|Oct-07-02|| ||refutor: i wonder how many times kramnik has played the exact same game v. fritz at home in the past year when he had the *exact same program* to play against as training...wouldn't surprise me if he wins every game with white, and draws every game with black ;) |
|Aug-01-04|| ||acirce: 25.h4?? is laughable. |
|Aug-01-04|| ||michlo: Why?? |
|Aug-01-04|| ||acirce: Because it leads to a simple draw and the program probably didn't even realize it; any human player would have tried to win the position (I don't know if it is possible but there are ways to try). |
|Aug-01-04|| ||Knight13: Now this game became a "pawn kill." |
|Dec-09-07|| ||dumbgai: This game illustrates the difference between machines and humans. Any human player knows that this game is probably drawn, but White can try to win with his extra kingside pawn with no risk, because black's queenside pawns aren't going anywhere. Therefore a strong human player would probably march the king forward, and aim for a f4-f5 pawn push. I think the game would still be drawn anyways, but 25. h4 creates an immediate trivial draw. Ironically, according to the article published in Chess Life, the computer thought it was winning at this point thanks to its extra f-pawn, but the operator knew better and sensibly offered a draw.|
|Jul-11-08|| ||apexin: Game One
Deep Fritz – Vladimir Kramnik (2807)
Brains in Bahrain (1), 04.10.2002 [C67]
Can Kramnik beat Deep Fritz and avenge Kasparov’s defeat against Deep Blue in New York in 1997? The tension was increased by the postponement of this match, but we can hope that Kramnik will strike back first and that Kasparov will win his upcoming match against Deep Junior, so that the ball will be back in the machines court.
A match against a machine is very different from a match against a human, so Kramnik’s strategy will be to aim at closed, controlled positions that are difficult to play for a tactical-oriented computer program. But he could also use his preparation from the World Championship match against Kasparov. The Berlin wall is a good weapon, if you are: A) Kramnik and B) the opponent is a computer program (it also worked very well in his match against Kasparov). See for yourself.
1.e4 e5 2.Nf3 Nc6 3.Bb5 Nf6
The Berlin Defense – also called the Berlin Wall. 3...a6 is the main line. 4.0-0 Nxe4 5.d4 Nd6 6.Bxc6 dxc6 7.dxe5 Nf5 8.Qxd8+ Kxd8
Black’s bishop-pair compensates for his weakened pawn structure on the queenside (the pure pawn ending would be lost!) and White’s advantage in development. There is a heated debate at the moment as to whether Black’s compensation is sufficient, but this is not the interesting question in this match. It is more important how Fritz plays such positions, if it feels at home or not. And so today Kramnik’s choice was correct, especially as I think that Kramnik is satisfied with four draws as Black and will try for more with White. We will soon find out.
6 9.Nc3 h6 10.b3
GM Ronen Har Zvi also played the Berlin in a Man vs. Machine event at Kasparov Chess. He did quite well with it, and I’m sure that Kramnik has studied this game carefully.
10.Bd2 Ke8 11.Rad1 Be6 12.Rfe1 Rd8 13.a4 Bb4 14.Ne4 Bxd2 15.Rxd2 Rxd2 16.Nfxd2 Ke7 17.f3 Rd8 18.Kf2 b6 19.b4 g5 20.g3 Rd5 21.c4 Rd3 22.Rb1 Ra3 23.a5 Ra2 24.g4 Nh4 25.Ke3 Ra3+ 26.Rb3 Rxb3+ 27.Nxb3 Bxc4 28.Nd4 Bd5 29.a6 Ng6 30.Nf5+ Ke6 31.Nxh6 Bxe4 32.Kxe4 Nxe5 33.Nf5 f6 34.Nd4+ Kd7 35.h3 c5 36.bxc5 bxc5 37.Ne2 c4 38.f4 Nd3 39.fxg5 fxg5 40.Kf5 Nf2 41.Kxg5 Nxh3+ 42.Kh5 Kd6 43.g5 Nxg5 44.Kxg5 Kc5 45.Kf4 Kb6 46.Ke5 Kxa6 ½–½, Deep Junior-Har Zvi,R KasparovChess, Internet 2000.
10.h3 Kramnik has done well with this move: 10...Bd7 (10...Ke8 11.Ne4 c5 12.c3 b6 13.Re1 Be6 14.g4 ½-½, Kasparov,G-Kramnik,V London 2000) 11.b3 Kc8 12.Bb2 b6 13.Rad1 Ne7 14.Rd2 c5 15.Rfd1 Be6 16.Ne2 g5 17.h4 g4 18.Nh2 h5 19.Rd8+ Kb7 20.Rxa8 Kxa8 21.Rd8+ Kb7 22.Nf4 Ng6 23.g3 c4 24.bxc4?? An unbelievable blunder by Anand. 24...Nxf4 25.gxf4 g3!! 26.Nf1 (26.fxg3 Bc5+ 27.Kg2 Rxd8o) 26...gxf2+ 27.Kh2 Bxc4 0-1, Anand,VKramnik, V Mainz 2001.
Judit Polgar was able to defeat Garry Kasparov with 10.Rd1+!? in a rapid event. Kasparov was apparently so impressed with Kramnik’s treatment of the Berlin Defense that he introduced it into his own repertoire.
|Jul-11-08|| ||apexin: 10...Ke8 11.h3 Be7 (11...a5 12.Bf4 Be6 13.g4 Ne7 14.Nd4 Nd5 15.Nce2 Bc5 16.Nxe6 fxe6 17.c4 Nb6 18.b3 a4 19.Bd2 Kf7 20.Bc3 Rhd8 21.Rxd8 Rxd8 22.Kg2 Rd3 23.Rc1 g5 24.Rc2 axb3 25.axb3 Nd7 26.Ra2 Be7 27.Ra7 Nc5 28.f3 Nxb3 29.Rxb7 Nc1 30.Nxc1 Rxc3 ½-½, Kasparov,G-Kramnik,V London 2000) 12.Ne2 Nh4 13.Nxh4 Bxh4 14.Be3 Bf5 15.Nd4 Bh7 16.g4 Be7 17.Kg2 h5 18.Nf5 Bf8 19.Kf3 Bg6 20.Rd2 hxg4+ 21.hxg4 Rh3+ 22.Kg2 Rh7 23.Kg3 f6 24.Bf4 Bxf5 25.gxf5 fxe5 26.Re1 Bd6 27.Bxe5 Kd7 28.c4|
c5 29.Bxd6 cxd6 30.Re6 Rah8 31.Rexd6+ Kc8 32.R2d5 Rh3+ 33.Kg2 Rh2+ 34.Kf3 R2h3+ 35.Ke4 b6 36.Rc6+ Kb8 37.Rd7 Rh2 38.Ke3 Rf8 39.Rcc7 Rxf5 40.Rb7+ Kc8 41.Rdc7+ Kd8 42.Rxg7 Kc8 1-0, Polgar,J-Kasparov,G Moscow 2002.
Kramnik voluntarily leaves the d-file and goes for a very solid setup with his bishops on e7 and e6, and pawns at h5 and a5. This position can only be stormed with very long-term plans that are difficult for the computer to find. So it is an ideal position against Deep Fritz, but a disappointment for chess fans who want to see fiery combinations.
Kramnik has certainly studied the following game: 11...a5 12.Ne2 a4 13.Nf4 Be6 14.g4 Ne7 15.Nxe6 fxe6 16.Nd4 (16.Kg2!?) 16...Kf7 17.c4 (17.f4 Nd5) 17...h5!= 18.Nf3 Ng6 19.Ng5+ Ke7 20.Ba3+ Ke8 21.Bxf8 Nxf8 22.Rad1 hxg4 23.Rd4 Rh6 24.Rxg4 axb3 25.axb3 Ra3= 0-1, Klovans,J-Dautov,R Minsk 1986.
Kramnik cleverly holds back his bishop on c8 to retain the bishop-pair. 12...Be6 13.Ne2 Rd8 14.Rxd8+ Bxd8 15.Rd1 Rg8 16.Nf4 g5 17.Nxe6 fxe6 18.g4 Ne7 19.Nd4 Rg6 20.c4 c5 21.Ne2 Rg8 22.Ng3 Ng6 23.Nh5 Be7 24.Kg2 Rf8 25.Kg3 Rf7 26.Rd3 Bd8 27.f3 b6 28.Kf2 Rd7 29.Rxd7 Kxd7 30.Ke3 Ke8 31.Ke4 Kf7 32.Bc1 c6 33.Be3 Nf8 Now the white pawn majority is a clear danger. 34.f4 gxf4 35.Bxf4 Kg6 36.Nf6 Bxf6 37.exf6 Nd7 38.h3 b5 39.f7 h5 40.gxh5+ Kxf7 41.Kd3 Kf6 42.a4 a6 43.a5 Kf5 44.Ke3 Nf6 45.h6 bxc4 46.bxc4 Kg6 47.Be5 Nh7 48.Bf4 Nf6 49.Kf3 Kf5 50.h4 Kg6 51.Be5 Ng8 52.Bg7 Ne7 53.Ke4 Nf5 54.h5+ Kf7 55.Be5 Nxh6 56.Kf4 Nf5 57.Kg5 Kg8 58.Kg6 Nh4+ 59.Kf6 1-0, Klovans,J-Reichenbach,W Berlin 1998. 8 13.a4
|Jul-11-08|| ||apexin: 16.h3 b6 17.Nfd4 Nxd4 18.Nxd4 c5 19.Nxe6 fxe6
The bishop-pair has been halved, but the black position is very solid. An important detail is that the e5-pawn is on the same color square as the bishop on b2. This makes it a weak piece with no prospects on either wing (an important restriction technique). Does Fritz see and understand this? 20.Rxd8+ Kxd8 21.Bc1 Kc8
Kramnik plans to exchange the other rook as well, so that Fritz only has the bishop left.
22.Rd1 Rd8 23.Rxd8+ Kxd8 24.g4 g6
After 24...hxg4?! 25.hxg4, White has the plan f4, Kg1-g2-h3 and Bc1-d2-e1- h4 etc. For this reason Kramnik leaves the white pawn on h3. 9 25.h4?
A human would not play this move, as it leads to an immediate draw. In the post-game press conference, GM Danny King was eager to find out whether such a “poor” move could be avoided by Fritz in future versions. Deep Fritz operator Matthias Feist acknowledged this and provided adequate explanations and excuses. Most were in agreement until the proverbial bubble was burst when Kramnik politely pointed out that 25.h4 was surely White’s best move in the position!
An alternative strategy would have been 25.Kg2 Ke8 26.Kf3 Kf7 27.Ke4, with the idea f2-f4-f5. Kramnik will certainly have checked such endgames, but I am unable to quickly find the correct defense. After seeing the analyses of Notkin and Stertenbrink, I reexamined the endgame and found the following lines: 27...Bd8 (27...hxg4?! 28.hxg4 White plans Kf3-g3-h3 followed by Bg3- h4. So 28...g5 (28...Bh4 29.Be3 (29.f4 Be1 30.f5 gxf5+ 31.gxf5 Ke7 32.Bg5+ Kd7 (Notkin) 33.f6 (33.Bf6 Bc3 34.Kf4 Bd2+ 35.Kg4 Be3=) 33...Ke8 34.f7+ Kxf7 35.Bd8 c6 36.Bxb6 Bb4 37.Kf4 Kg6 38.Kg4 Kh6=) 29...Ke8 (29...g5? is wrong, e.g. 30.Kd3 Kg6 31.Ke2 Kh6 32.Bd2 Kg6 33.Be1 Kh6 34.f3 Bxe1 35.Kxe1 Kg6 36.Kf2 Kh6 37.Kg3 Kg7 38.f4 gxf4+ 39.Kxf4 Kg6 40.Kg3 Kg5 41.Kh3 Kf4 42.Kh4 Kxe5 43.g5 Kf5 44.Kh5 ) 30.Kf3 (30.g5 Bxf2 31.Bxf2 is drawn (Stertenbrink): 31...Kd7 32.Kd3 Kc6 33.Kc2 Kb7 34.b4 axb4 35.a5 Ka6 36.axb6 cxb6 37.Kb3 Kb7 38.Ka4 Kc6 39.Be1 Kc7 40.Kb5 Kb7 41.Bd2 Kc7 42.Ka6 Kc6 and White can’t make progress as 43.Ka7?? b5o even loses.) 30...Kf7 31.Kg2 Be7 32.f4 Bd8 33.Kh3 Be7 34.Bf2 Kg7 35.Bh4 (35.Kg2 Kf7 36.Kf3 Bd8 37.Ke4 Be7 38.f5 gxf5+ 39.gxf5 Bg5=) 35...Bxh4 36.Kxh4 Kh6 37.Kg3 Kg7 38.Kf3 Kf7 39.Ke4 Ke7 40.f5 g5= and Black’s fortress is impregnable because of the reserve tempo c7-c6.) 29.f4 gxf4 30.Bxf4 Kg6 31.Kf3 c6 32.Kg3 Bd8 33.Kh3 with the idea Bf4-g3-h4.) 28.f4 Be7 29.f5 hxg4 30.hxg4 c6 31.f6 Bd8 and the black position should be quite impenetrable, since the following plan does not succeed: 32.Be3 Bc7 33.b4? axb4 34.a5? b3o; 25.gxh5?! gxh5 26.Kg2 Ke8 27.Kf3 Kf7 28.Ke4 Kg6 29.f4 c6 is totally drawn of course.
If Deep Fritz was a human being, one would think that it was hoping for 25...Bxh4??. The danger is 26.g5 and the bishop will never see the light of day. However, Kramnik later told us that he actually considered it, although White can capture the bishop with Kg1-g2-h3, this does not give him a path to penetrate. In fact, Guenter Stertenbrink, Bjoern Frank and Maxim Notkin proved that the position is still drawn after 26...Bxf2+ 27.Kxf2. 26.Bg5 Bxg5 27.hxg5
The pawn endgame is totally drawn, as there is no way for either king into the enemy camp.
27...Ke8 28.Kg2 ½–½
|Oct-13-08|| ||just a kid: The Berlin is very drawish at times.|
|Jul-08-11|| ||wordfunph: 25.h4? or 25.h4!