|
|
Man versus Machine / "Brains in Bahrain" Match; (Game # 2) Manama, Bahrain; 2002. |
                      |
Kramnik (as White) draws first blood.
The game - a seemingly very quiet
line of the Q.G.A, (Queen's Gambit
Accepted); seems all
set to repeat
the course of the first game and
end in another fairly lifeless draw.
But the game heats up, fueled by
a slight inaccuracy by Fritz. Then
Kramnik plays a stunning
combination, and seems all set to win material.
The computer deftly returns all the
material, and
nearly equalizes the
position. Once again, we seem
headed for another draw. But in a
Rook-and-Pawn ending ... where
the box must defend VERY carefully
and very accurately ...
Fritz goes
astray. Black weakens his Pawn
position, and Kramnik simply slowly
increases the
pressure, until the
computer breaks. In the end, White
decides to enter an easily won K+P
ending.
***
The comments that are contained
within the brackets, ... << blah-blah-blah-blah >>
are from the short version of
this game - that you can download
from my web site.
Click here to see a fairly complete explanation of the symbols that are used.
1.d4
d5; 2.c4
dxc4; {Diagram?}
A Queen's Gambit Accepted.
(Q.G.A.)
In this variation, Black gives up
the center - but gets much freer
piece play than he achieves
in
the more cautious lines of the
Q.G.D.
3.Nf3
Nf6; 4.e3
e6; 5.Bxc4
c5; 6.0-0
a6; {Diagram?}
Black prepares the pawn advance
(space grabber) of ...b5.
So far it has been all a (main) book line.
<< So far, so book. >>
7.dxc5!?, (Maybe - '!')
{Diagram?}
White exchanges in the center and
allows the Queens to be swapped.
Normally, I would greatly
disapprove
of such a method ..... but here it is a
fairly good anti-computer strategy.
Normally White plays Qe2 or a4 here.
The great Botvinnik's favorite move
from this position
was usually a4.
(But Kramnik has a lot of experience in this this line as well.).
Why is this a good idea? Programs
excel at tactics and sharp positions.
They play closed/quiet
positions
relatively poorly and also are not
at Kramnik's level in terms of
basic positional
understanding and
also are not great at many endings.
Basically Kramnik is intelligently playing to his strengths and to the program's potential weaknesses.
At the super-GM level, a move such dxc5!? would probably be interpreted as a draw offer.
Note that Kramnik played this same,
exact move against GM Garry Kasparov,
and almost won with it.
(Brain Games WCC, game # 4. A long & very difficult draw.)
<< An move obviously designed to deaden the game and take the emphasis off tactics. >>
***
[ The <main> book line is: 7.Qe2
b5; 8.Bb3
Bb7; 9.Rd1
Nbd7;
10.Nc3
Qb8;
{Diagram?} The new main line.
(As a youngster, I played these
lines. {As Black.} One book I had
at that time had me
convinced
that ...Qc7; was the best move
here. Black actually has maybe
more than
half-a-dozen different,
completely playable moves here.)
11.d5
Nxd5; 12.Nxd5
Bxd5; 13.Bxd5
exd5; 14.Rxd5
Be7; {Diagram?}
The end of the column. 15.e4
Qb7; 16.Bg5!
Nb6; 17.Rad1!!
f6!; 18.Bf4
0-0!;
19.R5d2
Rad8!?; "=" {Diagram?}
with an approximately level game.
GM L. Christiansen - GM R.
Hubner; Germany, (Bundesliga?) 1992.
(Black is fine here, and actually went
on to win a very long end-game. But
that does
not change the verdict that
this position is roughly equal.)
[ See MCO-14, page # 445,
column # 1, and note # (e.) ].
White could also play: 7.a4!?, ('!')
{Diagram?} which gains space
and also
pretty much stops ...b5.
(But greatly weakens b4.) ].
***
7...Qxd1; {Diagram?}
This is probably the best move
for Black in this position.
[ A clearly inferior line is: 7...Bxc5?!;
8.Qxd8+
Kxd8; 9.Bd2,
"+/" {Diagram?}
and White is much better.
Interesting is:
7...Nbd7!?; 8.c6!,
"+/=" {Diagram?}
This is the smartest move for
White; Black is left with a
broken Q-side pawn array.
(The continuation: 8.b4!?
a5; 9.Qa4
Qc7; 10.Na3
Qb8; 11.Nb5,
"+/=" {Diagram?}
gives rise to some wild tactics.)
8...bxc6; 9.Nc3,
"+/=" ("+/") {Diagram?}
White is clearly better.
Not only
is he ahead in development and
has more space, but Black's pawns
have
been busted up as well. ].
8.Rxd1, {Diagram?}
"Looks unambitious, but Kramnik
knows how to keep increasing {his
advantage} in such
positions." - GM Karsten Müeller.
[ A mistake is: 8.Bb5+?? Qd7; "-/+" {Diagram?} and Black wins. ].
8...Bxc5; {Diagram?}
A search of ChessBase's on-line
database reveals over 400 master
games have been played
from
this particular position!
It also reveals that Kramnik has
played the White side of this set-up
against such players as
GM's Kasparov,
Karpov, Anand, Lautier, and many
others.
(See ChessBase's on-line analysis
of this game for a few examples.)
9.Kf1!?, {Diagram?}
Several annotators call this an
attempt by Kramnik to get the
stupid box out of 'book' ...
and 'thinking on its own.'
(And they are correct.)
Kramnik can play slowly to increase
his advantage, and the computer will
have few games
in its memory on
which look to for guidance.
<< A move that is almost guaranteed to get the box out of its dreaded 'book' lines. >>
"The first new move of the game." - GM Danny King.
*****
[ The 'normal' book line is: 9.a3
b5; 10.Be2
Bb7; 11.b4
Be7; 12.Bb2
Nbd7;
13.Nbd2
Rc8; 14.Rac1
0-0; "~"
when the symmetrical and very
balanced
pawn-structure makes
it difficult for White to stir up a
real advantage.
(Most master games are drawn
not long after this position.)
See the encounter: GM L. Lengyel -
GM L. Barczay;
Hungarian Championship Tournament
Budapest/HUN/1977.
(Drawn in just a few more moves.)
***
The other main move here is: 9.Nbd2
b6!?; {Diagram?}
Black foregoes the big advance,
in favor of a more modest development on the Q-side.
The move, 9...b5!?; {Diagram?} will probably transpose to those lines with 9.a3.
THE (!) most theoretically significant
game - of current GM praxis - in
this line has to
be: 9...0-0; 10.Be2
Rd8; 11.b3
b6; 12.Bb2
Bb7; 13.Rac1
Nbd7; "=" {Diagram?}
(White may have a very, very
small edge in this position.)
1/2-1/2 (Drawn in 44 moves.)
GM A. Karpov - GM V. Ivanchuk;
Russia vs. The R.O.W;
/Moscow/RUS/2002.
(Kramnik also reached this
same position against GM V.
Anand in his Advanced
Game
{rapid} in 2002.)
10.Be2
Nbd7; 11.Nc4
Bb7; 12.b3,
"+/=" {Diagram?}
White only has the tiniest of
advantages here, but went on
to win in 53 pretty, but
somewhat grueling moves.
GM V. Kramnik - GM A.
Karpov; Frankfurt/GER/1999. ].
*****
9...b5; 10.Be2
Bb7; 11.Nbd2
Nbd7; {Diagram?}
Thus far White has played well,
but Black's defense is has been
more than adequate.
12.Nb3
Bf8?!;
(Maybe - '?')
{See the diagram, just below}
Kramnik's brilliant anti-computer
strategy is rewarded with this very
silly move.
The square e7 would
have been sufficient for the Bishop.
'?' - GM Danny King. '?!' - GM Karsten Müeller.
(One GM - in the pressroom - is
said to have practically fallen out of
his chair laughing
when he saw this move. Khalifman's on-line news service called it one of
the silliest
moves he had ever seen, While a commentator on London's
Chess Centre called this
move, "A gross blunder ... of only the kind a computer could
play.")
<< A stupid move for which there
is almost no rational justification
for.
(A human would be happy with placing this piece on b6,
or even e7.) >>
|
|
[ Much better would have been: >= 12...Bb6; "~" {A.J.G.} ].
***
<< White now correctly attempts to
open the game before Black
can maneuver his pieces
to
squares that will allow him to cope
with this threat. >>
13.a4!, {Diagram?}
White finds the correct method.
'!' - GM Karsten Müeller.
[ Also good is: 13.Bd2!?, "+/=" and White is a tad better. ].
13...b4!?; {Diagram?}
Black opts for keeping the game
closed. This might be the correct
decision, but
I am not 1000%
convinced of this.
[ The move 13...bxa4!?;
"~" {Diagram?}
would have avoided the weakness
on a5 that haunted Black for the
remainder of the entire game. ].
14.Nfd2!?, (Maybe - '!')
{Diagram?}
This is good, but I liked Na5 here.
"Kramnik
mobilizes his knight to
occupy the square c4, which he
has just secured.
From here it
can keep a dangerous watch on
the Black position."
- GM Karsten Müeller.
[ Interesting was: 14.Na5!?,
"+/=" (Maybe - '!') {Diagram?}
and White is also a little better
in this position. ].
14...Bd5!?; {Diagram?}
<< Perhaps a further waste of time? >>
Several problems for Black here:
# 1.) His earlier bone-headed play
of Bf8?! ('?')
# 2.) The programming prevents
Black from simply playing the move,
...Be7; and admitting
his earlier move
was a mistake.
# 3.) His pieces do not really work
very well together in this position.
The move ...Bd5 is only a short-term
fix! In the long-run, this move does
not really
help Black and could be
viewed simply as a loss of tempo.
***
[ Maybe better is: 14...Be7; "~" {Diagram?} and try to develop.
Also good was: 14...Nc5!?; {Diagram?} and Black is probably O.K. ].
***
<< White continues to simply try
and slowly improve his position. >>
15.f3!, {Diagram?}
White literally had MANY playable
moves here, but this is the most
forceful,
preparing e4 ... evicting
Black's well-posted Bishop on d5.
[ 15.Na5!?, "~" ].
15...Bd6!?; 16.g3
e5; 17.e4
Be6; 18.Nc4
Bc7; 19.Be3
a5!?; {Diagram?}
<< The pawn is unhappy here.
(It will need constant defense.) >>
(See GM's K. Müeller's rather
lengthy comments on the
excellent ChessBase
web page
on this game.)
[ After: 19...0-0!?; 20.Nba5, "+/=" {Diagram?} and White is a little better here. ].
<< White continues to play very
strongly and also try to improve
his position. >>
20.Nc5!?, (Maybe - '!')
{Diagram?}
This appears somewhat on the
over-aggressive side, but is
a very fine play by Kramnik.
[ Also good for White was:
20.Rac1!?,
"+/=" {Diagram?}
and White is just a little better.
- GM Karsten Müeller. ].
20...Nxc5; 21.Bxc5
Nd7; ('!?')
{Diagram?}
This is probably the best approach
for the second player here.
[ Even worse was the rather
inferior continuation of: 21...Rd8?!;
22.Nd6+
Bxd6;
23.Bxd6
Nd7; 24.Bc7
Ra8; 25.Bb5
f6; 26.Rd3,
"+/" {Diagram?}
and Black is pretty tied up. ].
22.Nd6+
Kf8; {Box? Diagram?}
This appears to be forced.
[ Much worse for Black was: 22...Ke7?;
23.Nf5+
Kf6; 24.Be7+
Kg6;
25.Rac1,
"+/-" {Diagram?}
White is very clearly better.
- GM Karsten Müeller. ].
23.Bf2!, {Diagram?}
White saves a valuable tempo.
(On e3 the Bishop might need
to be defended by Kf2 in many
lines.)
[ Probably not as accurate was:
23.Be3!?
Bb6; 24.Kf2,
"=" {Diagram?}
White might have a tiny edge
here. {A.J.G.} ].
23...Bxd6!?; {Diagram?}
Black surrenders the two Bishops.
[ Was 23...Ke7!?, any better? ].
24.Rxd6
Ke7; 25.Rad1!!,
{See the diagram, just below.}
This move was either a product of
Kramnik's brilliant calculations or
the genius of his intuition.
(I thought he might have to retreat the Rook to avoid any of the box's tricks in this position.)
|
|
[ Also good is: 25.Rd2!?,
"+/=" {Diagram?}
with a small advantage for the
first player in this position. ].
25...Rhc8; 26.Bb5
Nc5!?; {Diagram?}
<< Many players - including many
Masters - thought this was an
oversight.
But the computer will
not miss many tactical plays! >>
"This looks like a tactical blunder.
But we must remember that Black
is a computer that will
never overlook
something as shallow as this."
- GM Karsten Müeller.
27.Bc6!, {Diagram?}
White had several plausible moves
in this position but this is up and
away the best play
for White.
[ 27.Rc6!?, "+/=" ].
27...Bc4+!; {Diagram?}
<< Did Kramnik really miss this
simple move, as many reports
have claimed?
(I find this hard
to believe.) >>
The computer finds the best defense.
A couple of GM's thought Kramnik
might
now experience real
difficulties from this position.
[ 27...Ra6??; 28.Bxc5, "+/-" ].
<< One Internet player confided to
me {he thought} Kramnik's next
move was a blunder. >>
(The next series of moves is
actually a very fine combination
by Kramnik.)
28.Ke1!
Nd3+; 29.R1xd3!
Bxd3; 30.Bc5!!, {Diagram?}
<< After this astounding shot, I was
gleefully predicting a quick and easy
win for White.
Almost unbelievably -
the nasty little box finds a way to
wiggle out of almost all of its
difficulties in this position. >>
'!' - GM Karsten Mueller.
[ I think many players had really expected the move: 30.Rxd3, in this position. ].
30...Bc4!; {Diagram?}
The computer finds the best
defense - based on a hidden
tactical idea - that was
missed by
many (most) humans here.
[ White is much better after:
30...Bc2; 31.Kd2
Bb3; {Diagram?}
This looks forced.
(31...Bxa4??; 32.Rd5+
Kf6; 33.Bxa4,
"+/-" - GM K. Müeller.) 32.Rd3+
Kf6;
33.Bxa8
Be6; 34.Bxb4!
Rxa8; 35.Bc3,
"+/" {Diagram?} White is clearly a lot better
in
this ending. Additionally, many
of Black's weaknesses are fixed
on dark squares,
where White can
easily attack them. ].
31.Rd4+
Kf6; {Diagram?}
<< This appears forced. >>
[ Really a blunder would be: 31...Ke6?;
32.Bd7+, {Diagram?}
and White is much better here.
( "+/" ... maybe "+/-") ].
32.Rxc4
Rxc6; {Diagram?}
This appears forced. (Black tries to regain material equality.)
33.Be7+
Kxe7; 34.Rxc6,
"+/=" {See the diagram, just below.}
<< White maintains a small advantage
in this ending, but I do not think it
should
have been enough to win. >>
"The resulting rook endgame is better
for White because the Black Rook is
tied to the passive
defence of the
a-pawn. One of the main principles
of rook endings is activity. The rook
is
useless if it is engaged in passive
defensive tasks."
- GM Karsten Müeller.
|
|
34...Kd7; {Diagram?}
This looks like the only move
for Black in this position.
(...Ra7; possibly allows White
to later play Rc8.)
[ 34...Rd8!?; 35.Ke2!?,
"+/=" {Diagram?}
... "doesn't help."
- GM Karsten Müeller.
(But 35.Rc5, - or even to a6 -
seems a mite better. {A.J.G.)
Maybe Black should have at
least considered trying: 34...Ra7!?; {Diagram?}
with a passive but flexible
defense. {A.J.G.} ].
35.Rc5
f6!?; (Possibly
'?!') {Diagram?}
<< This move was harshly condemned
by many Masters, but is not as bad
as they
make it out to be. >>
(But ... to be honest, it is far from the best defense!)
'?' - GM Karsten Müeller.
To be fair, a computer is programmed
to find a material win. A computer would
NEVER
sacrifice a pawn for play.
(Not in an ending, anyway.) And in the
main lines of Noitkin's
analysis, Black
is three pawns down ... but will draw
the game. It would defy logic if Fritz
could play this way.
[ Much better is:
>= 35...Rc8!!; <=> {Diagram?}
- M. Noitkin.
See "Chess Today," # 699.
(His full analysis runs over three
pages ...
I will spare you the
complete version.) ].
36.Kd2
Kd6; 37.Rd5+
Kc6; 38.Kd3
g6?!; (Maybe - '?')
{Diagram?}
<< This only weakens Black's
position, according to several
GM's here. >>
'?!' - GM Karsten Müeller.
[ A much tougher defense here was: 38...Ra7!; {Diagram?} - GM Karsten Müeller. ].
***
<< In the next phase of the game,
Black continues to weaken his
King-side Pawn
formation ...
almost guaranteeing a loss. >>
39.Kc4
g5!?; {Diagram?}
<< Maybe dubious. >>
The more Black moves these
King-side pawns, the easier
it is for White to attack them.
This type of ending calls for a
passive defense, a concept the
program has yet to 'learn.'
[ Much better was:
>= 39...Ra7;
40.Rd8, "+/=" {Diagram?}
White is a little better ... but there
is no guaranteed win. ].
40.h3
h6?!; 41.h4!
gxh4!?;
{Diagram?}
Black continues to open lines ...
and also break up his King-side.
I think the average IM could have defended this ending better than Fritz did.
42.gxh4
Ra7; 43.h5;
(Maybe '!') {Diagram?}
<< Kramnik slowly increases his
advantage and now threatens
Rd8 and Rh8xP/h6 ...
gaining a dangerous passed pawn. >>
(GM K. Müeller says pretty much the same thing on the ChessBase page.)
[ Maybe also good for Kramnik
was the move: 43.Rd8!?,
"+/-" {Diagram}
with the better position for
the organic-based intelligence. ].
43...Ra8; 44.Rc5+!?,
('!') {Diagram?}
Kramnik starts a nice maneuver
that puts Fritz in a bind.
[ Maybe 44.Kd3!?,
"+/=" {Diagram?}
with the idea of Ke3 and f4.
- GM Karsten Müeller. ].
44...Kb6; 45.Rb5+, {Diagram?}
All part of the plan.
[ GM K. Müeller points out: 45.Kd5??
Rd8+; 46.Kc4
Rd4+;
47.Kb3
Kxc5; "-/+" {Diagram?}
when Black is winning. ].
45...Kc6; {Diagram?}
This looks forced.
[ Simply bad is: 45...Ka6?;
46.Rd5, "+/-" {Diagram?}
when White is much
better ...
and possibly even winning. ].
46.Rd5!, {Diagram?}
<< Black is in Zugzwang. >>
[ Also possible was:
46.f4!?
exf4; 47.Rf5,
"~" {Diagram?}
with an unclear position.
(Possibly this is just a tiny bit
better for White here.) ].
46...Kc7?!; (Maybe - '?')
{Diagram?}
<< After this extremely passive move,
White is probably winning. >>
'?!' - GM Karsten Mueller. '?' - many others.
[ Good, bad or ugly, Black
simply had to play: 46...Ra7;
which is practically the only
move {box} for Black here. ].
<< White's play in the remainder of
this game is flawless - and shows ...
at least one
reason - why GM
V. Kramnik is the World Champ. >>
47.Kb5
b3!?; 48.Rd3
Ra7; 49.Rxb3
Rb7+; 50.Kc4
Ra7; {Diagram?}
Black cannot exchange Rooks.
[ GM K. Müeller points out the
following variation:
50...Rxb3?; 51.Kxb3
Kb6;
52.Kc4
Kc6; 53.b4,
"+/-" {Diagram?}
and White has an easy win
with his
outside-passed-pawn
in the pure K+P end-game. ].
51.Rb5
Ra8; 52.Kd5
Ra6; 53.Rc5+
Kd7; 54.b3
Rd6+!?; {Diagram?}
The computer predictably goes
for some activity in this ending.
[ White also wins after:
54...Ra7!?; 55.Rb5
Ke7; 56.Kc6
Ra8; 57.Kb7
Rd8;
58.Rxa5,
"+/-" {Diagram?}
- GM Karsten Müeller. ].
55.Kc4
Rd4+; 56.Kc3
Rd1; 57.Rd5+, {Diagram?}
<< The King-and-Pawn ending
is - quite simply - hopeless for
the second player
from this position. >>
ChessBase
provides the following excellent analysis:
< 57.Rd5+ the transition into the
pawn ending is the simplest way
to victory, for example;
57...Rxd5; 58.exd5 Kd6; 59.b4 axb4+;
60.Kxb4 Kxd5; 61.Kb5 Kd6; (61...f5; 62.a5 e4;
63.fxe4+ fxe4;
64.a6 e3; 65.a7 e2; 66.a8Q+, "+/-") 62.a5 f5; 63.a6 Kc7; 64.Kc5 e4;
65.fxe4 fxe4; 66.Kd4 Kb6; 67.Kxe4 Kxa6; 68.Kf5 Kb6; 69.Kg6 Kc7;
70.Kxh6 Kd7;
71.Kg7, "+/-" >
- GM Karsten Müeller.
<< A very nice ending for Kramnik.
This win gave the first victory to
the humans in this contest. >>
(But surely the entire team of
Fritz's programmers had to be very
dismayed by the under-performance
of Deep Fritz in this ending!!)
Please visit my web page(s) at:
https://www.angelfire.com/fl5/
human_fan02/index.html
for a
better look at this whole match.
(The Kramnik - Deep Fritz event
has its own section.)
Every game
is annotated in depth.
***
Copyright (©) A.J. Goldsby I. Copyright (c) A.J.G; 2002.
(Code initially) Generated with ChessBase 8.0
1 - 0
I
went to great pains to annotate this game ... NOT to benefit a master, but to
make it possible
for the average player to follow this game. I also tried to make this a
game a brief survey of the
opening, for any aspiring player who is trying to learn - and might not have
access to many books.
Click HERE to return to the page for Kramnik vs. Deep Fritz, Game # 2.
Click HERE to go to (or return to) the main (parent) page for the Kramnik vs. Deep Fritz Match.
Click HERE to go to (or return to) my Main (home) page for this site.
This
is a game I annotated especially for these web pages. I eventually plan to
annotate - in a similar
manner ... all of these games. If you would like a copy of this game to help you
study, contact me.
Page first posted, late October, 2002. Page last updated: January 15th, 2003.
Copyright (c) LM A.J. Goldsby I
Copyright (©) A.J. Goldsby, 2001-2005.
Copyright (c) A.J. Goldsby, 2006. All rights reserved.
