The Instructor

Mate on the Back Rank

The Instructor
Mark Dvoretsky
The Chess Cafe
E-mail Newsletter
Each week, as a service to
thousands of our readers, we send
out an e-mail newsletter, This
Week at The Chess Cafe. To
receive this free weekly update,
type in your email address and
click Subscribe. That's all there is
to it! And, we do not make this list
available to anyone else.
EMail:
Subscribe

Combinations on our title theme may be found in any

manual or collection of exercises; so I'm not discovering
any Americas here. Nor have I set myself the task of
collecting a fresh set of positions, probably ones
unknown to the reader - what would be the point? If you
need to assimilate the typical middlegame combination,
one or two dozen examples (three dozen, at most) are
quite sufficient; there is no need to pile on more and
more. It's far more important to deal with high-quality
material, which will illustrate our theme solidly and give
esthetic satisfaction as well - thus, it will be securely
etched in the memory. I hope that my examples, culled
from various sources, meet this criterion.
A most important part of effective work on chess
consists of practical training - solving exercises on a
study theme. Under each diagram, you will find
indicated who is on move; and if the position works as
an exercise (which the majority of them do), then there
will be a question mark as well. Try solving these
yourself, and only afterwards, consult the following text.
I The first and largest part of this collection consists of
examples and exercises assisting in the development of
fantasy and combinative alertness.
Let's begin with a classic game, known to most players
from their earliest youth.
Bernstein Capablanca

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (1 of 27) [04/06/2003 7:15:22 PM]

The Instructor

Moscow 1914

Order
School of Chess Excellence,
Volume 2:
Tactical Play
by Mark Dvoretsky

Black to move
Black could have played 25...Rc5 26. Nb3 Rc7 here.
Capablanca preferred to set a trap, provoking his
opponent to attack the c3-pawn.
25...Rc7!? 26. Nb5 Rc5 27. Nxc3?
White had to bring his knight back to d4; but after 27...Rdc8 (threatening
28...Nb4), Black retains the upper hand, thanks to his dangerous passed
pawn at c3.

27...Nxc3 28. Rxc3 Rxc3 29. Rxc3

Black to move
?
Evidently, Bernstein expected only 29...Qb1+? 30.Qf1
Qxa2, with a likely draw.

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (2 of 27) [04/06/2003 7:15:22 PM]

The Instructor

29...Qb2! White resigned.

A most instructive fragment - later on, we shall see the
ideas contained in it again and again.
The first thing to note is the deflecting queen sacrifice - a
standard technique to exploit the weakness of the back
rank. Sometimes, the queen has to be sacrificed
repeatedly.
Rovner Kamyshov
Moscow 1946

White to move
?
1. Qa7! Qa5 2. Qxa6! Qc7 3. Qa7! Black resigned.
Vogt - Alexander
Berlin 1962

Black to move
?
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (3 of 27) [04/06/2003 7:15:22 PM]

The Instructor

Can White's queen be decoyed away from the defense of

the f1-square? The solution is a bit unexpected.
1...c4! 2. bc (2. Qxc4 Qxd5!) 2...Qa3! 3. Qd1 Qxc1!
White resigned.
Wolf - Spielmann
Moravska Ostrava 1923

White to move
?
Here, several enemy pieces are decoyed: first, the queen from the protection of the f3-pawn; then, the rook - from
the 8th rank; and finally, the bishop - from the e8 square.
36. h3! Qxh3 37. Qxf3! Black resigned, in view of
37...Rxf3 38. Rc8!+ Bxc8 39.Re8+.
Krause - Capelan
Solingen 1961

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (4 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
It's important to get the center pawns moving, but Black
has concentrated sufficient firepower against the e4
square. Nevertheless, the pawn break is still possible thanks to the 8th-rank weakness.
1. e4! fe 2. fe Rxe4?
2...b5 was necessary. White would have had to content
himself with a rather better position after 3. Bxd5 cd 4.
Qc3 Bg4 5. e5, or 3. e5 Qd7 4. Bd3, since the immediate
3. Bd3?! Nf6 leaves his center too shaky. For instance, 4.
Qc3 Ng4 5. Rf4 (5. Rfe2? Qf8!) 5...b4! (5...Rf8 is
inferior because of 6. e5, when 6...Qd5+? fails to 7.Be4
Rxf4 8. Bxd5 Nf2+ 9. Kg1 Rg6+ 10. Qg3 Rxg3+ 11. hg
Rxd4 12. e6!) 6. Qc4 (6. Qxb4? Qxf4! 7. Bxf4 Nf2+ 8.
Kg1 Nxd3 - such cases are referred to as "family forks")
6...g5! 7. Rxg4 (forced) 7...Bxg4, and White's
compensation for the exchange is problematic.
3. Rxe4 Rxe4 4. Bxd5 cd 5. Qc7!
This was the blow (analogous to the one Capablanca
delivered in our first example) that Black overlooked.
5...Bg2+ 6. Kg1 Black resigned.
In this case, the solution found by White objectively
should not have led to a win. In essence, what we have
here is a successful trap based on the 8th-rank weakness
(just as in the Bernstein - Capablanca game).
Here's another example of a successful trap.
Ganchev - Pipkov
Albena 1966
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (5 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
1. Nc5! Rxa3 2. Rxa3 (of course not 2. Nxe6?? Ra1+)
2...Rxc5?
The bait is taken! Black had to play 2...h6 3. Nxe6 fe 4.
Ra4, when White would stand a little bit better, but
nothing more than that.
3. Ra8!+ Bc8 4. Ra7! Qb6
Black had probably calculated this far, and had found
nothing better for his opponent than the repetition after 5.
Ra8.
5. Qd7!! Black resigned.
And now, we will examine two episodes from Alexander
Khalifman's games. In the first, the grandmaster found a
forced win; while in the second, the decisive
combination went undiscovered.
Khalifman - Ehlvest
Lvov 1985

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (6 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
30. Bf5! (decoying the bishop away from the square e8)
30...Bxf5 31. Qc7!! (and here, a standard deflection of
the queen) 31...Rxd1+ 32. Kxd1
Black might have ended his resistance here. Ehlvest
probably neglected to do so, only out of extreme
disappointment. His last move, leading to the diagram,
had been 29...Qd6-c6?? Had he exchanged rooks on d1
first, everything would have been fine.
32...Bxc2+ 33. Kc1 (33. Kd2 Nc4+ 34. Kc1 would have
been good, too) 33...Ba4+ 34. Qxc6 Bxc6 35. Re6 Bb5
36. Rxb6 Kg8 37. Rb7 d4 38. Kd2 Kf8 39. h4 d3 40.
b3 h6 41. gh gh 42. Rh7 Black resigned.
Khalifman - Hjartarson
Reykjavik 1991

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (7 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
The goal would have been attained after 31. Ng4!! Nxg4
(31...Rxg4 32. Bxf6) 32. Qd5+! Kh8 (32...Rxd5 33.
Re8+) 33. Rf7! (threatening, above all, 34. Bxg7+)
33...Bxc3 34. Rxc7.
The game actually continued 31. Nxg6? hg 32. Bxf6 Rf8
33. Qd5+ Qf7 34. Qd1 (34. Qxf7+ Rxf7=) 34...Ra1! (an
excellent counterstroke!) 35. Qxa1 Qd5+ 36. Kg1 Bxf6
In such a position, the exchange advantage is worth little,
in view of the open situation of the White king and the
queen's unfortunate position.
37. h4 Qd4+! 38. Kg2 Qd2+ 39. Kh3 (39. Rf2 Qxf2+!
40. Kxf2 Bxb2) 39...Qd7+ 40. Kg2 (40 g4 Qd3+)
40...Qd2+ 41. Kh3 Qd7+ 42. Kh2 Qd2+ 43. Kh3 Draw.
And here's another case in which a strong grandmaster
missed a standard combination.
Teschner - Portisch
Monte Carlo 1969

White to move
?
White's position is difficult - on 29. Rc1!? there is the
powerful reply 29...Kb8!. He decided to snatch a pawn
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (8 of 27) [04/06/2003 7:15:22 PM]

The Instructor

with 29. Rxd5?!, which could have been beautifully

refuted by 29...Qf2!! 30. Ng3 Qe1+!.
Portisch chose 29...Qa6? 30. Ng3! Red8 31. Rxd8 Rxd8
instead, and the game ended in a draw ten moves later.
Note that here (as in our first example, Bernstein Capablanca), both sides were speculating on the
weakness of the back rank. Such situations, which we
shall see more than once, require special attention and
farsightedness from both players.
Let me point out one more particular of the position we
just looked at. The move Portisch played was also aimed
at exploiting the back-rank weakness; but it proved
relatively ineffective - the opponent turned out to have a
defense. This can certainly happen to any chessplayer (to
some - more often than others): we see a tempting line,
and quickly bash it out, without checking the variations,
and without seeing if there might be other, perhaps
stronger, possibilities.
The examples that follow illustrate precisely this
situation.
Crouch - Speelman
Hastings 1992

White to move
?
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (9 of 27) [04/06/2003 7:15:22 PM]

The Instructor

The continuation was 26. Rxd7?! Rf8! 27. Rb2 Rxe5,

when a draw was agreed.
White had a far more effective means of exploiting the
back-rank weakness: 26. Rb8! Ra8 27. Ra1!! Qxa1 28.
Rxd8+ Rxd8 29. Qxa1.
Leonhardt - Fahrni
Carlsbad 1911

White to move
?
42. Qa2! Rd1+ (42...c2 43. Qxd5!) 43. Kh2 Qxa2 44.
Rc8+ would have won immediately.
White actually played something much weaker: 42.
Qxe4? Rd1+ 43. Kh2 Qd6+ (this check would not have
existed after 42. Qa2!, since the queen would have been
pinned) 44. f4 Qf8 45. Rxc3 (White gets nothing from
45. Qe6+ Kh8 46. Rc8 Rd8). He still stands better, but
Black can defend. True, the game ended quickly all the
same: 45...Rd8 46. Rc6 Ra8 47. f5 Qb8+ 48. g3 Qa7??
49. Rc7! Black resigned.
Capablanca - Thomas
Hastings 1919

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (10 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
The "conclusive blow" 29. Qa8? had the effect of
causing Thomas to resign; he failed to notice the
uncomplicated defense 29...Ra2!, liquidating all threats.
Commentators to this game have suggested 29. Rxe8!?
Qxe8 30. Qa4! (a typical queen deflection, coupled with
a double threat), when Black either gets mated or loses
his rook - on 30...Rc1+, there follows, naturally, 31. Kf2.
So would this have forced his resignation? Not
necessarily - he could still have tried 30...Rxg2+! 31.
Kxg2 Qxg6+ 32. Kh1 Qh5!? Black would have had four
pawns for the rook; and although objectively, his
opponent would retain good winning chances, there
would still have been a long fight ahead.
So, before executing his "conclusive combination", it
would make sense for White to find a different, shorter,
and more certain means to his end (following the
"candidate-moves" principle). And there is such a means:
after 29. Qb5!, a quick capitulation would indeed have
been unavoidable. On 29...Rc1+ 30. Kf2 Rxb1 31. Rxe8!
is decisive; 29...c6 30. Rxe8 Qxe8 31. Qb8 Rc1+ 32. Kf2
is hopeless too; and if 29...Rxb8 30. Qxb8 Kg8, then
either 31. Qb3+ or 31. Qa7.

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (11 of 27) [04/06/2003 7:15:22 PM]

The Instructor

And here, I must confess to a small falsification. The

diagrammed position differs somewhat from the actual
game: there White also had a pawn at a2, while Black
did not have a pawn at e5. Now the example becomes
uninteresting, since the book solution (29. Rxe8 Qxe8
30. Qa4) also leads to an easy win. In my version,
Whites task becomes more difficult, the exercise is
qualitatively improved, and becomes useful even to
players at higher rating levels. I believe that, in pursuit of
his training goals, a trainer is justified in the occasional
introduction of such alterations in his examples. For the
demonstration of old games is not an end in itself. The
fundamental principle of materials selection must always
be: Maximum benefit to the students from
learning/training exercises.
To conclude this first chapter - a combination I like very
much.
Schwager - Ackerman
Riga 1978

Black to move
?
1...Qe3+ 2. Kh1 Qf2! 3. Rg1
Black's first moves, of course, were obvious - but what
now? He certainly didn't invade the enemy position with
his queen so that he could exchange everything on d2.
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (12 of 27) [04/06/2003 7:15:22 PM]

The Instructor

3... Nh5!! 4. Rxe2 Ng3+! 5. hg Rc5!

White resigned, as mate is forced - this time, not on the
rank, but on the file.
II Our next cycle of exercises is aimed at developing
attention to the opponent's counterchances. This is a very
important habit for any chessplayer to develop: the
ability to see tactics not only for ourselves, but for our
opponent as well, the habit of accurately checking our
intentions, to avoid unpleasant surprises. You would
already have been well served by such an ability in some
of our previous exercises (such as Capablanca Thomas); now this ability will be placed at the forefront.
In some of the examples which follow, you will be
required to foresee and account for the enemy tactic of
exploiting your weak back rank in your plans; in others,
you will have to examine your own combination
critically for this theme.
Panno - Bravo
Fortaleza 1975

Black to move
?
The tempting 28...Ne2+? 29. Rdxe2 Qxe2 is a mistake,
in view of 30. Nf6+! Rxf6 31. ef, when White wins.
Black actually played 28...Qe2!!, and White resigned.
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (13 of 27) [04/06/2003 7:15:22 PM]

The Instructor

Short - Bareev
Tilburg 1991

Black to move
?
Having outplayed his dangerous opponent, the
Muscovite grandmaster slackened his attention for a
moment, and was severely punished: 27...Nxe5?? 28.
Nf5! Black resigned.
"Make haste slowly", said the ancients. The simple
27...Rcd7! (now 28...Nxe5 is threatened) would have led
to the win of a second pawn, without any unpleasant
consequences.
Fernandez - Marino
Zamora 1996

Black to move
?

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (14 of 27) [04/06/2003 7:15:22 PM]

The Instructor

"The back rank is weak - now's the perfect time for a

decoying blow" - is what Black must have thought.
20... Qxa2?? 21. Qd1! Black resigned.
The right move was 20...Bd6!, endeavoring to exchange
White's powerful knight. On 21. Qxd5, Qxa2! is now
good.
Alden - Nilsson
Sweden 1972

White to move
?
White could have played simply 1. h3, preparing for a
lengthy defense a pawn down. However, 1. Rxf6 looks
tempting. Which should he choose?
Long-suffering would have been correct in this case,
since the sharp move is immediately refutable.
1. Rxf6? Qb7??
Black wins with the spectacular 1...Qc6!! (double
decoying, coupled with a double attack); whereas the
text move leads to the opposite result.
2. Qf1 (2. h3 works equally well).
Bradford - Byrne

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (15 of 27) [04/06/2003 7:15:22 PM]

The Instructor

USA 1980

Black to move
?
1...Qd4?? (the consequences of 1...Qd5?? would be the
same) 2. Qxh7+!, and White wins, since after 2...Kxh7
he picks up the rook with check, and only then
recaptures the queen.
After the more accurate 1...Qd6!, Black would have won
instead - it's important that the queen protects the rook
from d6. Another strong continuation was 1...Qf5! 2.
Nxf8 Nxd1.
Mikenas - Bronstein
Tallinn 1965

White to move
?
Above all, Black threatens 24...Qxb2 or 24...Qe1+ 25.
Qf1 Qxh4. There is also another, hidden threat - and it
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (16 of 27) [04/06/2003 7:15:22 PM]

The Instructor

was this that Mikenas overlooked.

24. Rb4? Rxa3!! White resigned.
The same blow (24...Rxa3!!) would have been the reply
to 24. Qd2? or 24. Qc3? There were two ways of
warding off Black's pressure: 24. Qd4! Qe1+ 25. Rxe1
Rxe1+ 26. Qg1; and 24. Rd4! Qe1+ (24...Ra4 25. Rxa4
Qxb2? 26. Qd4; 24...Bb5 25. Qd2) 25. Qf1 - in both
cases, White obtains about even chances.
Seirawan - Lobron
Arnhem/Amsterdam 1983

White to move
?
On 22. Qa1! Rxc1+ 23. Qxc1 Qb6 or 22. Qd2! Rxc1+
23. Qxc1 Qb6, the game should end in a draw - Black is
only a little bit better. But couldn't White just get rid of
that passed pawn, by exploiting the back-rank weakness?
22. Rxa7? Qxf2+!! 23. Kxf2 (23. Qxf2 Rxc1+)
23...Rxb2+ 24. Kf3 Rxc1 White resigned. Black is a
rook ahead, and there is no mate (25. Ra8+ Rc8).
Ballona - Aristazabal
Colombia 1993

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (17 of 27) [04/06/2003 7:15:22 PM]

The Instructor

Black to move
?
1...Nd3+ 2. Kh1 Qe3? is no good: after 3. Rf1 (with the
threat of 4. Qc3) 3...Nxb2 4. Qxb2, it's White who wins.
1...Ne4+ 2. Kh1 Qe3? is equally useless: 3. Rf1, and
Black must consider both 4. Bc1 Qd4 5. Bb2=, as well as
4. Ba1 (preparing 5. Qb2).
Black executed a pretty combination:
1... Nd1+ 2. Kh1 Qxd5!? 3. Rxd1! (of course, not 3.
cd?? Re1#) 3...Bxa4!!
Black had to have foreseen this blow, as otherwise, he
would even lose: on 3...Qxg5 or 3...Qe4 there follows 4.
Qc3; if 3...Qd6, then 4. c5! (but not 4. h3? Qg3, and not
4. Qc1? Bxa4! 5. Rxd6 Rxd6) 4...Qd5 5. h3.
4. Qxa4 Qd2! White resigned, since he remains a piece
down, in view of the threatened 5...Re1+.
It's a well-known truism that it is never too late to resign.
It would have been interesting to see how Black would
have demonstrated his advantage after 5. h3 Qxb2 6.
Qxa5. 6...Qb6?! doesn't promise anything: 7. Rd8+ Kf7
8. Qd2! (threatening 9. c5); also unconvincing are
6...Re7 7. Qc5!, or 6...Nc6 7. Qxc7. Black would
probably have had to continue 6...Rd6!? (7. Rxd6 Qc1+
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (18 of 27) [04/06/2003 7:15:22 PM]

The Instructor

8. Kh2 Qf4+ 9. Kh1 Qxd6; 7. Re1 Qf2), but here too, the
outcome remains unclear: 7. Qe1 Rxd1 8. Qxd1 is only
"=/+".
In a recent training session, I offered this exercise to
three young grandmasters: Alexander Motylev,Vladimir
Potkin, and Ernesto Inarkiev. None of the three was
taken in by the spectacular-looking, but spectacularly
ineffective combination; all three preferred the simpler
and sounder plan 1...Ng4+! 2. Kh1 Qd6! 3. Bxe6+ (3. g3
Bc6) 3...Bxe6 4. g3 Qc6+ 5. Qg2 Nf2+!? 6. Kg1 Nh3+ 7.
Kh1 Bxc4 (or 7...Qb6!?), and wins.
III The third and last cycle of examples are what we call
"playing exercises". This is a form of training that I
worked out, and began using successfully, many years
ago. These are the kinds of positions in which it is not
necessary, and in some cases not possible, to foresee
everything from beginning to end. So a player must
make not just one decision (deep and difficult though it
may be), but a whole series of them.
A significant part of the skill of calculating variations is
the ability to come to a decision with maximum
economy - that is, without calculating any more than
necessary. In a practical game, choosing the first move
of the right plan doesn't necessarily require us to foresee
a refinement on the fourth move - which in turn means
that, in an analogous situation, a trainer also does not
need to require this ability in his training. Otherwise, the
student becomes enmeshed in an irrational algorithm in
order to come up with a move, which must inevitably
result in the time-pressure disease. You can read more
about this form of training in the first and second books
of my series, "School of Chess Excellence".
The following examples are very well suited for playing
out. Trainers may play them against students (setting the
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (19 of 27) [04/06/2003 7:15:22 PM]

The Instructor

time-control in accordance with the complexity of the

position and the strength of the opponent). If you are
training yourself, then cover the text of the "game" with
a piece of paper, and then find one move after the other,
after which you can move the paper just enough to
uncover the "opponent's" move. One good idea is to
establish the amount of time used, and then analyze how
rationally you worked at this or that moment of the
"game".
Our first example is probably the simplest. It lies on the
border between the solving and the playing kind of
exercises - if desired, you may calculate it at once to the
end.
I. Hoch
1973

White to move
?
1. Qxc2! Re8! Both sides are working with back-rank
mate threats. The question becomes, who will prevail in
this sharp duel.
2. Qc8!! Qd7!! (2...Rxc8 3. Rxa4) 3. Qc1! 3. Qa8?
would be a mistake, owing to 3...h6 (or 3...g6).
3...Qc7! 4. Ra8!! (but not 4. Qe3? Qd8!).

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (20 of 27) [04/06/2003 7:15:22 PM]

The Instructor

Lowtzky - Tartakower
Jurata 1937

Black to move
?
White's downfall is not only that his first rank is weak
(as far as this circumstance goes, both players are
similarly placed), but that his rook also stands badly at
g5. The experienced player will have no difficulty in
predicting that the game must be decided by a double
attack: the threat to mate (or to the queen, decoying her
away from the first rank), and simultaneously to the
rook. But accuracy is required, so as not to overlook the
opponent's resources. For example, 1...Qb6+ 2. Kh1
Qf6? would be a mistake, in view of 3. h4.
1...Qc5+ 2. Kh1 Qc4! 3. Kg1 Qd4+ 4. Kh1 Qe4! Just so
- it's important to drive the queen away from its best
square: e1. Black gets nothing from 4...Qd2? 5. Rxe5 or
4...Qf4? 5. h4; and 4...Qd3 5. Kg1 is also useless.
5. Qc1. The most stubborn defense. 5. Qd1 Qf4 loses at
once (here's the fatal double attack, now that White no
longer has h2-h4), as does 5. Qg1 Qe2. With the text,
White sets a small trap: the tempting 5...Qe2? would let
slip the win, because of 6. Rf5!.
5...Qd3! 6. Kg1 Qd4+ 7. Kh1 Qd2! White resigned.

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (21 of 27) [04/06/2003 7:15:22 PM]

The Instructor

The following is an example of the difficulties facing a

trainer searching for quality exercises for playing
purposes.
Lilienthal - Aronin
Moscow 1948

White to move
?
25. c5! Thanks to this pawn sacrifice, White's rook gets
through to the 7th or 8th rank, which Nimzovich called
the goal of any operation on an open file.
25... Qxc5 As K.-D. Maier has pointed out, Black could
have offered stiff resistance - perhaps even saved the
game - by giving up his queen for rook and minor piece:
25...Bxc5! 26. Rxb7 Nxd5! 27. Rxc7 Nxc7. This
circumstance may lessen the impression of the exercise;
but it does not negate its value entirely, since there seems
to be no equivalent alternative to the move 25. c5!
26. Rc1 Qa5 27. Bxf6 Rxf6 28. Rc8+ Bd8. White's last
two moves could be transposed, but that's an
insignificant demerit in this exercise. What' s much
worse is that, by rejecting the standard text blow in favor
of 29. Qd1!, White could have won instantly, thanks to
his irresistible threats on the d-file (29...Kf8 30. Bc6!;
29...Rd6 30. Bxf7+ Kxf7 31. Qxd6), and all his
refinements in the game would have been rendered
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (22 of 27) [04/06/2003 7:15:22 PM]

The Instructor

unnecessary. Which means that this position isn't a very

good one for training play - finding the first moves
(including 29. Qd1!) is rather simple.
29. Qc3?! Qb6 30. Qb2! Qd6 30...Qa5 is met decisively
by 31. Qd2!
31. f4! A excellent stroke! The inviting 31. Qxe5 would
have been much less energetic - after 31...Kf8 or
31...Qd7, Black could still defend. The same evaluation
is true after 31. Qxb7 Kf8.
Such would have been my comments - before I read
Maier's analysis. The German analyst has shown that
after 31. Qxb7! Kf8 (on 31...Qf8, 32. f4! becomes still
stronger) 32. Rb8 Ke8 33 Bb3! or 32...g6 33. Qc8 Ke8
34. Bc4, the unbreakable pin on the 8th rank must decide
the game.
31...ef If 31...Rxf4, then 32. Qxe5 (a double - no, more
exactly, a triple attack: Black's queen and rook no longer
defend one another). 31...Qf8 (threatening 32...Bb6+) is
also hopeless after 32. Qxb7 Rd6 33. Kh2!? However,
according to Maier's analysis, Black can put up stiff
resistance with 31...Qd7!? Which means that the
spectacular move 31. f4!? was not the strongest.
32. e5 Qd7 33. Qxb7! Qe8 34. Qb8 Rd6 35. Qxd6
Bb6+ 36. Qxb6 Qxc8 37. e6 g5 38. e7 Kg7 39. Bc6
Black resigned.
P. Rakolta, E. Janoczy

1978

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (23 of 27) [04/06/2003 7:15:22 PM]

The Instructor

White to move
?
Both kings are in trouble, so White must act with great
energy. Black's queen is taboo, because of the first-rank
mate.
1. b4+! Kb6! The pawn cannot be taken either with the
rook or with the queen, in view of 2. Qxf2. And 1...cb is
refuted by 2. Re5+. On 1...Ka4, the winning path is only
slightly longer: 2. Qa6+ Kxb4 3. a3+ Kb3 4, Qb5+ Kc2
(4...Rb4 5. Qd3+) 5. Re2+.
2. bc+ Kb7 2...Kc6 doesn't help: 3. Qa6+ Kd7 (3...Kd5
4. Qe6+ Kxc5 5. Re5+ Kb4 6. Qb3 is mate) 4. Qe6+ Kd8
5. Qg8+! Kd7 6. Qg4+ and 7. Rxd4 - the check is no
longer a threat, as White interposes on d1.
3. c6+! Ka8! Here (as opposed to the starting position),
the back rank is vulnerable on both sides. There follows
a hot finishing duel.
4. Rg4! Rd8 5. Rg8! Qd4 (5...Rxg8 6. Qxf2) 6. Qg4!
Qd2 7. Qc8+! Rxc8 8. Rxc8 mate.
We conclude with two short games, ending with
spectacular thematic attacks (these may also be used for
play-action training, beginning in each case with the
diagrammed positions). It should be noted that the first
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (24 of 27) [04/06/2003 7:15:22 PM]

The Instructor

and most widely known game was probably never

actually played - according to many sources, it was made
up. But this should not prevent us from enjoying this
brilliant combination - or better still, finding it ourselves,
step by step.
Adams - Torre
New Orleans 1920

1. e4 e5 2. Nf3 d6 3. d4 ed 4. Qxd4 Nc6 5. Bb5 Bd7 6.

Bxc6 Bxc6 7. Nc3 Nf6 8.
0-0 Be7 9. Nd5 Bxd5 10. ed 0-0 11. Bg5 c6 12. c4 cd
13. cd Re8 14. Rfe1 a5?! 15. Re2 Rc8? (15...h6) 16.
Rae1 Qd7

White to move
?
17. Bxf6! Bxf6 17...gf was objectively better, when 18.
Rxe7? fails to 18...Qxe7! (18...Rxe7? 19. Rxe7 Qxe7 20.
Qg4+) 19. Rxe7 Rc1+. However, after the simple 18. h3,
Black's position is not to be envied.
18. Qg4! Qb5 19. Qc4!! Qd7 20. Qc7!! Qb5 Now the
direct 21. Qxb7? is refuted by 21...Qxe2! 22. Rxe2 Rc1+.
Nor is 21. Qxa5? good: 21...Qxe2! 22. Rxe2 Rxe2.
White has an elegant way of getting his rook off the e2
square with tempo.
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (25 of 27) [04/06/2003 7:15:22 PM]

The Instructor

21. a4! Qxa4 22. Re4! (threatening 23. Qxc8!) 22... Qb5
23. Qxb7! Black resigned.
Korchmar - Polyak
Ukrainian Championship 1937

1. e4 e5 2. Nf3 Nc6 3. Nc3 Nf6 4. Bb5 d6 5. d4 ed 6.

Nxd4 Bd7 7. 0-0 Nxd4?! 8. Bxd7+ Qxd7 9. Qxd4 Be7
10. Rd1! 0-0 11. e5 Ne8 12. Bf4 a5 13. Rd3 Ra6 14.
Re1 Qf5 15. Nd5 Bd8 16. ed Nxd6 17. Rg3 f6 18. Bh6
Rf7

White to move
?
19. Nb4! The start of the decisive combination. White
gets nothing out of 19. Qc5?! Rd7.
19...ab 20. Qxd6! Qd7! 21. Qd5!! The key move in
White's attack, creating the threat of 22. Rxg7+. On
21...g6, there would follow 22. Rge3.
21...Kf8 22. Rxg7! Qxd5 23. Rg8+! Kxg8 24. Re8+ Rf8
25. Rxf8 - mate.

Copyright 2003 Mark Dvoretsky. All rights reserved.

Translated by Jim Marfia
file:///C|/Cafe/Dvoretsky/dvoretsky.htm (26 of 27) [04/06/2003 7:15:22 PM]

The Instructor

This column is available in Chess Cafe Reader format. Click here for
more information.

[ChessCafe Home Page] [Book Reviews] [Bulletin Board] [Columnists]

[Endgame Studies] [The Skittles Room] [Archives]
[Links] [Online Bookstore] [About The Chess Cafe] [Contact Us]
Copyright 2003 CyberCafes, LLC. All Rights Reserved.
"The Chess Cafe" is a registered trademark of Russell Enterprises, Inc.

file:///C|/Cafe/Dvoretsky/dvoretsky.htm (27 of 27) [04/06/2003 7:15:22 PM]