted, and anyone

hoping to use 4 Bxc6/5 0-0 on a regular basis must expect to face
these positions from time to time. The two main seventh move alter­
natives are (a) 7 ... Ne7 and (b) 7 ... Bd6.

Other moves fail to coordinate Black's pieces swiftly enough

to prevent White from amassing too much strength in the center e.g.
7...g6?! 8 Be3, Nh6 9 h3, Nfi 10 Nd2, Bd7 (There is no better
square for this bishop. Against 1 0 Bg 7 White would continue as in
...

the game) 11 f4, Qe7 12 Qf3, Bg7 13 Rael and White was obviously
better in Grodzensky-Lapygin, Correspondence 1975.
Chapter Fiye: Other 6
... exd4 Lines 73

(A)
1 e4 es
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6
S 0-0 f6
6 d4 exd4
7 Nxd4 Ne7

Originally popularized by Lajos Portisch in the late 1 960s,

Black's last move controls f3 while retaining the possibility of forcing
an endgame with a later cs. (By comparison, 7. Bd6 relinquishes
... ..

the endgame option.)

8 Be3

By no means the only way of handling the position (8 Na3 is

among good alternatives). The bishop belongs on e3 to safeguard the
diagonal leading to his king in case he advances his f-pawn. He also
watches the cS square in case Black should advance.

8... Ng6
74 Winning With The Ruy Lopez Exchange Variation

Black doesn't have much choice here since 8 cS allows 9

..•

QhSch!, Ng6 (9. . . g6? J O Qxc5) 10 Nb3! and Black is pressured on

both wings. An illustration of how quickly he can go downhill is
10 c4 1 1 NaS, Qe7 (or 1 1 . . . b5 12 /4 and 13 /5) 12 Nc3, Qe5 13
•..

QxeS, Nxe5 14 NdS, Bd6 15 f4, Nf7 16 Nxc4, Kd7 17 Radl, Re8
18 eS!, fxeS 19 fxeS, Nxe5 20 Nxd6, cxd6 21 Nb6ch and Black
could have resigned in Kaplan-Balinas, Lone Pine 1979.

9 Nd2

Again there is plenty of choice. White retains the option of

planting his queen at hS or his Knight at fS. He avoids the immediate
9 f4 because the endgame created by 9 ...cS! is more double-edged that
usual because of the now-weakened White e-pawn.

9 ••• Bd6

This is where Black has expected to post the bishop all along.
It controls some key dark squares here, whereas on e7 it allows White
a free and easy initiative, e.g. 9...Be7 10 QhS!, 0-0 1 1 NfS, Rf7 12
f4, Bf8 13 Rael (a typically strong kingside assault force is assem­
bled), Ne7 14 Nd4, fS 15 N(4)f3!, h6 (else 1 6 Ng5) 16 NeS, Rf6 17
Rdl, Be6 and now 18 g4!, fxg4 19 fS! gave White a terrific game in
Hansen-Harandi, Haifa 1976.

Generally, the White queen is so strong on hS that Black

needs access to f7 for his own queen for defensive purposes. In fact,
the immediate 9 Qe7 and 10 Qf7 has been tried even without the
.•. •..

provocation of a QhS. That maneuver, however, seems a tad prema­

ture, e.g. 9 Qe7 10 NfS, Qf7 1 1 Nb3!?, b6 12 Rel, cS 13 Qd2,
•••

Bb7 14 Radl, NeS 15 Bf4, g6 16 Bxe5!, fxe5 17 Ne3 with a slight

pull for White (Suradiradja-Mabbs, Hastings Challengers 1973-4).

10 Nc4
Chapter Fiye: Other 6,,,exd4 Lines 75

This follows Fischer's play in his game with Unzicker, Siegen

1 970 (see Chapter Nine). The immediate 10 Nf5, or Karpov's 10 c3,
0-0 11 Qb3ch, Kh8 12 Nf5 have also achieved slight advantages for
White.

A good example of the immediate occupation of f5 was Bed­

narski-Dobrynski, Lublin 1973: 10 Nf5, 0-0 (1 O. . . Bxf5 1 1 exf5 and
12 Qh5ch is disconcerting) 11 Nxd6, Qxd6 (or l l . . . cxd6 12 Nc4!
hitting at d6 and b6) 12 f4, b6 13 Qh5, Bd7 14 Rf3! and White had
the kind of attack that plays itself. Black has missed his opportunity
to halt him on the kingside with f6-fS.
...

The game ended shortly after 14...Rfd8 15 f5, Nf8 16 Qh4,

Be8 17 Rg3, Rd7 18 Rfl, Kh8 19 Nc4!, Qe7 20 e5!, fxe5 21 Qxe7,
Rxe7 22 Bg5, Rd7 23 Nxe5, Rd5 24 f6!, Ne6 (24. . . g6 25 j7) 25 f7.

10... 0-0

11 Qd3!

This improves on previous play which failed to take advan­

tage of Black's various weaknesses. The queen can now go to b3 to
pressure the queenside and c4-g8 diagonal. And by overprotecting the
e-pawn, it also prepares an advance of the f-pawn.
76 Winnin2 With The Ruy Lopez Exchao2e Variation

1 1... Ne5

If Black waits for White to build up with Radl or Rael and

f2-f4 he may leave himself confined to his first and second ranks.
Note that Black's queen bishop has no future unless he decides to
push his c-pawn. But 11 cS 12 Nf5 or even 12 Nb3 is difficult to
...

defend for Black

12 Nxe5 Bxe5
13 f4 Bd6

And now with 14 fS! White severely restricts Black's pieces

and gains the space for a kingside or central attack. For model play,
see the Fischer-Unzicker game (Chapter Nine).

(b}
(After 1 e4, e5 2 Nf'3, Nc6 3 Bb5, a6
4 Bxc6, dxc6 5 0-0, f6 6 d4, exd4 7 Nxd4)

7 •.• Bd6

The bishop move, at first glance, appears weak because of 8

Nf5, Bxf5 9 exf5 followed by queen and/or rook checks. However,
Chapter Fiye: Other 6 exd4 Lines
... 77

9...Qd7 10 Relch, Ne7 1 1 Qh5ch and now 11 ...g6 12 fxg6, 0-0-0 is

actually a dangerous pawn sacrifice. And 1 1 ... KfS! is even better
because White has no other way of defending f5 other than the highly
weakening 12 g4. Then 12 ... NdS should be at least equal for Black.

In general White has to time the Nf5 move accurately in such

positions.

8 Qh5ch

White could proceed as in (a) with 8 Be3, 9 Nd2 and 10 Nc4,

but here he has the more forceful method of loosening up Black's King
position. In the absence of a knight at e7, Black must use his pawn
structure to block the check.

8... g6
9 Qf3

The queen should retreat to this square in order to support

Bf4. For example, Black can win a pawn with 9 ...Bxh2ch 10 Kxh2,
Qxd4 but then 1 1 Rdl, Qe5ch 12 Bf4 renders a nice initiative
(12. . . Qxb2 13 Bxc7!, Qxal 14 Rd8ch, Kf7 15 Qa3! and wins, or
12. . . Qe7 13 Qg3). Best in this line is 1 1 ...Qc4! and then 12 Bf4, Qf7
when 13 Qb3, Qxb3 14 axb3, Be6 15 Bxc7 is only a small edge for
White.

9... h5!?

Black needs to take some risks since otherwise he will not be

able to complete his development normally (9 Ne 7? 1 0 Qxf6). After
...

9 ... Qe7 10 Bf4, Bd7 11 Nd2, 0-0-0 12 Nc4 or 9 ... c5? 10 Nb3, Be6
1 1 Rdl, Qe7 12 e5 Black's troubles are only beginning.

10 Qd3
78 Winning With The Ruy Lopez Exchange variation

Clearly not 10 Bf4?? because of 10 ...Bg4. The text may

threaten 1 1 eS, Bxe5 12 Qxg6ch and also prepares the Fischer cen­
tralization plan of Be3, Nd2-c4 and Rael along with the advance of
the f-pawn.

10... Qe7
1 1 f4

Now that the Black queen can block checks at f7, 1 1 e5?
lacks point.

Black has some freedom here but also many weaknesses. He

can develop with Nh6-f7 and ... Bd7/...0-0-0 but should never quite
•••

equalize, e.g. 1 1 ... Nh6 (this covers f7 in case White pushes his e­
pawn) 12 Nd2, Bd7 13 Nc4, Bc5 14 Khl followed by Be3 with an
edge.
Chapter Six: Mod.em Main Line (5..,f6/6...Bg4) 79

CHAPTER SIX : Modern Main Line

(5 ... f6/6... Bg4)
Before Fischer sprang his improved version of the Exchange
Variation on a largely unsuspecting chess world in 1966, there was
virtually no mention of the defense that has since become the standard
reply. Perhaps this is because it appears so illogical -- Black
undermines his light squares at move five, then prepares to trade off
the bishop that defends them at move six. And, perhaps worst of all,
this defense leaves White with the key choice at move seven -­

whether to play a double-edged middlegame or a slightly superior

ending.
Nevertheless, a long line of defenders -- including Garry
Kasparov, Vasily Smyslov, Svetozar Gligoric, Yefim Geller, Jan
Timman and Boris Spassky -- have shown this is the variation that
truly tests the Exchange Variation.

1 e4 es
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4
80 Winning With The Ruy Lopez Exchange Variation

7 dxe5

There is no definitive word on whether this endgame is essen­

tially better than 7 c3. It is, however, easier to learn and easier to
play.

7 ... Qxdl

Black has no choice: 7 ...fxeS? allows White to get out of the

pin (8 Qxd8ch) and win a clear pawn (Nxe5).

8 Rxdl fxeS

This is more exact than 8 Bxf'3. That exchange doubles

..•

White's pawns before he can reinforce f3 with Nbd2. But it prema­

turely releases tension in the position. After 9 gxf'3, fxeS White should
not rush into the undoubling 10 f4? because his center is not yet
secure: 10 Nf6 11 f'3, Bd6! or 1 1 Nc3, Bd6 12 fS?, g6! with advan­
•••

tage to Black. Not much better is 1 1 Nc3, Bf6 12 f'3, exf4 13 eS after
13 BcSch 14 Kg2, NhS lS Ne4, Be7 and Black keeps his pawn
•••

(Berg-Shirov, Kerteminde 1991).

However, 10 Be3 is a well-timed preparation. If Black then

develops his knight at e7 he is vulnerable to the advance -- 10 ... Ne7
11 f4! after which the sacrificial attack of 1 1 ...exf4 12 Bxf4, Ng6 13
Bxc7, BcS 14 Nc3, 0-0 was ably handled by lS Bg3, Rae8 16 Rd7 in
Kagan-Westerinen, Siegen 1970.

For 10 Be3, Bd6 1 1 Nd2, Ne7 see Fischer-Rubinetti in

Chapter Nine. Perhaps Black's best after 8...Bxf'3 9 gxf'3, fxeS 10 Be3
is 10 Bd6 11 Nd2, Nf6 (or 1 0. . . Nf6 and l l . . . Bd6) when 12 Nc4, 0-
•••

0-0 13 Rd3, Rd7 14 Radl, Rhd8 holds White's edge to a minimum.

But it is an edge nevertheless (e.g. 15 Bg5 with the idea of f3-f4 or of
Bh4-g3).
Chapter Six: Modern Main Line (5 ,f6/6,,,Bg4)
.. 81

9 Rd3!

The only way to hope for an advantage is to get out of the pin
and, at the same time, safeguard the kingside against Bxf3. Clearly
••.

9 Nbd2, 0-0-0! , creating a new pin, is the wrong sequence.

9 ••. Bd6

Once more there is a temptation to exchange on f3. The main

reason for this is that 9 ... Bxf3 10 Rxf3, Nf6 turns the focus of atten­
tion from Black's e-pawn to White's. After 11 Nd2, 0-0-0 12 Kfl,
BcS or 11 Nc3, Bb4! 12 BgS, Bxc3 13 bxc3, Rf8! -- keeping the king
in the center -- Black has survived the worst. For more on this, see
Fischer-Smyslov in Chapter Nine.

The correct reply to 9 ...Bxf3, therefore, is 10 gxf3! after

which the lack of a target for his pieces dooms Black to passivity:

(1) 10...BcS 11 Nd2, Ne7 12 Nb3, Bd6 (or 2. . Bb6 13 Be3)

.

13 NaS!, 0-0-0 (l 3. . . b5 14 Nb7) 14 BgS with a nice edge (Adorjan­

Hernandez, Graz 1972). White can retreat the bishop effectively to
g3 to attack e5 or trade down into a good-N-vs.-bad-B ending.

(2) 10 Nf6 11 Nd2, bS 12 a4!, Bd6 13 Nb3 (or 13 . 0-0 14

•.. ..

Rc3! with strong queenside pressure), cS 14 NaS, c4 lS Rdl with

advantage to White according to Gipslis. If Black allows the White
knight to go to c4, he faces more problems: 10 ...Nf6 1 1 Nd2, Be7 12
Nc4, Nd7 13 Be3, 0-0-0 14 Kfl , Bf6 lS Ke2, Rhe8 16 NaS, Nb8 17
Radl, Rxd3 18 Rxd3, b6 19 Nc4, cS 20 c3 with a slowly improving
game (Szmetan-Sully, Nice 1974). The White N heads for d5 or g4
in such endings.

(3) 10...RdS 11 Rxd8ch, Kxd8 12 Be3, Ne7 13 Nd2, Ng6

14 Nc4, Be7 1S Kg2!, Rf8 16 Kg3!, b6 17 Rdlch, Kc8 18 Kg4! and
the king illustrates what happens when Black gives up his good
bishop.
82 Winning With The Ruy Lopez Exchange variation

The game Iskov-Villeneuve, Stockholm 1974-S saw the king

march to glory on the light-colored squares: 18 ...Rd8 19 Rxd8ch,
Kxd8 20 f4, exf4 21 Bxf4, Nxf4 22 Kxf4, Bd6ch 23 es, Bes 24 Ne3,
Ke7 2S Ke4, Ke6 26 f4, g6 27 fSch, gxfS 28 NxfS, Bf2 29 Ng7ch,
Kfi 30 NhS, Bgl 31 h3, Kg6 32 Nf4ch, KgS 33 Ne6ch, Kh4 34
Nd4!, KhS (else the pawn queens) 3S Nf3, BcS 36 KfS, Be7 37 Ke6,
Bes 38 Kfi Resigns.

(4) 10 ...Bd6 1 1 Nd2, bS (11.. .Ne l transposes to Fischer­

Rubinetti) and now either 12 Nb3, cS 13 NaS, c4 14 Rdl, Ne7 lS
Nb7!, Kd7 16 f4 with advantage (Tisdall-Yilmaz, Thessaloniki
1988) or the immediate 12 f4, exf4 13 Nf3 and White regains his
pawn favorably, e.g. 13 Ne7 14 eS, BcS lS Bxf4, NdS 16 Bg3, h6
•..

17 Rel, 0-0-0 18 Re4 -- Sorokin-Jovcic, Correspondence 1977.

10 Nbd2

White's basic plan is to exploit the inability of the enemy to

refend his e-pawn with any remaining pawn. The attack on eS can be
mounted with Nc4 and a queenside fianchetto. There is also the
prospect of a queenside raid with Nb3-aS or with c2-c4-cS.

10... Nf6
Chapter Six: Modem Main Line (5..,f6/6...B�4) 83

When a player as significant at Garry Kasparov varies from

the known paths of theory, the rest of us take notice. During his match
with Jan Timman at Hilversum 1985, the future world champion
played 10 ... bS, a move that had a somewhat questionable reputation.
White should then continue 11 b3, Ne7 12 Bb2, Ng6 13 g3 (to keep
the knight off /4), 0-0 14 Kg2 and now in earlier games White had
pushed Black slowly back after 14 ... Rad8 lS h3, Bc8 16 Nfl, Rde8
17 h4, hS 18 Ne3, BcS 19 NgS, Be7 20 Radl, Red8 21 Nf3! or
14...Rf6 1S h3? (15 Ngl directly makes more sense), Bd7 16 Ngl, cS
17 Ne2, Raf8 with pressure along the f-file (Nunn-Portisch, Wijk
aan Zee 1989). The right procedure for Black's rooks is not at all
clear and he can even play 14 Rae8 lS c4, Rf6 16 Rel, cS 17 cxbS,
••.

axbS 18 Ba3, b4 19 Bb2 and Nc4. In fact, the latter line follows a
1 977 game by a 14-year-old named Garry Kasparov.

But in the Timman game, he varied with 14...cS and then lS

c4, Rab8!. White has a broad choice here and probably best is 16
Rel followed by Ba3 to pressure cS, or 16 h3 followed by the repo­
sitioning of the knight to c3 via gl and e2.
However Timman chose an intriguing alternative, a pawn
sacrifice: 16 a4!?, bxc4 17 Nxc4, Bxf3ch 18 Rxf3, Rxf3 19 Kxf3,
Rxb3ch 20 Ke2. He had ample compensation because of the weak
enemy pawns, bad bishop and prospect of Kd3-c4, following 20 ••• Kf7
21 Kd2, Ke6 22 Bc3, Ne7 and now 23 f4, exf4 24 Kc2 virtually
forced 24 ... Rxc3ch 2S Kxc3, fxg3 26 Nxd6, cxd6 27 hxg3 after
which Black managed a draw. Improvements abound in such a line
(23 Kc2!?)

A new idea, in place of the above 11 b3 (after 1 0... b5) is 11

b4! with the idea of trading White's b-pawn for Black's e-pawn. After
11 Nf6 12 Bb2, Nd7 White can force matters with 13 c4!, Bxf3 14
•••

Nxf3, bxc4 lS Rc3, Bxb4 16 Rxc4, Bd6 17 Rxc6, 0-0 18 Nd2 and
White won in Dvoretsky-Southam, World Open 1991.

11 b3!
84 Winnini: With The Ruy Lupez Exchange Variation

This is now preferred to 1 1 Nc4, 0-0! because it has been

found that 12 Nfxe5, Be2! 13 Re3, Bxc4 leads nowhere for White.

11_. 0-0-0

In the 1 970s Black used to castle on the other wing and de­
velop the QR at e8. But after a number of setbacks it was concluded
that this did not generate enough counterplay, e.g. 1 1 ••• 0-0 12 Bb2
and now:

(1) 12 ... Bxf3 13 Nxf3, Nxe4 liquidates the center but after 14
Re3, Nf6 15 Nxe5 all things are equal -- except the pawn structure.
And that means White is better (15... Rae8 16 Rael, Bc5 1 7 R(3)e2,
Nd5 18 Nd3, Rxe2 19 Rxe2, Nf4 20 Nxf4, Rxf4 21 Kfl, Rf5 22 J3,
Bd6 23 h3, Kf7 24 Re4, Ra5 25 a4, Rc5 26 c4, b5 27 Bc3, Rf5 28
Ke2, Bg3 29 Bb4, h5 30 Re7ch and White won in Bordonada-Coo­
per, Nice 1974).

(2) 12...Rae8 13 h3, Bh5 14 Rel, b5 15 Nfl, Bg6 16 Ng3,

Nh5?! 17 Nxh5, Bxh5 18 Nh4, Bc5 19 Nf5 with an edge in
Matanovic-Nemet, Yugoslav Championship 1975.

12 Bb2
Chapter Six: Modem Main Line (5 f6/6,,,Bg4)
... 85

Here there are two stem games to follow. Neither one shows
model play but they illustrate how White can maneuver and maneuver
with little risk -- and then make progress at a later stage.

The first, Velikov-Lengyel, Reggio Emilia 1979-80, saw

Black firmly defending e5 while White maneuvered and maneuvered
until there was breakthrough: 12 ... RheS 13 Rdl, b5 14 h3, Bd7 15
Rel, Nh5 16 Nfl, Nf4 17 Rddl, Re7 18 Ne3, Rf8 19 Nh2, Bb4 20
c3, Bc5 21 Rd2, Ng6 22 b4, Bxe3 23 Rxe3, Be6 24 Bel , Bc4 25
Rel, Rd7 26 Rxd7, Kxd7 and the bishops of opposite color proved
to be insufficient to draw following 27 Be3, Ra8 28 a3, a5 29 Nf3,
axb4 30 axb4, h6 31 h4, Ke6 32 h5, Nf8 33 Nh4, Kf6 35 Nf5 1 -0.

In the second, Lautier-Dolmatov, Marseille 1988, two fu­

ture stars once again found that the liquidation of the center can be
dangerous for Black: 12 ...Bxf3 13 Nxf3, Nxe4 14 Nxe5, Rhf8 15 f3,
Nc5 16 Rddl, Ne6 17 Kfl, g5 18 Rel, Nf4 19 Re4, Bc5 20 g3, Ng6
21 Ke2, Rf5 22 Nd3, Bd6. Here 23 Re6 is a worthwhile bid for pro­
gress.

As played, White preferred to defend his kingside pawns by

bringing his king back to g2. Play then continued: 23 Kf2, Rdf8 24
Re3, c5 25 Kg2, b5 26 c4, Kd7 27 Rel, Kc6 28 Rc2, b4 29 Rce2,
a5 30 Nf2!, a4 31 Ng4 with a slight edge that eventually won in 65
moves.
86 Winning With The Ruy Lo_pez Exchange Variation

CHAPTER SEVEN: Bronstein' s 5 0d6 ...

Virtually ignored twenty-five years ago when the Exchange

Variation was reborn, this defense of the Black e-pawn, recommended
by former World Championship Challenger David Bronstein, has be­
come, far and away, the most popular of Black's ambitious plans.
Black acknowledges that he is willing to play an endgame -- but on
his terms. And if White plays passively during moves 7- 1 5 , Black's
pieces and pawns will so quickly take over the board that White won't
realize it until it's too late.

1 e4 e5
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6
5 0-0 Qd6

Black defends his attacked pawn with a queen but also pre­
pares to exert pressure along the half-open file by castling queenside.
The queen may end up on g6 where it attacks e4 and looks in the
direction of g2.
Chapter Seven: Bronstein's 5 Qd6
... 87

The move S Qd6 was the only defense to S 0-0 that did not
•..

turn out to favor White in Kasparov and Keene's first edition of Bats­
ford Chess Openings. They gave 6 d4, exd4 7 Nxd4, Bd7 8 Be3 and
now not 8...Qg6 9 Nd2, 0-0-0 1 0 Qe2, which favors White according
to a frequently quoted Soviet analysis by Aivars Gipslis -- but 8...cS!
9 Nb3, BbS! as leading to even play.
Immediately going after Black's bishop with 6 Na3, Be6 7
NgS may turn out to be White's best policy. This scored a major suc­
cess in lvanchuk-Short, Amsterdam 1994 after 7 ... Ne7 8 d3, Ng6 9
Nxe6, Qxe6.
White was able to paralyze the enemy queenside while en­
gineering f2-f4 by way of 1 0 Nc4, BcS 1 1 Be3, Bxe3 12 Nxe3, 0-0
and now 13 a4, Rad8 14 aS and later Qel-c3-cS.

6 d3

This leads to positions reminiscent of the delayed fom1s of the

Exchange Lopez (e.g. 1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Ba4, Nf6 5 0-
0, Bel 6 Bxc6, dxc6 7 d3, Nd7 8 Nbd2, 0-0 9 Nc4, f6). White retains
the choice of where to open the game -- on the kingside with f2-f4, in
the center with a delayed advance of the d-pawn, or on the queenside
with an all-out advance of the a- and b-pawns, usually after Black has
castled there.

White can also try to delay d2-d3 profitably. For example, 6

Na3, threatening 7 Nc4, is often met by 6 ... Be6. But 7 NgS, threaten­
ing 8 Nxe6, and if 7 ...Bd7, then 8 d3! leads to better play than in our
main 6 d3 line, e.g. 6 Na3, Be6 7 NgS, Bd7 8 d3, f6 9 Nc4, Qe7 10
Nf3, Qf7 11 Be3, Be6 12 Nfd2 and 13 f4 as in Christiansen-Pinter,
Manila 1992.

6 ..• f6

This solves the problem of defending eS once and for all , and
enables Black to launch a kingside pawn storm with a subsequent
88 Winning With The Ruy Lopez Exchange variation

... g7-gS. Other moves such as 6... Bg4 or 6... Be6 will almost certainly
transpose into the main line that follows.

But there are a few wholly independent policies for Black:

(a) 6...g6 7 Nbd2, Bg7 justifies the queen's presence on d6

but loses time after 8 Nc4, Qe7. Then 9 b4!? followed by 10 a4 and
Ba3, whether or not Black takes the b-pawn, is a promising method of
exploiting the absence of Black's bishop from its natural diagonal.
Another, more traditional, policy is 9 a4, Bd7 10 b3, cS 1 1
Bb2 after which Black has to play 11 ...f6 after all . In Ostos-Silva,
Students Olympiad 1977, White correctly turned to the kingside with
12 Ne3, Bc6 13 Nd2, Nh6 14 Ndc4, aS lS Qd2, b6 16 Rael, 0-0 17
f4 and doubling along the f-file.

(b) 6... cS hopes to bring the remaining Black knight to c6 in

two additional moves. (In the comparable Doubled Delayed Exchange
Variation position, Black ends up playing . . . Nf6-d7-b8-c6 !) But after
7 Nbd2 Black has to take time out to protect against 8 Nc4 and loses
time e.g. 7 ... Qe6 8 Nc4, f6 9 Nfd2 followed by 10 f4 with a good
position for White.

(c) 6...Ne7, however, is a major alternative since it allows

Black to defend the e-pawn and discourage f2-f4 with ... Ng6. Its
drawback is that Black is just one tempo away from stopping d3-d4:
7 Nbd2, Ng6 8 Nc4, Qf6 9 d4!, exd4 10 Qxd4 or 10 BgS, Qe6 1 1
Qxd4. White i s better able to play endgames now that he has an extra
tempo or two of development, e.g. 10 Qxd4, Qxd4 1 1 Nxd4, Bd7 12
a4, 0-0-0 13 Be3, cS 14 NfS!, Be6 lS Nd2, Ne7 16 Nxe7ch, Bxe7 17
f4 and in Gaprindashvili-Tuzofsky, Tbilisi 1974 White won after
making inroads on both wings (1 7. . .f5 18 RJ2,, b6 19 a5, Kb7 20 Rel,
Bh4 21 g3, Bf6 22 e5, Bel 23 Nf3, Rd5 24 Rd2 etc.).

But since 7 Nbd2, f6 8 Nc4, Qe6 is better than appearances,

some players will prefer 7 Be3, which promotes a central advance
better. For example, 7 ... Ng6 8 Nbd2, Be7? 9 d4! with advantage. In
Chapter Seven: Bronstein's 5 Qd6
... 89

Benjamin-Kamsky, U.S. Championship 1991 Black was in trouble

after 8 cS 9 Nc4, Qe6 10 NgS, Qf6? 1 1 QhS!, Bd6 12 f4!, exf4 13
•..

eS! with a powerful attack that should have won.

7 Be3

Once Black has solidified e5 with the f-pawn, there is no fu­

ture in a queenside fianchetto. On the contrary, there is a premium on
White pushing his d-pawn to the fourth rank. If Black does nothing to
prevent that advance, he can quickly obtain a bad game, e.g. 7 ... Be6 8
Nbd2, Qd7?! 9 d4, exd4 10 Nxd4, Bfi 1 1 Qe2, cS 12 N(4)b3, b6
13 Radl, Ne7 14 Nbl!, Qc6 15 Nc3, Ng6 16 f4, Bd6 17 eS! with a
strong attack (Adorjan-Timman, Hastings 1973-74). Similarly,
8 ... Ne7 9 d4, Ng6 10 Qe2 or simply 10 dxeS and 1 1 NgS.

7 ••• Bg4

Stopping 8 d4. The chief alternative is 7

cS, which serves
••.

that function more directly. Then 8 Nbd2, Be6 9 Nc4, Bxc4 leads to
drawish endgames (but not 9. . . Qc6 J O Nfd2, Ne7 I I a4, b6 12 f4,
exf4 13 Bxf4, Ng6 14 Bg3, Bd6 15 Bxd6 which was excellent for
White in Schneider-Romanishin, Buenos Aires 1978).
90 Winning With The Ruy Lopez Exchange Variation

Better than 9 Nc4 is 9 Qe2 followed by Nc4, when ...Bxc4

does not lead to an endgame, or perhaps Nb3 and an attack on c5 in
preparation for an advance of the ct-pawn. For example, 9 Qe2, Ne7
10 c3, Nc6 11 Nb3, Bxb3 12 axb3, Rd8 and in Kir. Georgiev­
Short, Plovdiv 1984, White was ready for a push in the center with
13 Rfdl, Qe6 14 d4, cxd4 15 cxd4, exd4 16 Nxd4, Nxd4 17 Bxd4,
Bd6 18 Qc4 but after 18 .. Kf7 he had little to show. A better strat­
.

egy, once Black has relinquished the right to castling long, is 13 Radl
and if 13 ... Qe6, then 14 Nd2 and 15 f4.

Fischer had considered similar positions back in the 1 960s

but believed that queenside castling was best for Black, e.g. in this
last line 10 ...0-0-0 (instead of JO. . Nc6). But in that case the liquida­
.

tion that follows 1 1 d4, cxd4 12 cxd4, exd4 13 Nxd4, Nc6 14 Nxc6,
Qxc6 favors White's heavy pieces after 15 Rfcl!.

A drastic example of this was Dvoretsky-Romanishin,

Vilnius 1975: 15...Qb5 16 Nc4, Bc5 17 a4, Qb4 18 Bd2, Qb3 19
Ba5!, Bxc4 20 Rxc4, Bd6 21 Rael, Rd7 22 g3, Kb8 23 R(l)c3,
Qa2 24 Qc2, Rc8 25 Rb3, Be5? and now 26 Rxb7ch! won since
26... Kxb7 27 Rb4ch is a quick mate.

8 Nbd2
Chapter Seven: Bronstein's 5 Qd6
... 91

8... Ne7

A variety of other moves were tried in the 1980s and found

failing:

(a) 8 ... Qd7, once considered the main line, loses time and
also sets up a problem because of the question-putting 9 h3!. Then the
attacked bishop cannot retreat to h5 without incurring the tactical
wrath of Nxe5!. So 9 h3, Be6 10 d4! would follow and after
10 exd4 11 Nxd4, Bf7 12 Qe2, 0-0-0 13 Rfdl Black is about to
•..

lose more valuable time.

In fact, he is one or two mistakes away from disaster, e.g.
13 c5 14 N(4)b3, Qc6 15 Na5, Qe6 16 Ndb3, Rxdlch 17 Rxdl,
•••

Qxe4? 18 Qd2!, Qe8 19 Bxc5 and in the junior game Malishauskas­

Khalifman, Leningrad 1983 Black collapsed with 19 ...Nh6 20 Bxf8,
Rxf8 21 Qb4!, b6 22 Nc6!, Bxb3 23 Rd8ch.

(b) 8 c5 has the disadvantage of creating a target for a line

•••

opening advance of the White b-pawn. This comes about, for exam­
ple, via 9 Qe2, b6 10 Nc4, Qd7 1 1 Rfdl, Ne7 12 h3, Bh5 13 Rd2,
Bf7 (13... Nc6 14 Nfxe5!) 14 c3, Qe6 15 b4 -- Sherzer-Torre, New
York Open 1988.
A more direct method is, 9 a3, which when followed by Qe2
and c2-c3, will make b2-b4 readily available. Black may then try to
pressure the d3 pawn -- but that is another two-edged blade: 9 a3, b6
10 Qe2, a5 11 c3, Rd8 12 h3! , (Black's many light-square weak­
nesses make this offer sound) Bxf3 13 Qxf3, Qxd3 14 Rfdl, Qb5 15
a4! and in Egin-Kaidanov, Tashkent 1985 Black died quickly --
15 ...Qxb2 16 Nc4, Qb3 17 Rxd8ch, Kxd8 18 Qe2, Bd6 19 Qd3, f5
20 f3, f4 21 Bf2 resigns.

(c) 8 ... 0-0-0 is the most natural response but it invites a

queenside pawn storm with 9 Rbl (also good is 9 Qbl e.g. 9 .. Qd7 J O
.

b4, g5 1 1 a4, Nh6 12 b5!, cxb5 1 3 axb5, Qxb5 1 4 Qa2 and 1 5 Rjbl
as in Petrushin-Kaidanov, Bryansk 1984), Ne7 10 b4 and now
10 g5 11 a4, Ng6 12 b5! favors White's faster attack while 10 Ng6
••• •••
92 Winning With The Ruy Lopez Exchange Variation

1 1 h3, Be6 12 a4, Qd7 13 d4! can lead to a debacle like Wahl-Lane,
Malmo 1985-86 which ended prenily with 13 ... Bd6 14 b5, axb5 15
axb5, exd4 16 Ral, Kb8 17 Nxd4, cxb5 18 Ra8ch!, Kxa8 19
Qalch, Kb8 20 Qa7ch!! resigns (20... Kxa7 21 Nc6ch and 22 Ral
leading to mate).

9 h3

A good move to insert here since 9 ... Bh5 creates tactical

problems again after 10 Nc4 (10. . . Qd7 1 1 Nfxe5; J O. Qe6 11 Ng5).
. .

Another good idea, taking advantage of Black's last move, is

9 b4, and then 9...Ng6 10 h3, Be6 11 d4 (Dolmatov-Smagin,
Erevan 1988).

9... Be6
10 d4

Now that the pin on f3 has been broken, this break in the
center makes sound policy. On 10...exd4 11 Nxd4, Bf7 White has a
considerable edge in time.

10... Ng6
Chapter Seven: Bronstein's 5 Qd6
... 93

White has more than the glimmer of an initiative, a more

ambitious center and better development. He can force matters now
with 1 1 dxeS, NxeS 12 Nd4, Bt7 13 f4 as in Anand-Molobyev,
Frunze 1987, which continued 13 ...Nd7 14 N(2)f3, 0-0-0 lS Khl (15
c3!) cS 16 NfS.

Appearing less forceful -- but perhaps much more promising -

- is 1 1 c4. A key idea then is to meet 1 1 ... b6 with 12 cS!, bxcS 13
dxcS, Qd3 14 Qa4, QbS lS Qc2!, leaving the enemy queen some­
what out of play. If Black avoids the exchange on cS, with 12 cS,
Qd7, then 13 Qc2 offers White an obvious positional edge because of
the attack on c6.

In short, Bronstein's S... Qd6 remains the most dangerous of

the major sub-variations of the Exchange Lopez. When masters begin
to have doubts in the 5 ... f6/6 ... Bg4 of Chapter Six they turn to
s. .Qd6.
.
94 Winning With The Ruy Lopez Exchange Variation

CHAPTER EIGHT : The Irregulars

5 Qe7?!, 5 Qf6?, 5 Bd6! ?, 5 Ne7, 5 Be7?
••. ... ... ... ••.

(1 e4, es 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0)

Although White makes a threat with his fifth move -- attack­

ing the enemy e-pawn -- Black has much more choice than the main­
lines we've considered would suggest. There are five reasonable ir­
regular sub-variations. Three of the irregulars defend the e-pawn in
what seems to be a discreditable manner: (a) Smyslov's 5 Qe7, (b) •••

5 Qf6?! , and the 1 9th Century favorite (c) S

... Bd6!?. There are also
•••

two that don't protect the e-pawn at all, (d) Keres' S Ne7 and (e)
••

s. .Be7.
.

None offers the opportunities for equality that S f6 or •••

S Qd6 do, but in their own way, each can be dangerous. Failing to
•••

open the center at the right moment -- a common error by White in the
Exchange Lopez -- can allow Black an excellent game.
Chapter Eii:ht: The Irrei:ulars 95

(a}
SMYSLOV'S 5 Qe7 ...

Like a lot of moves in the Exchange Lopez, this means of

defending the e-pawn looks like a beginner's mistake. But it is harder
to crack than expected -- and White should not expect a very large
advantage.

1 e4 eS
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 Qe7

This is a modified version of Bronstein's 5 ••• Qd6 with a num­

ber of minor differences which, adding up, make Black's fifth move a
bit questionable. Nevertheless, it has been played every so often by
former World Champion Vasily Smyslov.

6 d4!

This is the most natural way to obtain an advantage against

5 Qe7. In a closed position, the queen's presence on e7 would not be
...
96 Winning With The Ruy Lopez Exchange Variation

such a hindrance because rapid development is not at such a high

premium.

Unlike the comparable position in the Bronstein line, 6 d4

does not permit Black to force a quick endgame or allow him rapid
queenside castling. Instead it invites the White queen to take a pow­
erful centralized post at d4.

6... exd4

Trying to maintain e5 as a strong point, with 6 J6?!, is

••

questionable. Then 7 b3 is a promising idea, and so is 7 dxe5, fxe5 8

Qd3, preparing 9 Qc3.
For example, 8 ... Bg4 9 Nbd2, Nf6 10 Qc3, Nd7 11 Nc4 led
to a good endgame for White after 1 1 Bxf3 12 Qxf3, 0-0-0 13 Qg4!,
•••

Re8 14 Rdl, Qe6 15 Qxe6 in Malyutin-Katishonk, Moscow 1991.

7 Qxd4

Here 7 Bg4 is supposed to be bad because 8 NeS, Rd8 9

..•

Qc3, Bc8 creates problems for Black in completing his development.

However, he has no real weaknesses and his absence of kingside
development may be offset by an exchange of queens (. . . Qb4/. . . Qc5).

In practice, it has been more common for White to play 8 Bf4

(rather than 8 Ne5) and permit the doubling of his f-pawns. After
8...Bxf3 9 gxf3, Nf6 10 Nc3 we have Dvoretsky-Smyslov, Odessa
1974 which favored White clearly after the exchange of queens:
10 NhS 1 1 Bg3, Rd8 12 Qe3! (not 1 2 Qa7, Nxg3 13 hxg3, Qb4!),
.•.

Nxg3 13 hxg3, QcS 14 Radl, Qxe3 15 Rxd8ch, Kxd8 16 Rdlch,

Kc8?! 17 fxe3, g6 18 e5!.

A short time later, Gaprindashvili-Bokhosian, Tbilisi 1974,

saw 11 ...Nxg3 (instead of 1 1 . . . Rd8) 12 hxg3, QcS 13 Radl, Qxd4
14 Rxd4, BcS but the transition to a minor piece ending (15 Rd3, Ke7
16 Rfdl, Rhd8 1 7 Na4!, Ba7 18 b3, Rxd3 19 Rxd3, Rd8 20 Rxd8,
 97

Kxd8 21 Nb2, Bd4 22 Nd3, Bc3 23 f4, b6 24 Kg2, c5 25 g4, Ke7 26

Kf3) only helped White to win in 57 moves.

Another, perhaps superior procedure for Black is 10 RcS 11

•••

Radl, QcS with the idea of trading queens immediately. But in

Rabinovich-Steinsapir, 3rd WorJd Correspondence Cup 1981-87,
White maintained his slight edge for another 50 moves to force resig­
nation: 12 Ne2!, Be7 13 BeS, Qxd4 14 Nxd4, Nd7 lS Bg3, NcS 16
Rfel, 0-0 17 f4, Rfd8 18 b3, Rd7 19 NfS, Bf8 20 Rxd7, Nxd7 21
Ne3, Re8 22 f3, BcS 23 13!, etc. The target pawn at c7 won't go
away.
In these lines the Black queen impedes the smooth develop­
ment of his pieces on e7 and, unlike in the Bronstein line, it cannot
easily go to an aggressive square such as g6.

7 ••• Qf6

In the Bronstein line we avoided an early advance of the

White d-pawn. Now if Black tries to reach that kind of position with
7 ... Qd6, White will keep wood on the board and harass the enemy
queen -- 8 Qe3!, Ne7 (8. . . Qc5 9 Nc3, Qxe3 is consistent, but gives
White too much of a lead in development) 9 Nc3, Ng6 and in Ligter­
ink-Kurajica, Amsterdam 1976 White obtained a slight edge with 10
e5, Qe6 1 1 Rel, Be7 12 Nd4, Qd7 13 e6.
8 a4
98 Winning With The Ruy Lopez Exchange Variation

Also good is the 8 Qxf6 and 9 Bf4 endgame. We are follow­

ing Vasiukov-Zaitsev, U.S.S.R. 1971 which went 8 Bg4 9 BgS?,
•••

Qxb2 10 Nbd2, Qb4! and White was searching for compensation for
his missing pawn.

8... Bg4
9 eS
But this is stronger. Now 9 Qg6 10 NgS, BfS 11 Qb3!
•••

(11 . . 0-0-0 12 Nxj7, Be6 13 Nxh8) grants White the initiative, ac­
.

cording to Geller. Although this has not been tested at the grandmas­
ter level, it appears sufficiently convincing to discourage most players
from trying s . Qe7.
. .

(b}
THE THIRD QUEEN MOVE
(After 1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0)

s... Qf6

This queen defense of the e-pawn seems to be the weakest of

the trio (5... Qd6 and 5. . . Qel). It was apparently used successfully by
Max Harmonist, a 19th century master-ballet dancer. It's greatest
success was Schallopp-Harmonist, Frankfurt 1887: 1 e4, eS 2 Nf3,
Nc6 3 BbS, a6 4 Bxc6, dxc6 S 0-0, Qf6 6 d4, exd4 7 eS?!, Qg6 8
Nxd4, Bh3 9 Qf3, Bg4 10 Qg3, 0-0-0 11 c3, BcS 12 Be3, Ne7 13
h3?, Bxd4! 14 Bxd4, hS! lS hxg4, hxg4 16 f3, Qd3 17 Qxg4ch, NfS
18 Na3, Rxd4! 19 cxd4, Rh4 20 QgS, Qxd4ch 21 Rf2, QxeS 22 f4,
Qe4 23 Nc2, f6 24 Qg6, Rh6 2S Qf7, b6 26 Rel, Ng3!! 27 Qf8ch,
Kb7 28 Rfe2, Qxf4 White Resigns.

Later the line was tried a few times in master events in the
early days of the Exchange Variation revival, but the queen's vulner­
ability to harassments (Bg5, etc.) has caused it to virtually disappear.
Yet the lack of grandmaster games with S...Qf6 indicates there may
be plenty of unexplored territory here.
Chavter Eight· The Irregulars 99

6 d4!

Once again, White can exploit the queen position only with
vigorous play. If he allows Bg4 (e.g. 6 d3 ?!, Bg4!) Black has
•••

weathered the worst of it.

6 .•. exd4

Of course, 6 Bd6 7 dxeS, BxeS 8 NxeS, QxeS 9 Nc3 or

•..

6...Bg4 7 dxeS, Bxf3 8 Qxf3 surrenders his bishop-versus-knight

compensation. (And in the last line 8 exj6! is even better for White,
winning a pawn.)

However, Black can improve after 6 Bg4 7 dxeS with

...

7 Qg6!. In Blokhin-Aguilar, World Open 1994 Black regained his

•••

pawn with gocxl play following 8 Qd3, Rd8 9 Qe3, BcS! 10 Qf4 (10
Qxc5, Bxf3) , Ne7 11 Be3, Bxe3 1 2 Qxe3, QhS and Ng6. .•.

If White is going to give back the pawn, he should try 8 e6!,

e.g. 8 .fxe6 9 NeS! ; 8...Qxe6 9 NgS.
.•

7 Bg5 Qg6
100 Winning With The Ruy Lopez Exchange Variation

This was Black's intention when he played his fifth move. The
queen stands nicely on g6, a safe square (since White has no light­
squared bishop) and a menacing square (since Black's light-squared
Bishop can now go to g4 or even h3 with effect).

8 Qxd4 Bd6

Black can reduce the enemy initiative through exchanges -

8 Be7 9 Bxe7, Nxe7 10 Ne5, Qd6. But White generally benefits
••.

from an exchange that denies Black a pair of bishops. An example of

this endgame occurred in Poutinen-Takemoto, Teeside 1974 and fell
into White's favor: 11 Rdl, Qxd4 12 Rxd4, f6 13 Nc4, Kf7 14 Nc3,
Be6 15 Ne3, Rhd8 16 Radl, Rxd4 17 Rxd4 and White owns the d­
file and is ready to push his kingside pawns.

Another method of simplifying is 8 ... Bg4 9 Nbd2, Qd6. Here

White probably does better by keeping the heavy wood on the board:
10 Qe3!, Qc5 11 Nd4 and now 11 .f6 12 Bf4, 0-0-0 13 N(2)b3, Qb6
••

14 Qg3 was a 1972 Soviet game, Fershter-Kaminsky won by Black

only after several major errors by White.

9 Nbd2

Also good is 9 Qe3 with the idea of 10 Nh4/11 Nf5 and of

stopping Black's kingside development (9. . . Ne7 JO e5!?, Bb4 1 1 a3,
Ba5 12 Qc5!).

Then on 9...Nf6 the fork 1 0 e5 is not conclusive (10. . . Nd5)

but 10 Bxf6, forcing 10 ... gxf6, is promising, e.g. 11 Nc3, Rg8 12 g3,
Qh5 13 Rfel, Bc5 14 Qf4, Bg4 15 Nh4 and 16 Nf5 is Doroshk­
evich-Nikolaevsky, Tbilisi 1966-67.

9 ••• Be6
Chapter Eight: The Irregulars 101

Black cannot prevent Nc4 indefinitely. One faulty line runs

9 Bh3?! 10 gxh3!, f6 1 1 eS, Be7 12 Qe4! with a favorable end­
.••

game.

lO Rfel Ne7
1 1 Nc4!

An improvement suggested after Gipslis-Romanishin,

U.S.S.R. 1972 saw Black equalize against 1 1 c4?, Qh5 12 Qe3,
Ng6.
1 1... Bxc4
Not 1 1 ...0-0-0 12 Nxd6ch and 13 Qa7!, marauding the
queenside. If he stays in the center with 1 1 ...RdS, White exploits the
king position with 12 Nxd6ch, cxd6 13 e5, dxeS 14 Qc5 or 13 ...d5
14 Bxe7! and 15 Qb4ch.
12 Qxc4 f6

And now 13 eS!, fxeS 14 Bxe7, Bxe7 15 RxeS or 13...fxg5

14 exd6, cxd6 15 Re6 favors White distinctly.
102 Winning With The Ruy Lopez Exchange Variation

(c}
THE 19TH CENTURY FAVORITE 5 ... Bd6
(After 1 e4, eS 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0)

This 5...Bd6 is one of the oldest of the 5 0-0 su bvariations. In

fact, it was the sole line to appear in Chess Openings: Ancient and
Modern by Freeborough and Ranken ( 1 893) -- one of the first books
to take 4 Bxc6/5 0-0 seriously. It is a most natural method of defend­
ing the e-pawn and is the method Black usually employs in the De­
layed Exchange Variation (1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Ba4, Nf6
5 Bxc6! ?, dxc6 6 0-0 and now since 6. . ./6 is impossible, 6... Bd6 is
seen).

During the late 1 970s, 5 ... Bd6 in answer to 5 0-0 suddenly

became a focus of major attention after it was found that White gets
very little from the most aggressive, line-opening plan (8 e5 below).
But other methods of exploiting Black's weakenesses were found and
today 5... Bd6 has receded in popularity.

5... Bd6

6 d4
Chapter Eight: The Irregulars 103

Anything else would reward Black for his slightly stodgy fifth
move. Now Black must exchange pawns since 6 f6? 7 dxe5, fxe5 8
•••

Nxe5! once again wins a pawn (8. .. Bxe5 9 Qh5ch). And even Free­
borough and Ranken knew that 6 Bg4 7 dxe5, Bxf3 (forced, since
•••

7. . . Bxe5 8 Qxd8ch costs a piece) 8 Qxf3, Bxe5 favors White's supe­

rior pawns and development. They regarded 9 Qb3 as correct but 9
Nd2 and 10 Nc4 makes more sense.

6... exd4
7 Qxd4

Best. After 7 Nxd4 Black can calmly complete his develop­

ment (7. . Ne7) and castle.
.

7 ... f6

This move may not be artistic but it accomplishes Black's

immediate goals. Black can afford the slight weakening of e6 in many
Exchange Variation lines because White no longer has a light-squared
bishop to exploit it. Moreover, the attempt to use White's initiative in
the center -- with 8 e5, dxe5 9 Nxe5 -- tends to liquidate the situation
in Black's favor following 9 Qf6.
••.

8 b3!

Fischer had dismissed 5 Bd6 in My Sixty Memorable

•••

Games, with the comment that here 8 Nbd2 and 9 Nc4 gets White a
clear edge. but with the White bishop temporarily blocked, Black can
reply 8 Nh6! and reinforce d6 with 9 Nf7.
... •••

The text, with its ideas of Na3-c4 or Ba3xd6 has greater

punch.

8... Qe7
104 Winning With The Ruy Lo_pez Exchange Variation

This is now reguarded as more promising than the former

favorite, 8 Be6 (and much preferable to 8...Bg4?, which allows 9
•..

e5!, Bxf3 JO exd6, Bd5 1 1 Reich or 9. .. Bxe5 JO Qxg4, Bxal 1 1

Qxg7). Black bides his time with 8 Qe7, preparing to castle on either
.•.

wing. He also stops 9 Na3 and enables himself to meet 9 Nbd.2 with
9 Nh6.
•••

On 8 Be6 White can mount the assault on d6 with 9 Ba3,

•..

e.g. 9...Ne7 10 Bxd6, Qxd6 11 Nc3, 0-0 12 Na4, Bg4 13 Qc4ch

with a slight edge (lvkov-Westerinen, Hamar 1979). Similarly,
9 Nh6 10 Bxd6, cxd6 11 c4!, 0-0 12 Nc3, Nf7 13 Qe3 and Nd4
.•.

(Kagan-Zwaig, Hasti ngs 1976-77). Even though Black's queenside

majority has been repaired by . cxd6, he has created a weak ct-pawn
. .

and surrendered half of the ct-file.

9 Nbd2

White obtains less than nothing from 9 Rel because of

9 Bg4 10 eS?!, Bxf3! and the e-pawn is pinned.
•••

9... Nh6

The attempt to attack a knight when it reaches c4 with 9...

-

Be6 10 Nc4, BcS 11 Qd3, Rd8 12 Qe2, Qf7 - turns out to be a

dubious pawn sacrifice after 13 NaS!.

In Timman-Piket, SWIFT (rapid) 1992 play continued 13 ...

Bb6 14 Bd2, Ne7 15 Nxb7, Rb8 16 NaS and Black never achieved
equality.

10 Nc4 Nf7
11 Nxd6ch cxd6
12 Bf4
Chapter Eight: The Irregulars 105

Once again Black has corrected his pawn structure at the cost
of incurring some holes and other weaknesses, particularly the d6
pawn. In van der Wiel-Pinter, Rotterdam 1988 White maintained a
slight but significant pull with 12 ... Be6 13 Qb6, 0-0 14 Nd4, Bc8 15
Rfel, Qd8 16 Qxd8, Rxd8 17 Radl.

(d}
KERES' CHALLENGE
(After 1 e4, e5 2 Nf3, Nc6 3 BbS, a6 4 Bxc6, dxc6 5 0-0)

5... Ne7
106 Winning With The Ruy Lopez Exchange Variation

Around 1970 Paul Keres issued this analytical challenge to

the Exchange Variation: He argued that Black can temporarily sacri­
fice the e-pawn by closing the e-file with his knight. If White lhen
keeps the center closed (6 d3) Black will development smoothly with
•.•Ng6. If White, on the other hand, opens matters with 6 d4, exd4 7
Nxd4, then 7 ... c5! 8 Nb3, Qxdl 9 Rxdl, Nc6 or 9 Ng6 is an obvi­
•••

ous improvement over Chapter Four's endgames.

To obtain an advantage, therefore, White may have to accept

the challenge and take the e-pawn.

6 Nxe5

One almost unexplored but promising alternative is 6 Na3

with the idea of 6 f6 7 d4, exd4 8 Nxd4 (8. . c5 9 Qh5ch, Ng6 J O
••• .

Nf5) or 6 Ng6 7 Nc4 followed by 8 d4 i n case the e-pawn i s de­

•..

fended. If Black continues in the Keres spirit with 6 Ng6 7 Nc4,

.••

Be6, then 8 Ncxe5, Nxe5 9 Nxe5, Qd4 10 Nf3, Qxe4 11 Ng5 is more
dangerous than in comparable lines below (Rosentalis-Zaitsev, Bu­
dapest 1989).

6... Qd4

Otherwise Black remains a pawn down.

7 Qh5 g6

If Black defends f7 with 7 ... Be6 White has time to defend his
e-pawn. A rare example of this was I.Gurevich-Hauchard, World
Junior 1990, which went 8 d3, g6 9 Qg5, h6 10 Nf3!, Qd7 11 Qe5
(or just 1 1 Qe3), Rg8 12 Qd4 and left Black with scant compensation
for his pawn.

8 Nf3
Chapter Eight· The Irregulars 107

This seems more promising than the subject of so much

1 970s study, 8 QgS, Bg7 9 Nd3, fS 10 eS, cS.

8... Qxe4

Material equality is re-established but White has something of

an initiative based on his lead in development and the vulnerability of
Black's queen and c7 and f6 squares. Here 9 Nc3 has been suggested
by Geller, presumably with the idea of 9 ... Qxc2 10 QeS or 9 ... QfS 10
Qh4.

9 Qa5!

And this artistic move (with threats of JO Qxc7 and, to a

lesser extent, JO Qc3), forces Black's hand.

9... Qf4

On 9 ... b6 White has a nice idea in 10 Qc3, Rg8 11 Rel, QdS

and now 12 NgS! (because of 12 . . . Qxg5 13 Qxc6ch).
If Black continues 12...Bg7 13 Qg3, Qd6 White keeps the
upper hand with 14 Qf3, BfS lS d3, h6 16 Ne4 (I. Burevich-Bron­
stein, Hastings 1991-2).
108 Winning With The Ruy Lopez Exchange variation

Perhaps better is 9...Bg4 10 Rel, Qd5. Then 1 1 Qc3 obtains

little. White should probably settle for the slight endgame advantage
of 1 1 Qxc7, Bxf3 12 gxf3, Qxf3 13 d3, Qg4ch 14 Qg3.
After 14...Qxg3ch 15 hxg3, 0-0-0? 16 Bg5 and 17 Bf6 fa­
vors White significantly (Brunner-Yilm�, Manila 1992). Black can
reduce but not entirely eliminate that edge after 15 f6 16 Nc3, Kf7
••.

17 Be3, Nf5 16 Bb6.

10 d3

Not 10 d4 because White would like tactical chances along

the c3-h8 line, rather than a c2 pawn that can be attacked by Bf5.
•••

10... Qd6
1 1 Nbd2

White need not hurry into seizing the diagonal, since now
1 1 Bg7 would allow 12 Ne4 and 13 BgS very strongly.
•••

1 1... Nd5

Now in Timman-P. Nikolic, Brussels White demonstrated

his endgame was quite nice after 12 Ne4, Qb4 13 Qxb4, Nxb4 14
Bd2!, Be7 (14 ... Nxc2 ? 15 Rael, Nb4 1 6 Bc3) 15 a3, NdS 16 c4,
Nb6 17 Bf4.
Chapter Ei2ht: The Irregulars 109

A similar, but more spectacular attempt to bust the line was

Lein-Biyiasas, New York 1977, which went 12 Relch, Be7 13 Ne4,
Qb4 14 Qxb4, Nxb4 15 Bh6!, Nxc2 16 Bg7, Nxel (16... Nxal 1 7
Nf6ch, Kd8 18 Ng5!) 1 7 Rxel, Rf8 18 Nf6ch, KdS 19 Ng5.

(e)
THE RARE 5 ... Be7
(After 1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4 Bxc6, dxc6 5 0-0)

Aside from the Soviet player Arkady Novopashin, who tried

it a few times in the 1970s, this has almost disappeared from practice.
5... Be7
6 Nxe5!
If there is any way to refute the bishop move, this is it. After
quieter play, such as 6 d3, Black will defend the e-pawn with 6...Bf6
and ckvelop his knight at e7. Compared with normal positions of the
Double Delayed Exchange Variation (1 e4, e5 2 Nf3, Nc6 3 Bb5, a6 4
Ba4, Nf6 5 0-0, Be7 6 Bxc6, dxc6 7 d3, Nd7 8 Nbd2, 0-0 9 Nc4, Bf6)
White has gained some time, but Black's knight will stand well on g6.
6... Qd4
7 Nf3 Qxe4
110 Winning With The Ruy Lupez Exchange Variation

8 Rel

The queen has no really gcxxl squares that are secure from
further harassment. Fischer, in My Sixty Memorable Games, offered
this supporting analysis to claim White is better: 8.. . Qf5 9 b3, Nf6 10
Re5!, presumably 10...Qd7? 11 Ba3 or 10...Qg6 11 RgS, Qh6 12
d4.

Better is 8 ... Qg6, intending to develop his bishop, after which

9 NeS, QfS 10 d4 creates developmental problems for Black. His
bishop is vulnerable to discoveries ( 10. . . Nf6 11 Nc3, 0-0? 12 Nxc6,
bxc6 13 Rxe7 or, better yet, 12 g4!, Qe6 13 Ng6).

Black has to block the e-file with ••• Be6 at some point -- but
must still watch out for White's g-pawn, e.g. 10...Be6 1 1 Nc3, 0-0-0
12 Ne4 with a threat of 13 g4 or 13 Ng3. Also in that line, White
12 Re4 and 13 Rf4.
might try to attack the queen with
Chapter Nine: How Fischer Played It 111

CHAPTER NINE : How Fischer Played It

Bobby Fischer only played the Exchange Variation of the

Ruy Lopez nine times in competition over a six-year period. but bis
games had a most impressive -- and instructive -- impact on the rest of
the chess world. The games that follow show White's many roles. He
is sometimes on the attack and sometimes on the defense, sometimes
sacrificing his pieces and sometimes sacrificing his opponents', cre­
ating a mating attack against Svetozar Gligorich and finding an
endgame-saving defense against Vasily Smyslov. These are not uni­
formly well-played, but they are uniformly useful to the student.

(1)
FISCHER-PORTISCH
Havana Olympiad 1966

To fully appreciate Fischer's use of the Exchange Variation it

is worth considering not just how he played it, but when and against
whom. For instance, as word of his success with it in the Havana
Olympiad spread, it was widely assumed that he sprang it the first few
times he had the opportunity. Not at all.

Fischer didn't want to waste the novelty value of his new

weapon by using it against an opponent he should beat through nor­
mal (i.e. 4 Ba4) means. And he didn't use it when the points were not
so precious. So, when Sven Johannessen of Norway defended the
Lopez with 3 a6 in the preliminaries, Fischer retreated and achieved
...

a substantial positional edge by move 19 (and won). Against Fridrik

Olafsson of Iceland, who had lost the only previous time he played
1 eS against Fischer, Bobby also routinely played the bishop to a4.
. ..

Later still in this team tournament, Fischer also played 4 Ba4 against
Boris Spassky, perhaps because he was expecting a Marshall Gambit
-- as the Russian had played against him exactly three months before -
- and had an improvement. (What was the improvement? We'll prob-
1 12 Winning With The Ruy Lupez Exchange Variation

ably never know, since Fischer never played the White side of the
Marshall again.)

1 e4 es
2 Nf3 Nc6
3 BbS a6
4 Bxc6!

But Lajos Portisch must have seemed like the perfect oppo­
nent for the Exchange Variation: he was an openings theoretician who
had variations prepared for virtually any likely grandmaster opponent
-- in the main lines. Taking him on again in the Breyer Variation of
the Lopez, as Fischer had at the Piatigorsky Cup the previous
summer, would have been playing into his strength. But in the Ex­
change Variation?

4... dxc6
5 0-0 f6

An intriguing choice considering the establishment theory of

1 965 . At the time, the books almost unanimously recommended
5 ..Bg4, and of course, if 6 h3, then 6 ... hS!. Fischer was probably not
.

the only player who already knew the defects of the bishop-pin.

6 d4 exd4
Chapter Nine: How Fischer Played It 1 13

7 Nxd4 cs

Portisch later turned to 7 Bd6 and 7 ... Ne7, staying in the

•••

middlegame, against lesser opponents.

8 Nb3 Qxdl
9 Rxdl Bd6

Clearly, the Hungarian GM knew enough of "book" to rec­

ognize this as superior to 9 ... Bd7. A contemporary work on the
Lopez, Max Euwe's Theorie der Schach-Erojfnungen, cited some
Barendregt games that went 9...Bd7 10 Bf4, 0-0-0 1 1 Nc3, Re8 12
NdS with advantage to White.

10 NaS! bS

This move, which was known even before 1 966, was later
criticized by Yefim Geller, who claimed sufficient counterplay for
Black from 10 Nh6, and if 1 1 Bxh6, gxh6 12 Nc4, then 12 Be7 13
.•. •••

Nc3, Kf7 14 NdS, Be6 (Bagirov-Keres, Moscow 1967). But now 15

NaS! , instead of Bagirov's 15 Nxe7?, offers excellent chances for an
advantage.

1 1 c4

Remarkably, this was the final position of Barendregt-Sz­

abo, Marianske Lazne 1961 - a "grandmaster draw" in which there
was only one grandmaster, but he (Black) must have known he was
worse. Another game, Perez-Spassky, Havana 1962 had allowed
Black to equalize with 1 1 a4, Ne7 12 Nc3, Rb8.

11... Ne7
12 Be3 f5
114 Winning With The Ruy Lopez Exchange Variation

Since the cS-pawn cannot advance and cannot be protected

further, Black needs some distraction to keep White from playing
Nd2-b3xc5 or 13 Nc3, b4 (13... Bd7 14 Nb7) 14 Na4.

13 Nc3 f4
14 eS!

Portisch probably underestimated the force of this thrust. It

offers to exchange off one bishop-pawn for another (14 . . .fxe3 15
exd6, e:if2ch 1 6 K:if2, 0-0ch 1 7 Kgl, cxd6 8 Rxd6) when Black's
queenside pawns remain under pressure. That, however, is better than
the exchange of his bishop that follows.

14... BxeS?!
15 BxcS Bxc3

More or less forced in view of the threat to the b-pawn and,

more directly, the threat of 16 Rel .

1 6 bxc3 Ng6
17 Nc6 Be6
18 cxbS axbS

White's a-pawn seems to be as weak as the b-pawn. And with

bishops of opposite color present, Black seems to have survived the
queenside crisis. Fischer's deft next few moves show how misleading
that impression is.
Chapter Nine: How Fischer Played It 115

19 Na7! Rb8
20 Rdbl! Kti
21 Nxb5 Rfd8
22 Rb4 Bxa2

Did Fischer overlook this simple trick (23 Rxa2 ? ?, Rdl

mate)? No, he comes up with a surprising method of ending the enemy
tactics.

23 Nxc7 Rbc8
Not 23 Rxb4 24 cxb4, Rd2 because of 25 b5 with what has
•..

become a won endgame.

24 h4! Rd2
Not 24 Rxc7 because of 25 Bb6, Rcd7 26 Bxd8, Rxd8 27
.•.

Rxa2. White also threatened 25 h5 and 26 Rxf4ch here.

25 Bb6 f3?!
26 Be3 Re2
27 Nb5!

Suddenly both the fork on d6 and the powerful 28 Nd4 are

threatened.
27... Ra8
28 h5 Ne5
29 Rf4ch Ke7
116 Winning With The Ruy Lopez Exchange Variation

30 Rdl !
Another deadly idea: 31 BcSch and 3 2 Nc7ch. Whi te can win
material in a variety of ways. The method he uses leads to a mating
net.
30... Re8
31 Re4 Kf6
32 Rd6ch KfS
33 Rf4ch KgS
34 Rxf3ch

Black Resigns

Not wanting to play out 34 ... KxhS 35 RfSch, Kg4 36 RgSch,

Kh4 37 Rd4ch.
Chapter Nine: How Fischer Played It 1 17

(2)
FISCHER-GLIGORIC
Havana Olympiad 1966

The victory over the Hungarian Candidate was dazzling and

Fischer followed it up quickly. Before his opponents had a few weeks
to work out a defense, Fischer struck again using his new weapon,
two rounds -- and two days -- after the Portisch game.

1 e4 e5
2 Nf3 Nc6
3 BbS a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4

All this had been played before -- in fact it had been played
against Gligoric. The bishop pin seems a bit inconsistent, but it makes
a good impression since 7 dxe5 (which Gligoric had played earlier
that year at Hastings) promises only a small endgame edge to White
and 7 c3 certainly looks like a reasonable middlegame for Black.

7 c3
118 Winning With The Ruy Lopez Exchange Variation

7 ... exd4

But this, and his next move, comprise an error. Black's

thinking is simple: after 8 cxd4, Qd7 Black will be able to attack the
d4-pawn at least one more time, with a rook at d8. Since White's
center pawns are contained by Black's own pawns at c6 and f6, White
will have begun defending d4 in some fashion.

8 cxd4 Qd7?
9 h3!

Pointing out Black's tactical slip: He overlooked his inability

to retreat to hS (9. . . Bh5 1 0 Ne5! with advantage). Incredibly, this
game -- certainly the most arresting victory of that day's round -­

seems to have eluded Fischer's final opponent in Havana. See Game

(3).

9... Be6

The Yugoslav grandmaster seems to have been in a bit of

shock for the rest of the game, even playing on when two pieces down.
Clearly, 9...Bxf3 10 Qxf3, Qxd4 was no road to equality, e.g. 1 1
Rdl, Qc4 12 Bf4, Q f7 1 3 Qg3! and Black has too many weaknesses.

10 Nc3 0-0-0
1 1 Bf4
Chapter Nine: How Fischer Played It 119

11... Ne7?

His last bid for a playable game lay in 11 ... Bd6. Fischer, in
his notes, pointed out the neat, forcing variation 1 1 gS 12 Bg3, h5
.•.

and now 13 dS!, exdS 14 Rel!, dxe4 15 Na4!, Kb8 16 Rxe7!, Qxdl
17 Re8eh!, Ka7 ( 1 7. . . Kxc8 18 Nb6 mate) 18 Bb8ch, Ka8 19 Nb6
mate!

Similarly, 14 ... Bd6 in the last line is handled by 15 Na4!,

Kb8 16 Ne5, Qe7 17 Nxa6eh! and wins 17 ...bxa6 18 Nd4, Bd7 19
--

Qb3ch, Ka7 20 Rxe7eh!, Bxe7 21 Bxe7, BbS (or 2 1 . . . Qc5 22 Qe3!

and wins) 22 Ne6eh!, Bxe6 23 Qb6eh and mate next.

12 Rel Ng6
13 Bg3 Bd6
14 Na4!

The first major threat is 15 d5, exd5? 16 Nb6ch. It won't be

the last.

14... Bxg3?
120 Winning With The Ruy Lo12ez Exchange variation

And after this there is nothing left to capture the knight that
arrives on c5 with devastating impact. The weakness on g3 is imagi­
nary. Black should have moved his king to greater safety at b8.

15 fxg3 Kb8
16 Nc5 Qd6
17 Qa4!

17... Ka7?

And this final error virtually ends the game here. Not much
better was 17 Bc8 18 Rc3, Qxg3? because of 19 Ne5!, Qh4 20
•••

Nxc6ch! , or 17 Bc8 18 Rc3, Ne7 19 Rfcl, Kb8 20 d5!, cxd5 21

•••

Nd4 with a deadly attack.

But Black can improve his defensive chances greatly in that
variation with 18 Nf8! (instead of 18. . . Qxg3 ? or 18. . . Nel).
•••

18 Nxa6!

That does it. Black loses his queen or king after 18 bxa6 19
•..

Rxc6. The rest can be kindly described as "desperation."

18... Bxh3

Hoping against hope for 19 gxh3, Qxg3ch.

Chapter Nine: How Fischer Played It 121

19 e5! NxeS

Played because the intended 19 fxeS allows 20 NcSch, Kb8

...

2 1 Rc3 and 22 Ra3! with a mate.

20 dxeS fxeS
21 NcSch Kb8
22 gxh3 e4
23 Nxe4 Qe7
24 Rc3 bS

Now just about anything wins, including Fischer's alternative

of 25 Qa6.

25 Qc2

Black Resigns.
122 Winning With The Ruy Lupez Exchange Variation

(3)
FISCHER-E. JIMENEZ
Havana Olympiad 1966

The novelty value was clearly exhausted by now. Surely, the

first board for Cuba must have known what to expect? But what hap­
pened in the final round in Havana, eight days after the Gligoric
game, is hard to explain.

1 e4 e5
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4
7 c3 exd4
8 cxd4 Qd7
9 h3 Bh5

Fischer gave this move an exclamation point, claiming the

endgame that now results was defensible.
Chapter Nine: How Fischer Played It 1 23

Either Jimenez was somehow unaware of Game 2 above,

Fischer's victory over a world-class opponent in 25 moves, or -- more
likely -- he had figured out a way to play the endgame. But he quickly
goes astray.

10 NeS! Bxdl

Not 10 fxeS 1 1 QxhSch and 12 QxeS(ch), which favors

..•

White in a big way, or the desperado 10 ... Qxh3, which favors him in
a smaller way following 1 1 gxh3, Bxdl 12 Rxdl, fxeS 13 dxeS, BcS
14 Kg2.

11 Nxd7 Kxd7
12 Rxdl Re8
13 f3

Black has no compensation for the inferior pawn structure

now that his light-colored bishop is off the board. White can play
quietly with Be3, Rael and then (a) Na4-c5 with an attack against b7
as in the Gligoric game, or (b) pushing the d-pawn to exploit the light­
square weaknesses that no longer have a bishop to coddle them.

13... Ne7
14 Nc3 Kc8
15 Be3 fS
124 Winning Witb The Ruy Lopez Exchange variation

Black's last move is his only way of changing the pawn

structure -- and thereby buying space for his rooks. Better that than
waiting for White to change it.
16 Rael fxe4
17 fxe4 g6?

The last real mistake Black makes. He had to play 17 ...Ng6

to stop. . .
1 8 Bf4! Bg7
19 d5!

The obvious target is c7: 19...cxdS is answered by 20 Na4!,

Nc6 (else Rxc7ch or Nb6ch) 21 exd5, NeS 22 d6, c6 23 Nb6ch, Kd8
24 BgSch or 23 ... Kb8 24 d7 and wins.

White has his choice of winning methods after 19 ...cS 20

Na4, b6 either 21 b4 or 21 d6.
--

19... Rd8
20 Na4 Rhf8
21 g3

White takes his time, secure in the knowledge that moves

such as 21 ... bS only make Black's life worse because of 22 NcS fol­
lowed by 23 Nxa6 or 23 Ne6.
Chapter Nine: How Fischer Played It 125

21... g5?

A bid for rook counterplay that only temporarily diverts

White's attention from c7. One wonders what Black would have done
if White had retreated the bishop to g3 at move 2 1 . Surely stodgy
defense with something like 21. Rfi was best.
••

22 Bxg5 Rt7
23 Kg2 cxd5
24 exd5 Kb8

Naturally, not 24 Rxd5? because of the knight fork.

•.•

25 Rel Bf8

Black's pieces become clumsier and more passive. But the

alternative 25 ••• Re8 sets up an e-file pin, while 25... Rd7 loses mate­
rial directly to 26 Nc5.
26 Rfl Rg7
27 Bf6 Rg8
28 Rcel Rd7

White's last three moves make a nice impression. With

Black's pieces under heavy pressure, he reorganized his own forces to
increase the pressure. Black is just holding on (28. . . Rd7 29 Nc5 ?,
Rxd5).

29 d6! cxd6
30 Bxe7 Bxe7

Or 30 Rxe7 31 Rxf8ch and 32 Rxe7.

•••
126 Winning With The Ruy LQpez Exchange Variation

31 Rf7

Black Resigns

Thanks to his 29th move White forks pieces decisively:

31 ...Re8 32 Nb6, Rc7 33 Nd5, Rc2ch 34 Kf3.
Chapter Nine: How Fischer Played It 127

(4)
FISCHER-SMYSLOV
Monaco 1967

This was played in a nine-round international invitational

after Fischer had won his first five rounds and was coasting towards
victory. The surprise value of 4 Bxc6 long gone, he plays conserva­
tively -- and nearly gets a lost endgame.

l e4 e5
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4
7 dxe5

Fischer gave both this and 7 c3 an exclamation point in My

60 Memorable Games, but began to rely on this transition into the
endgame from 1 967 on.

7... Qxdl
8 Rxdl fxeS

Fischer regarded this as an error and felt that 8...Bxf3 9 gxf3,

fxeS was best for Black, after which he liked 10 Be3. Then if Black
plays 10 Ne7, his knight is misplaced and 1 1 f4! is strong (compare
...

with the immediate 1 0f4, when 1 0. . . Nf6 is effective).

9 Rd3! Bxf3?!
10 Rxf3 Nf6
11 Nc3 Bb4
128 Winning With The Ruy Lo.pez Exchange YariatiQn

Black can even take time out for 1 1 h6! since in the absence
•••

of a Bg5xf6 threat, White's initiative rumbles to a halt: 12 Rd3, BcS

13 Be3, Bxe3 14 Rxe3, Rd8 15 Rdl , Rd6! and White must cede lhe
d-flle or undouble Black's c-pawns (Hecht-Matanovic, Berlin 1971).

12 Bg5

12... Bxc3

It's surprising that such an openings aficionado and sharp

tactician as Yefim Geller should have fallen -- three years later -- into
a quickly lost position. But in Bednarski-Geller, Siegen 1970 he
played 12 ... 0-0-0?, overlooking 1 3 Rxf6!, gxf6 14 Bxf6, Bxc3 15
bxc3, Rhf8 16 Bxd8, Rxd8 17 Kfl, Rd2 18 Rel with a clear advan­
tage for White. Once White cut off the enemy king and created a
passed pawn -- 18 ... cS 19 Ket, Rd6 20 Rdl, Rb6 21 Rd5, Rblch
22 Kd2, Rfi 23 Ke2, Rel 24 Rd2, c4 25 f4?, exf4 26 e5! - Black -

could resign.

13 bxc3

The desperado 13 Bxf6? fails to 13 ...Bxb2 14 Rbl, gxf6 15

Rxb2, 0-0-0 with a clear edge for Black (or, as Bobby put it, the
"cheap trap" of 14 Bxg7??, Bxal 15 Bxh8, 0-0-0!).
Chapter Nine: liow Fischer Played It 129

And 13 Rxc3 allows the fork on e4 with a quick resolution

into equality: 13 ...Nxe4 14 Re3, Nxg5 15 Rxe5ch, Kf7 16 Rxg5,
Rhd8 as in Gipslis-Smyslov, Moscow 1967 -- a grandmaster draw.

13... Rf8!

A key defensive move which greatly reduces any chance for a

White advantage. The Black king needs to stay in the center, particu­
larly to prevent the invasion of a rook at d7.

Compare with 13 ... 0-0? 14 Bxf6, Rxf6 15 Rxf6, gxf6 16

Rdl! and White dominates the d-file, or 14 ... gxf6 15 Rdl, Rad8 16
Rfd3, Rxd3 17 Rxd3, Rf7 18 Rd8ch! (18. . . R/8 19 Rxf8ch and 20/4!
is a won king-and-pawn endgame), Kagan-Kostro, Siegen 1970.

14 Bxf6 Rxf6
15 Rxf6 gxf6
16 Rdl Ke7

Also OK is 16... aS with the unusual, but not at all unreason­

able, idea of activating the rook at a5 or a6, e.g. 17 Kfl, a4 18 Ke2,
RaS . This was later tested in a few games, such as Markland-van
Scheltinga, Wijk aan Zee 1973: 16 aS 17 Kfl, Ra6 18 Rbl, b5 19
••.
130 Winning With The Ruy Lupez Exchange variation

Ke2, a4 20 c4, Rb6 21 c5, Rb8 22 a3, Ke7 23 c3, Rg8 24 g3, h5 25
f'3, Ke6 26 Rdl, Ke7 and neither side could make progress.

17 Rd3 Rf8
18 Kf1 a5
19 g4

This fixes the Black pawn on f6 but gives him the opportunity
to open the h-file.
19... Rg8
20 h3 b5
21 Ke2 Ke6
22 Kf3 Rb8!

Black, with a slight inferiority, must keep his rook active.

23 Ke3 c5

In Informan t #3, Editor Aleksander Matanovic questioned

this and recommended the waiting move 23 ... Rg8 because . . .

24 c4!
Chapter Nine: liow Fischer Played It 131

... of this move, a temporary sacrifice of an almost-meaning­

less pawn that gives White's rook access to the queenside. In contrast,
24 RdS allows Black to defend with 24... c4 25 Res, Kd6.

24... bxc4
25 Ra3?!

But this is wrong, according to Matanovic, who preferred 25

RdS and said it led to a slight but clear edge for White. The edge is
apparent after 25 Rd5, Rb5 26 a4, Rb4 27 Rxc5, Rxa4 28 Rxc7,
Kd6 29 Rxh7, Ra2 30 Ra7 because his h-pawn runs fast.
But what about active, rather than passive, defense? Is White
accomplishing much after 25 ... Rb2 26 Rxc5, Kd6? That is not clear.
25... Rb5

Now Black should be equal. Bobby rarely agreed to draws in

positions that were merely equal.

26 Kd2 Kd6
27 Rf3 Ke6
28 Kcl Rb8!
29 Ra3 Rb5
30 Rc3 Kd6!

Black will not allow himself to be zugzwanged.

132 Winnin2 With The Ruy Lopez Exchan2e Variation

31 a3?
After this, Fischer begins to get into trouble. Without this
move White cannot capture on c4 (31 Rxc4, Rb4!). It also keeps the
Black king from ever invading at b4. But at the same time the pawn
move denies his rook access to a3.
31... Rb8
32 Rxc4 hS
33 f3 Kc6
34 Rc3 Rd8
Suddenly Black stands slightly better because of his much
more active rook. It's easy to see here that 35 gxhS?, Rh8 causes his
own kingside to self-destruct.
35 Rd3 Rh8
36 Kd2?

And this repositioning is very risky. With 36 c4! he could

have equalized.
36... c4
37 Re3 KcS

Black cannot lose now, and in fact has excellent winning

chances.
38 Re2 hxg4
39 hxg4 Rhl?
Chapter Nine: How Fischer Played It 133

Smyslov, the past master of rook endings, misses a little fi-

nesse after which Fischer safeguards his pawns. With 39 ... Rh3! 40
Ke3 inserted first, the rook can invade effectively with 40 ... Rhl 41
Rg2, Ral since the king is now at e3 rather than d2 and the a-pawn
falls with check.
Geller, who also played in the tournament, gave these lines:
39...Rh3 40 Ke3, Rhl 41 Rg2, Ral 42 g5, Rxa3ch 43 Kd2, fxg5 44
Rxg5, Kd4! or 39...Rh3! 40 Rf2, Rg3 (pass) 41 Kc3 (or 41 Rfl,
Rg2ch 40 Kc3, Re2 with approaching zugzwang), Rgl and White
soon runs out of moves.
40 Rg2!
Draw agreed. The threat is 41 g5, which would activate his
rook and probably force Black's king to retreat (40... Kd4 41 c3ch).
134 Winning With The Ruy Lopez Exchan2e Variation

(5)
FISCHER-ANASTAPOLOUS
Athens 1968

This was played as part of a simultaneous exhibition, with

clocks, pitting Fischer against the top Greek players, including Black
here, a national champion.
1 e4 e5
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 exd4
7 Nxd4 cs
8 Nb3 Qxdl
9 Rxdl Bd7
10 a4
Of the various alternatives to 10 Bf4, this (and the other move
order of JO Be3, b6 1 1 a4) is the most attractive. If Black forestalls
the attack on his c-pawn by way of 10 a4, b6 1 1 Be3, aS, then his
queenside holes at b5 and c4 are highly vulnerable, e.g. 12 Nc3, 0-0-0
13 Bf4, Ne7 14 Rd2, Ng6 15 Bg3, Bc6 16 Rxd8ch, Kxd8 17 Nd2,
Bd6 18 Bxd6, cxd6 19 Nc4, Kc? 20 Rdl, Rd8 21 Na3! (Hurnik­
Lachut, Poland 1975).
Chapter Nine: How Fischer Played It 135

10 ••• 0-0-0?!

11ris cooperates with White, as now 11 a5! and 12 Be3 would

put the queenside under immediate pressure. Fischer seems to play the
opening without his usual exactness.

Black should meet 10 a4 with 10 b6 after which Black can

..•

cover his queenside with his king (11 Nc3, 0-0-0 12 a5, Kbl) or
complicate matters, such as in one of the few times Anatoly Karpov
found himself with the Black pieces (vs. Gipslis, U.S.S.R. Champi­
onship 1970) and ckfended against 10 a4, b6 11 Nc3, 0-0-0 12 Bf4
with 12...c4! 13 Nd2, Be6 with equality. (He should have been put to
the test by 13 Nd4!, when 13...Ne7 14 a5, b5 allows 15 Ndxb5).

1 1 Be3 b6
12 Nc3

The Soviet Grandmaster Eduard Gufeld later said Fischer

mistimed an opportunity here. Once Black has played his pawn to b6,
White should immediately reply 12 a5! since 12 ...c4 13 axb6!, cxb3
is much better than in the game (14 cxb3 with advantage).

12... Bd6
13 as c4!

Neatly played, especially considering the rating difference. In

many similar positions Black must make this attempt to seal the
queenside in order to survive.

Objectively, 14 Na4 is now the best reply. But that would

allow Black to seal matters with 14 ... bS! , after White has holes to
exploit, but it is hard to construct a plan for substantial White pro­
gress.

14 axb6!?
136 Winning With The Ruy Lopez Exchange Variation

Fischer rejects the tacit draw offer and sacrifices a knight for
two pawns and a queenless attack on the enemy king.

14... cxb3
15 Rxa6 Ne7!

Brings a key piece to the defense. If it reaches c6, the attack

is extinguished.

16 Ra8ch Kb7
17 Ra7ch Kb8

18 NdS!

This forces the exchange of Black's incoming defender before

it reaches c6 and also creates the prospect of a phalanx of queenside
pawns (exd5 followed by c2-c4-c5).
18... NxdS

Black doesn't have time for 18 ... bxc2 19 Rdal, Rc8 20

bxc7ch, Bxc7 21 Nxe7 or 19 ... NxdS 20 exdS threatening mate with
two rook checks.

19 exdS Bc8?
Chapter Nine: How Fischer Played It 137

A natural defensive move which fails only because Black

overlooks White's 2 1 st move. The correct defense, which leads to at
least a slight Black advantage, was 19 ...cxb6! and then 20 Bxb6,
bxc2!, attacking the rook at dl. If White replies 21 Rel, then
21...Bf5! 22 Bxd8, Rxd8 ends the attack and favors Black's bishops
and the powerful c-pawn. And on 21 Rdal, Black responds
21 ... Rc8!, threatening to promote the c-pawn. His king can then just
walk away from rook-checks.

20 bxc7ch Bxc7
21 c4!

Black had counted on 21 Rd3 (intending R(x)b3ch), against

which Black plays 21 ... bxc2! 22 Rb3ch, Bb6!!, e.g. 23 Rxb6ch,
Kxa7; 23 Bxb6??, cl(Q)ch; 23 Ral, Rxd5, etc.

21... Bb7
22 Rd3 Rd7
23 Rxb3 Bd6

24 Ra5!

Black's rook and bishops are well-placed to meet any chal­

lenge from the enemy rooks along the a- or b-files -- but not to answer
138 Winning With The Ruy Lopez Exchange Variation

a pawn storm. He can't, for instance, stop the c-pawn from reaching
c6.
24... Kc8
25 Rab5 Bb8
26 g3

Taking time to create luft (26 c5?, Rxd5!).

26... Re8
27 c5 Ba7
28 c6 Bxc6
29 dxc6 Rc7

Two pawns down, Black is lost on any series of exchanges.

But at least he seems to have escaped immediate loss (30 Bxal,
Reich! 31 Kg2, Rxa7)
30 Rb7!

Black Resigns

The bishop can't move because of mate on b8, and 30 ...Rxb7

31 cxb7ch, Kb8 32 Bf4ch is hopeless.
Chapter Nine: How Fischer Played It 139

(6)
FISCHER-RUBINETTI
Buenos Aires 1970

Fischer had played against 3 a6 three times in this tourna­

•••

ment before this game and observers might have wondered whether he
had given up on 4 Bxc6. He hadn't.

1 e4 eS
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4
7 dxe5 Qxdl
8 Rxdl Bxf3

By the time this game was played, this capture had replaced
8 fxeS (as Smyslov played) as the fashionable method of defending.
•••

It fell out of fashion after this game.

With his eighth move, Black gives up his bishop -- his one
significant advantage in the position -- to double White's pawns and
relieve pressure on eS. The resulting position remains slightly unbal­
anced because (a) White can easily undouble his pawns with f3-f4,
but (b) Black has more immediate use of the half-open d-file than
White has of the half-open g-file.

9 gxf3 fxeS
10 Be3

At first glance, 10 f4 looks correct, immediately undoubling

the f-pawns. However, 10 ... Nf6! 1 1 f3, Bd6 12 Nc3, exf4! should
equalize quickly.
140 Winning With The Ruy Lo_pez Exchange variation

10... Bd6
1 1 Nd2 Ne7
12 Nc4 0-0-0

13 Rd3!

A nice building-move, which threatens to pressure d6 further

with a doubling of rooks. Black's reaction creates a new pawn weak­
ness, but he could hardly play 13 Kb8 14 Radl, Nc8 15 Nxe5.
•••

13 ••• b5
14 Na5 Bb4
15 Nb3 Rxd3
16 cxd3 Ng6
17 Kfl Rf8
18 Ke2 Nf4ch

A highly committal move, as it leads to a good-N-vs.-slightly­

bad-B ending. Black hopes that his rook at f4 will provide enough
counterplay. And it was hard to suggest an active alternative
(18. . . Nh4 19 Rg1, g6? 20 Rg3, threatening 21 Bg5 or 21 Rh3).

19 Bxf4 Rxf4
20 Rgl Rh4?
Chapter Nine: How Fischer Played It 141

Black evidently didn't like the looks of 20...g6 21 Rg4, but

the text loses quickly.

21 Rxg7! Rxh2
22 a3 Bd6
23 f4!

lbis sacrifice doesn't require much thought. Once White sees

that 23 Na5 is bad because of 23 ... BcS!, he realizes that he must find
another way of realizing his advantage. But wherein lies his advan­
tage, he asks himself? To some degree its in the advanced rook and
the possible N-vs-B battle.
But it also lies in the potential pawn majority in the center.
Up to 22... Bd6 White's four center pawns have been held in check by
one opponent, at e5. After 23 f4, a temporary sacrifice, the pawns are
freed and, since BcS is averted, so is White's knight.
...

23 ... exf4
24 d4 Kd8
2S NaS! cs
26 eS Bf8
27 Nc6ch Ke8
28 Rxc7
Black Resigns
Perhaps a bit premature, but 28...cxd4 29 e6 will win a piece.
142 Winning With The Ruy Lnpez Exchange Variation

(7)
FISCHER-UNZICKER
Siegen, Olympiad 1970

This was the last serious tournament game in which Fischer

played 4 Bxc6. In tournament play, he had faced virtually nothing but
Sicilians and Caro-Kanns between the 1967 Smyslov game and the
Vinkovci tournament of 1968 when he reverted to normal (4 Ba4)
Lopezes -- and the Bishop's Gambit. The Exchange Variation seemed
to have been forgotten.

1 e4 es
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 exd4
7 Nxd4 Ne7

8 Be3!

When Black's seventh move first began to appear in master

games, White tried routine development such as R Nc3 and then
Chapter Nine: How Fischer Played It 143

8 Ng6 9 f4 in order to mobilize his kingside majority. However,

•.•

9 ... cS! reaches a better-than-usual version of the endgame line: 10

Nf3, Qxdl 11 Rxdl, Bg4 12 Kf2?, Bd6 13 NdS, 0-0-0! and White's
f-pawn and e-pawn are more liabilities than assets.

Later, 8 Na3 gained popularity, e.g. 8...Ng6 9 Nc4, Bd6 10

NfS! or 8 cS 9 QhSch, Ng6 10 NfS. But Fischer's common-sense
.••

move works most smoothly in terms of coordinating White's forces.

The d4-knight is protected and the gl-cS diagonal is safeguarded.

8... Ng6
9 Nd2

Clearly 9 f4 is again risky because of the endgame (9... c5 J O

Nb3, Qxdl 1 1 Rxdl, b6 12 Nc3, Bb7 13 a4, Bd6 14 a5, 0-0-0 15
Na4, c4! with advantage to Black, Kupper-Prameshuber, Lugano
1968). On 9 QhS Black has equalizing chances by playing 9 ... Qe7,
attacking the e-pawn, and then 10 Qf7.
...

Fischer prefers to develop his last minor piece. Now on

9 ...Be7, the Qe7-f7 maneuver is impossible and 10 NfS becomes
..•

strong. And since 9 ...BcS?? allows 10 Nxc6!, bxc6 1 1 BxcS, Black

decides to develop the bishop on the only other square of reason.

9•.• Bd6
10 Nc4

This knight is the one that supports the e-pawn, so if White is

planning on pushing his c-pawn and then seizing the b3-g8 diagonal,
now is the time to do it. (See next paragraph for comparison). 10 c3,
0-0 1 1 Qb3ch, Kh8 12 NfS, as played by the young Karpov, also
favors White slightly.

10... 0-0
1 1 Qd3!
144 Winning With The Ruy Lopez Exchange Variation

Seeking the weakened b3-g8 diagonal. Previously, White had

played either 1 1 Nxd6?! , which gave up any hope of advantage
(J l . cxd6 1 2 Qd2, Re8 13 f3 ?!, Ne5 14 Bj2, d5! - Barczay-Portis­
.. -

ch, Sousse 1967) or tried to reach the diagonal by way of 1 1 c3 and

Qb3.

The latter plan had one memorable fiasco, Damjanovic­

Portisch, Palma de Mallorca 1967, which went 11 c3, Re8 12 f3,
Bf8 13 Qb3, Kh8 14 Rfel? and now Black retook the diagonal with
a pin and a vengeance: 14 ...cS! 15 NfS, Be6 16 a4, NeS 17 Radl,
Qc8! and White resigned.

1 1... NeS

A natural move, exchanging off the more active White knight

for the Black one. But it encourages a move that Black believes to be
bad -- and White proves to be good.
12 NxeS BxeS
Or 12 ...fxeS? 13 Qb3ch and 14 Ne6, leading to a position in
which Black has a bad bishop, bad pawns and nothing as compensa­
tion.

13 f4 Bd6
Chapter Nine: How Fischer Played It 145

14 f5!

Fischer once belittled Emanuel Lasker as a dull, coffee-house

player, but he had learned a bit from the Old Man's great games. In
particular, he must have recalled how Lasker defeated Jose Capab­
lanca to overtake him in the final rounds of St. Petersburg 1 9 14. In
that historic struggle, a slight variation on our endgame variation,
White obtained a major opening advantage after 1 e4, e5 2 Nf3, Nc6
3 BbS, a6 4 Bxc6, dxc6 5 d4, exd4 6 Qxd4, Qxd4 7 Nxd4, Bd6 8
Nc3, Ne7 9 0-0, 0-0?! 10 f4, Re8 1 1 Nb3, f6 12 fS! and then 12 ... b6
13 Bf4, Bb7? 14 Bxd6!, cxd6 15 Nd4, Rad8?! 16 Ne6 and 17
Radl.

Fischer's plan is similar, although he will try to keep queens

on the board. He will reduce the effect of Black's bishop pair by (a)
exchanging off one of them with Bf4, and (b) restricting the other by
the White pawns at e4 and fS.

14... Qe7
15 Bf4 Bxf4
16 Rxf4 Bd7
17 Rel QcS
18 c3 Rae8

TI1is is the drawback to 14 fS. The e-pawn is backward,

cannot advance safely and is easily attacked.
19 g4! Qd6
20 Qg3 Re7?!
Logical, but the immediate 20 cS with the idea of Bc6
•••. •••

was preferable. He allows White to solve a positional problem tacti­

cally.

21 Nf3 cS
146 Winning With The Ruy Lopez Exchange Variation

22 e5!

I didn't say the pawn couldn't legally advance. This fine move
allows White to liquidate his only weakness and create and mobilize
his kingside majority which can al so serve as a shield for his heavy
pieces.

22... fxeS
23 Rfe4 Bc6
24 RxeS Rfe8
25 Rxe7 Rxe7
26 Ne5

Black has that fine diagonal leading from c6 to hl, but it does
him little good since White can capture on c6 whenever it gets too hot.
See next note.

26... h6
27 h4 Bd7?

This move was criticized after the game for its passivity.
much more natural -- and active -- is 27...Qd2 or 27 ...QdS 28 Nxc6,
Rxelch 29 Qxel, Qxc6 30 Qe3, QdS. In the latter endgame White's
advantage -- a kingside majority -- would be difficult (yet not impos­
sible) to convert because of the presence of queens on the board.
Chapter Nine: How Fischer Played It 147

28 Qf4 Qf6
29 Re2! Bc8

The trap was 29 Qxh4? 30 Ng6.

..•

30 Qc4ch Kh7
3 1 Ng6 Rxe2
32 Qxe2 Bd7
33 Qe7!

Forcing queens off the board to reach a knight-versus-good

bishop endgame in which Black is lost because of the inability of his
king to play, without making concessions.

33... Qxe7
34 Nxe7 gS
35 hxg5!

Not 35 fxg6ch, Kg7 after which the g4 pawn is hanging.

Fischer prefers to have a passed pawn that is easily protected -- and
also gain time to win a queenside pawn.
35 ... hxg5
36 Nd5! Bc6

The main threat was 37 Nf6ch. After Black's next move it

seems as if Bobby has miscalculated and will lose the more valuable
g-pawn. But he has seen further: a series of traps based on knight
forks.

37 Nxc7 Bf3
38 Ne8! Kh6

Such as 38...Bxg4 39 Nf6ch.

39 Nf6 Kg7
148 Winning With The Ruy Lopez Exchange variation

40 Kf2!

"Magic" was the tournament book's only -- and sufficient -­

comment about this move.

40... Bdl
41 Nd7! c4
It's also over after 41 ...Bxg4 42 f6ch, Kg8 43 f7ch!, Kxf7 44
Ne5ch.
42 Kg3

Black Resigns

After a night's sleep, Black decided not to resume this ad ­

journed position.
Chapter Nine: How Fischer Played It 149

(8)
FISCHER-KRAMER
Manhattan Chess Club Blitz Tournament 1971

In this double-round, seven-minute-a-game event, held to in­

augurate the Manhattan C .C.'s new quarters, Fischer nearly swept the
field. He was held to one draw. This is his only published game with
the pin variation (Chapter Two).

1 e4 eS
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 Bg4
6 h3 Bxf3
7 Qxf3 Qd7
8 d3 0-0-0
9 Be3

Did Fischer think he needed to stop BcS? Or did he want to

•••

avoid 9 ... cS followed by . .Ne7-c6? Or did he have some other reason

.

for avoiding the "knights-before-bishop" rule that suggests 9 Nd2 is

more exact. The best answer lies in the way Fischer clears his first
rank of minor pieces by move 1 0, thereby allowing his queenside
demonstration (moves 1 1 - 1 4).

That's the most ambitious plan. The most conservative is the

instant endgame, 9 Qg4, with a position similar to the Bernstein­
Benkner game from Chapter One. Another version, Pinter-Swanson,
Teeside 1974, showed how quickly White can achieve a won game:
9 Qxg4? 1 0 hxg4, BcS 1 1 Nd2, Nf6 12 gS, Ng4 13 Nc4, Rd7 14
•••

g3, h6 15 gxh6, f6 16 Kg2!, Nxh6 17 Rhl , Nf7 18 Rxh8ch, Nxh8

19 Bd2, Nf7 20 Rhl, Kd8 21 f4, Ke7 22 Rh7, etc.

9 .•• f6
150 Winning With The Ruy LQpez Exchange variation

10 Nd2 Ne7

11 b4!

Seeing that Black has made a pawn storm difficult for himself
with his 1 0th move (Jl g5 ? 1 2 Qxf6), Fischer accepts the invitation
...

to seek a mate.

11... Ng6
12 Rfbl

The right rook, since White may need the other one on a po­
tentially opened a-file.

12 ••. Be7
13 a4 b6
14 Qe2 Kb7
15 Nc4 Nf4
16 Qf3

Fischer probably planned to capture on f4 when he made his

14th move, leaving Black with a slightly "bad" bishop against the
knight. But here he changes his mind.

16... Ne6
Chapter Nine: How Fischer Played It 151

17 Rb3! Nd4

Black couldn't visualize a comfortable defense to threats such

as 18 Rabi and 19 bS.

18 Bxd4

Now White has no choice. The White knight may not seem so
well placed now, but when it reaches d5 it will.

18... Qxd4
19 Rabi h5
20 bS cxbS
21 axbS a5

22 Ne3!

When it lands on dS, the knight cannot be driven off by ...c6

without a serious loosening of his king security.

22... Qd7
23 NdS BcS
152 Winning With The Ruy Lupez Exchange Variation

Black has kept the queenside lines effectively closed to the

heavy pieces, but White now turns to a different play -- pushing the d­
pawn in order to attack c7.

24 Rc3 Qd6
25 Rc4 Rd7
26 c3 f5
27 QxfS!
Even in a seven-minute game, the future world champion
wouldn't pass up an opportunity like this.

27 .•• Rf8
28 Qxh5 Bxf2ch
29 Kh1 Rdf7
30 Rc6 Bg3

Threatening a last-rank mate and leading to. . .

3 1 Qxf7 Rxf7
32 Rxd6 cxd6

. . . a favorable liquidation that leaves White a pawn up with

the d6 target also doomed.

33 Ne3! Rf2
Chapter Nine: How Fischer Played It 153

34 Nf5 Bf4
35 Nxd6ch Kc7
36 Nc4 Rc2
37 g3

Based on the knight fork that arises in two moves (38. . . Rxc3
39 Nd5ch). But if Black had played 38...Rd2, the game would have
remained a struggle . But the clock begins to take its toll.
37... Bxg3
38 Ne3 Rh2ch?
39 Kgl Rxh3?
40 Kg2!

Black Resigns
154 Winning Witb The Ruy LQPez Exchange Variation

(9)
FISCHER-SPASSKY
16th Game
World Championship Match 1972

On his road to the world championship, Fischer avoided the

Exchange Variation in the 1 970 Interzonal at Palma de Mallorca -­
and never got the chance to play a Lopez in his subsequent Candidates
Match victories over Mark Taimanov, Bent Larsen and Tigran Pet­
rosian. But late in his championship challenge to Spassky -- after win­
ning a memorable Breyer Variation and establishing a three-point
lead, Fischer took one opportunity to try his pet variation. It was his
last Exchange Variation and, in many ways, his least.

1 e4 e5
2 Nf3 Nc6
3 Bb5 a6
4 Bxc6 dxc6
5 0-0 f6
6 d4 Bg4
7 dxe5

Fischer had given 7 c3 an exclamation point when annotating

his Havana experience, but his recommendation in My 60 Memorable
Games of 7 c3, exd4 8 cxdS, cS! led him to adopt this version of the
endgame against Smyslov and again here.

7... Qxdl
8 Rxdl fxeS
9 Rd3 Bd6!

This was now recognized as the main line, since 9...Bxf3 had
been played and found wanting in the previous three years.

10 Nbd2 Nf6
Chapter Nine: How Fischer Played It 155

1 1 Nc4

This liquidates the remaining pawns in the center and in­

creases the likelihood of a draw -- not a problem considering Fischer
needed seven <lraws in the remaining 1 1 games to win the match.

1 1... Nxe4

Nowadays 1 1 ... 0-0! is regarded as more exact because 12

NcxeS, Bxf3 transposes into the game and 12 NfxeS allows 12...Be2!
13 Re3, Bxc4 14 Nxc4, Bc5 and Black regains his pawn with equal­
ity. The drawback to the immediate 1 1 ... Nxe4 is revealed in the next
note.

12 Ncxe5

After this game 12 Nfxe5! was discovered and opinions about

the variation changed once more. After 12 ... Bf5 13 f3 (13. . . Bc5ch 14
Be3 or 13... Nc5 14 Nxd6ch, cxd6 15 Rxd6, Bxc2 16 Be3) White holds
a clear edge. Black can reduce that advantage to something
manageable after 12...Be6 13 f3, Bxe5 14 Nxe5, Nc5 but he is still a
bit worse after either 15 Re3 or 15 Rdl. (This order of Black moves
is neater than the immediate 12 . . . Bxe5 13 Nxe5, Bf5 because of 14
g4!, Be6 15 f3 with advantage).
156 Winnin2 With The Ruy Lopez Exchan2e Variation

12... Bxf3!
13 Nxf3 0-0
14 Be3 b5!

Praised at the time for gaining space and throwing White

slightly off-balance. Now 15 Nd2, Nc5 16 Bxc5, Bxc5 17 Ne4, Bb6
leaves Black with a dangerous threat of Rae8. And 16 Rc3, Na4 is
...

nearly as unpleasant.

15 c4?

Fischer probably underestimated the strength of Black's

"antipositional" 1 6th move. Better was 15 Nd2 with the idea of
15 NcS 16 Bxc5, Bxc5 17 Ne4, Bb6 18 Kf1 and 19 Rel, or
••.

15 Nf6 16 Nb3 with a slight pull in either case.

.•.

15... Rab8
16 Rel bxc4!

As in the Smyslov game, the tripled pawns are not as weak as

they seem -- while the open b-file yields immediate benefits. Now 17
Rxc4, Rxb2! favors Black.

17 Rd4 Rfe8
18 Nd2 Nxd2
19 Rxd2 Re4
20 g3

Not 20 Kf1? because of 20 Bxh2! -- an echo of the faulty

.••

•.. Bxh2 capture that cost Fischer the first game of the match.

20... Bes
21 Rcc2 Kfi
Chapter Nine: How Fischer Played It 157

Black has to clear his first rank before initiating the mini­
combination that follows, a combination Fischer must have expected
but could not avoid.

22 Kg2 Rxb2!
23 Kf3!
This denies Black the slightly improved piece placement re
would have gotten from 23 Rxb2, c3 24 Rbc2, cxd2 25 Bxd2, Ra4.
23... c3
24 Kxe4 cxd2
25 Rxd2 Rb5
26 Rc2 Bd6
27 Rxc6 Ra5
28 Bf4 Ra4ch
29 Kf3 Ra3ch
Black is beginning to realize he has no more than a draw:
29 Bxf4 30 gxf4, Rxa2 3 1 Rxc7ch, Kg6 32 Rc6ch and on 32... Kh5
.••

33 Rc7, Kh6 34 Rc6ch, g6 White liquidates more pawns with 35 f5.

30 Ke4 Rxa2
31 Bxd6 cxd6
32 Rxd6 Rxf2
33 Rxa6 Rxh2
158 Winning With The Ruy Lopez Exchange Variation

34 Kf3!

The last, real winning chance lay in Black being able to play
34 Rf2 , which would have cut the White king off from his pawn and
...

threatened to win with a march of the Black king and pawns. But after
34 Kf3 the game must surely be drawn because Black cannot create a
passed pawn without exchanging down from 2-pawns-to- 1 to 1 -0.
And since Philidor's day, we've known how to draw 1 -0 Rook end­
games. (Yes, I know, Jose Capablanca used to win 4-3 endgames,
with all the pawns on one side of the board, but the extra pairs of
pawns greatly increase the ways to win.)

The rest of the game was uneventful and, in fact, something

of a disappointment. I remember analyzing the moves as they came in
on Shelby Lyman's live television broadcast of the match and trying to
find something significant to say about the final 26 moves. There
wasn't much:
34... Rd2 35 Ra7ch, Kf6 36 Ra6ch, Ke7 37 Ra7ch, Rd7 38
Ra2, Ke6 39 Kg2, Re7 40 Kh3, Kf6 41 Ra6ch, Re6 42 Ra5, h6 43
Ra2, Kf5 44 Rf2ch, Kg5 45 Rf7, g6 46 Rf4, h5 47 Rf3, Rf6 48
Ra3, Re6 49 Rf3, Re4 50 Ra3, Kh6 51 Ra6, Re5 52 Kh4, Re4ch
53 Kh3, Re7 54 Kh4, Re5 55 Rb6, Kg7 56 Rb4, Kh6 57 Rb6, Rel
58 Kh3, Rhlch 59 Kg2, Ral 60 Kh3, Ra4.

Draw agreed.