Mihail Tal
(2757) - Viktor Kortschnoj
(2759)
[E19]
FIDE Candidate
Match,
(Semi-finals) / Game # 01
Moscow, U.S.S.R.
(Russia); 1968.
[A.J. Goldsby I]
This is a very famous draw ...
my main focus here was obviously the K+P endgame.
I saw it analyzed in several
books ... and also a great number of magazines over the years.
(Several authors ... Furman, Averbakh, Smyslov, Botvinnik, Keres, and Gligoric, just to
name a few "chess stars" ...
claimed that Tal missed a clear win, but none of the chess engines really
substantiate this point of view. 5-10 years ago,
the engines were quite unreliable
in the endgame, but, in recent years, the chess programs have made huge strides and
{today} they can
quickly calculate all of the important lines.)
********************************************************************************************
[
The ratings come from the historical ratings website, "Chess Metrics," by the
respected statistician, Jeff
Sonas.
According to that website, Fischer was the #1 player in 1968, Korchnoi was #2, and Tal was the #3 player.
]
1.d4 Nf6; 2.c4 e6; 3.Nf3 b6; 4.g3, (fianchetto)
This is a solid play but it
is also the older move.
(This game was played well before Kasparov came on the scene and showed that a3!
<on move four or five> to
be a path that generally gives White the best chance
at a solid advantage.)
rnbqkb1r/p1pp1ppp/1p2pn2/8/2PP4/5NP1/PP2PP1P/RNBQKB1R b KQkq - 0 4
[ For more on this opening,
see MCO-15, beginning on page # 563. ]
[
I prefer the try: 4.a3, see any good reference work.
{This is the Petrosian System ... many books have
been written on this particular line.}
]
4...Bb7; 5.Bg2 Be7; 6.0-0 0-0; 7.Nc3 Ne4!;
A known antidote, after
this well-timed play by the second party, White has problems trying to really
demonstrate a clear edge.
8.Qc2 Nxc3; 9.Qxc3 c5; ('!?')
So far, this is all a <<standard>>
book line ... ... ...
[ See MCO-15, page # 574; columns # 19-22, and all of the respective notes.
]
rn1q1rk1/pb1pbppp/1p2p3/2p5/2PP4/2Q2NP1/PP2PPBP/R1B2RK1 w - c6 0 10
This move, (9...c7-c5!?) is
probably sufficient for equality, any exchange in the center will lessen
White's dominance there.
(DF14 prefers 9...Be4.)
A fairly good and relatively
recent example of this game would have to be:
GM Evgeny Bareev (2645) - GM Judit Polgar
(2727); / World Champ. Candidates
/ Semi-Finals (4) /
2007.
{Bareev won a long game, 1-0 in 54 moves, however the machine shows many
improvements for Black.}
10.Be3 Bf6; 11.Rfd1!? Bxf3!;
12.Bxf3 Nc6; 13.Bxc6 dxc6;
Temporarily, Black has doubled Pawns, but they are about to be liquidated ...
r2q1rk1/p4ppp/1pp1pb2/2p5/2PP4/2Q1B1P1/PP2PP1P/R2R2K1 w - - 0 14
The current position makes for an interesting diagram. {See
just above.}
White now penetrates to the 7th rank, the result of which is just more exchanges.
14.Qd3 cxd4; 15.Bxd4 c5;
16.Bc3 Qe7; 17.Qd7 Rfd8;
18.Qxe7 Bxe7; 19.e4 h5!?;
This is not a bad move but Houdini's plan of >/= 19...f6; and then 20...Kf7; has to be a little better.
20.Kf1 Bf6!?;
(Too risky?)
Korchnoi gives himself doubled Pawns on the King-side. While not losing, I think it was unnecessary ... the plan outlined after the previous ply-set has to be less risky. (Editor's
note ... the Informant awards this move a full question mark. However,
as this move only causes a tiny disturbance in the engine's overall
evaluation, this is obviously much too harsh a judgment!)
21.Bxf6 gxf6; 22.Ke2 Kg7;
23.Rxd8, (paring-down)
M. Tal, a great master of the endgame, knows that the doubled Pawns in Black's camp means that White will have some real winning chances.
[ RR
23.a3 Kg6; "~" (unclear) ]
23...Rxd8; 24.Rd1 Rxd1; 25.Kxd1,
("=")
A king-and-pawn-endgame has been reached, most engines show that the current position is very close to dead level.
(This would be a good place to have another diagram.)
8/p4pk1/1p2pp2/2p4p/2P1P3/6P1/PP3P1P/3K4 b - - 0 25
If White has any advantage at all, then engines like Houdini 3.0, DF14 and also Deep Shredder ...
show that it is literally only 10-20 one-hundredths of a point. (Which is clearly NOT enough to win the game!)
I spent almost a full day analyzing this K+P endgame ... with the help of several strong chess engines.
Black can probably draw with 25...e5; but I much prefer Korchnoi's move.
25...Kg6!; 26.Ke2 Kg5!?;
There is probably nothing seriously wrong with this move, but here it just doesn't feel right.
(White can kick the BK with h4+ or even f4+ at some point.)
To be fair, however, I should point out that 26...Kg5; IS the top choice of several good chess engines - like Stockfish.
(And a few others.)
[ Deep Fritz 14 gives the following line here: (>/=)
RR 26...f5;
27.Ke3 Kg5;
28.exf5 Kxf5; 29.f3, "+/="
White has a tiny edge ... but it will never add up to anything
that the first player can deposit into the chess bank.
(IMO, its an iron-clad draw.) ]
27.Kf3 f5!?; (hmmm)
Probably ... eventually ... Black has to play this, but it is all a question of timing.
8/p4p2/1p2p3/2p2pkp/2P1P3/5KP1/PP3P1P/8 w - - 0 28
We have reached the critical position, and it is definitely time for another diagram.
In this position, Smyslov once published an analysis in one chess magazine ... which ran several pages. In that work, both Smyslov and Furman strongly asserted that 28.e5! would have won the game for White.
[ RR
27...a6;
28.h4+, "+/="
]
28.h3!?, (Ha, ha, ha!)
The play of h2-h3 ... originally given a question mark by Smyslov and Furman and then the same judgment was agreed upon by Paul Keres ...
turns out to be the #1 or #2 move choice of many chess engines! (As I write this note - several days ago I did an overnight D.P.A. and 28.h3, WAS the FIRST choice of DF14! (It is a flexible, it keeps just about all of White's options open, and this is probably why Mikhail Tal played it in the first place.)
The
section (just below)
is my detailed examination of the main line of this very complex K+P
endgame ...
and is the result of nearly 3 weeks of work. (Refer
to the diagram - just above.)
[ For a very long time, I follow the analysis of Smyslov and Furman, quoted in the book,
"Power Chess," by GM Paul Keres:
RR
28.e5! f6!; 29.h4+! Kg6;
30.Kf4 a6!; 31.a3! b5!!;
Normally, it is a loss to give your opponent an OUTSIDE PASSED PAWN ...
but Black has all of the bases covered here.
32.cxb5 axb5; 33.b3! fxe5+;
34.Ke3!!, (Amazing!)
It is extremely rare for a tactic like this to occur in a pure K+P endgame.
(Normally, to not recapture is a disaster, but here White holds the trump
card of the outside passed pawn.)
(
The following line - found mostly by the chess engines -
is so fascinating that I have
to show it:
RR
34.Kxe5 Kf7; 35.f3 Ke7;
36.f4 Kf7; 37.a4?, (error)
Normally, getting an outside Passed Pawn will guarantee
one side a virtual win, but here is a wonderful exception!
(All White had to do to draw this ending was to find the following line:
>/= 37.b4 c4;
38.Kd4 Ke7; 39.a4 bxa4; 40.Kxc4 Kd6;
41.b5 a3;
42.Kb3 Kc5; 43.Kxa3 Kxb5; 44.Kb3,
"="
with a
fairly obvious draw.)
37...bxa4!;
38.bxa4 Ke7;
39.a5 Kd7; 40.a6 Kc6;
41.a7 Kb7; (zugzwang)
And here ... White finds himself in a rare form of a move dilemma. If the first
player could hop his King over to the c4-square, he would win easily. Since
he cannot, he must move forward ... & lose the game!
42.Kxe6 c4; "-/+"
Black is winning easily. )
Now a careless move ... like playing the Black King to the h6-square ... will lose to a4!
But all is not lost, to maintain the balance, all Black has to do is to return the extra material.
34...f4+!; (Correct!)
Black returns the material, other lines are not convincing. (See the analysis,
just below.)
*** *** *** *** *** *** *** *** ***
*** *** *** *** *** *** *** *** *** *** *** ***
*** *** ***
(
Yuri Averbakh and I Maizelis, in the book: "Pawn
Endings,"
(Chess Digest, published in 1974.) <position/example # 802, beginning on page #267;>
give only the following line of analysis ...
(I stick to just their main variation, and pretty much ignore the many branches here.):
</= 34...Kf6?;
(bad)
This is wrong, the engine
throws a fit just after this move is played!
(Going from equal, to "plus-slash-minus, + 2.52.)
35.a4 bxa4;
36.bxa4 Ke7;
37.Kd3 Kd6; 38.Kc4 Kc6;
39.a5 f4;
40.gxf4 exf4;
41.a6 f3!?;
42.a7 Kb7;
43.Kxc5 Kxa7; 44.Kd6 Kb6;
45.Kxe6 Kc7;
46.Kf5 Kd6;
47.Kf4 Ke6; 48.Kxf3, "+/-"
(+-)
8/8/4k3/7p/7P/5K2/5P2/8 b - - 0 48
White is winning ... you can check this position with any engine you like.
)
*** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** ***
Going back to the our original line of analysis,
(White has to capture, 35.Kd2?!, fxg3;
'/+' will favor
Black.):
35.gxf4 exf4+; 36.Kxf4 Kf6;
37.Ke4 Ke7; 38.a4 c4!;
39.a5 cxb3; 40.Kd3 Kd6;
41.Kc3 Kc6!; (correct)
This simple move draws easily, and all of the engines will confirm this.
*** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** ***
( Our famous crew of analysts only give the following
(grossly inferior) line:
</=
41...Kc5??;
IMO, it is rare scenario when a King move ... that immediately grabs the opposition ...
loses by force!!! (All the strong chess engines "see" this as a blunder, the evals going
from near equal to "+12 to +15!")
42.Kxb3 e5;
43.f3!, "+/-" when White wins easily.
(In case you are not yet convinced that Black is completely lost ...
I offer the following two lines as our ... "the body of proof" ...
43.f3 b4!?;
(human)
I showed this position to a few guys at chess club tonight, they all wanted to play this move.
(Just blindly following the engine moves yields the following variation here:
RR 43...Kd5;
44.a6 Kc6;
45.Kb4 Kb6; 46.a7 Kxa7;
47.Kxb5 Kb7;
48.Kc5 e4;
49.fxe4 Kc7; 50.Kd5 Kd7;
51.Ke5 Ke7; 52.Kf5 Kf7;
53.Kg5 Kg7; 54.Kxh5, +-)
44.a6 Kb6;
45.Kxb4 Kxa6;
46.Kc5 Kb7; 47.Kd5, +-) )
*** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** *** *** *** ***
*** *** ***
(We return to the main stem of our analysis line here.)
42.Kxb3 Kc5; 43.Kc3 b4+;
44.Kb3 Kb5; 45.a6 Kxa6;
46.Kxb4 Kb6; (stand-off)
This is the point of 41...Kc6!
---> When the dust finally settles, the Black King can hold the opposition and still maintain the draw.
47.Kc4 Kc6; 48.Kd4 Kd6;
49.Ke4 Ke7; 50.Ke5, (opposition)
The most direct try.
( Or White could try: RR
50.Kf4 Kf6; 51.f3 e5+; 52.Ke4 Ke6;
"="
which is also a draw. )
50...Kf7; 51.f3 Ke7; 52.f4 Kf7;
"=" (100% equal!)
It is a DEAD DRAW ... all of the engines evaluate this position as "0.00," which is as drawn as it gets!!!
8/5k2/4p3/4K2p/5P1P/8/8/8 w - - 0 53
This analysis diagram should convince even the most hardened skeptic.
(Even if White exchanges off the f-Pawns and wins Black's h-Pawn, it is still a draw as long as the
Black King can reach the important f8-square.) ]
After a deep analysis of some of the potential K+P endings, we return to the actual game.
28...Kf6; 29.Kf4 e5+; 30.Ke3 a6;
31.b3,
This is actually the first choice of DF14, but Smyslov, Furman and Keres seem to think it was clearly inferior!
[ In the book, "Power
Chess," Keres claims that the following
line was a big improvement over the actual game:
31.a3! Ke6!?;
This is OK but Houdini shows that 31...b5! (right away) might be even better.
32.exf5+ Kxf5; 33.f3
f6; 34.g4+!?, (Premature?)
The line given by Furman, and later improved upon by Smyslov, but the engines don't seem to like it.
(Probably a little more tricky was the following line here:
RR 34.h4 Ke6;
35.g4 f5; 36.g5 Kf7;
37.b4 a5;
38.bxa5 bxa5;
39.Kd3 Ke6; 40.Kd2, "+/="
The engines give White almost a third of a Pawn ...
but its is all nonsense,
as it is an easy draw ... by "mutual terror,"
as as GM Arty Bisguier once described a similar situation of a
K+P endgame at a U.S. Open.
(Neither King can leave the square of the pawn, the BK has to keep
an eye on the g-Pawn, while the WK has to stick around to watch
the Black Pawn duo on e5 and f5.) )
34...hxg4; 35.hxg4+ Ke6!;
This is a HUGE improvement over the line, (with 35...Kg6?!); as given by Keres, Smyslov, etc.
36.Ke4 b5!; 37.cxb5 axb5;
38.b3 Kd6; 39.Kd3 Kc6!;
"="
with a VERY drawn position. (Check it with any engine.)
*** *** *** *** *** *** *** *** *** ***
*** *** *** *** *** *** *** *** *** ***
The continuation of: 31.exf5 Kxf5;
32.f3 Ke6;
33.g4 f5; 34.g5,
"+/=" 34...Kf7;
transposes to a variation already examined here. ]
The rest of the game is pretty much a draw, I saw no "blips" in the engine evals ... which would indicate that one side (or the other) missed an easy win or made a significant mistake in this endgame.
31...Ke6 ; 32.exf5+ Kxf5;
33.f3 Ke6 ; 34.g4 f5; 35.gxf5+ Kxf5;
36.h4 Kf6; 37.Ke4 Ke6; 38.a3 b5;
39.cxb5 axb5; 40.Kd3 Kd6; "="
8/8/3k4/1pp1p2p/7P/PP1K1P2/8/8 w - - 0 41
Its now a draw ... (Engines at "0.00" again.)
A wonderful endgame ... but it seems only one that a chess engine could really fathom.
Copyright (c) A.J. Goldsby, 2014. All rights reserved.
˝ - ˝
BIBLIOGRAPHY:
-
The
book:
"Power Chess,"
by GM Paul Keres.
(Chapter/game #19,
page # 212.)
Copyright 1991. ISBN: 0-812-91949-1 David McKay,
publisher.
-
The
book:
"Pawn Endings," by GM Yuri Averbakh and
I. Maizelis,
(Published by Chess Digest, 1974. Copyright (also)
in 1974.) <position/example # 802, beginning on page #267.>
(The link given here is NOT the
same book as mine. However, I have both books and I am pretty sure
that the info is pretty much the same.)
-
This
game is also analyzed in the <Yugoslavian publication> The INFORMANT.
(Informant #05, game # 680.)
Click here
to see a {full pictorial} copy of that
analysis.
Further Study / and More Resources
-
A very nice, instructive article
on the endgames. (Wikipedia.)
-
A few of the
most basic endgames ... all are thoroughly analyzed. (More
basic endgames.)
-
My endgame
school has many K+P endgames and also many basic
positions that every player should know.
-
The
top ten rules of endgames ... basic endgames for all
beginner's.
-
Hundreds (thousands?) of endgame
puzzles ... play against the computer! (Click here.)
-
Practice your basic endgames and basic mating techniques. (Click here.)
-
Click
here
to see a blog I did on another website ... many interesting K+P
endgames are covered there.
(It was also that blog that was the inspiration for this web
page.)
Below is a copy of an email that I received, and as it pertains to this
web page, I did not think Pal would mind if I posted it.
(I invited USCF
to have another GM write about this web page, but I guess they were too
chicken to take on such a complicated task.)
|