Saturday, November 25, 2006

Man versus Machine: Kick-Off!

Everyone's talking about it.

No, not EastEnders, not the cricket, and not Streatham & Brixton's new chess blog. I mean of course the Kramnik - Deep Fritz match, game one of which kicks the match off today at 2pm. I recommend the excellent site Doggers-Schaak's rather interesting preview - it includes predictions from top players, a profile of the computer, details of the unusual rules, some slightly odd photo's, and quotes from Kramnik's press-conference. Let's hope this one in particular proves accurate:

"This computing monster keeps getting better year by year, month by month, day by day: My opponent will be incredibly strong. But I think I can still beat it."

Of course, there'll be live coverage all over the internet and chess servers. I'll be watching it here - whilst Chessgames's coverage might (who knows?) offer a more sophisticated level of kibitzing, compared to the usual in-jokes and chatter on the ICC & Playchess. Apart from that, you can find out more about Kramnik from his home-page, whilst the official site for the match is here. I'll post the moves of the game up once it's done.

> > > UPDATE > > >

And so - it was a draw.

Kramnik played white, and here are the moves:

1. d4 Nf6 2. c4 e6 3. g3 d5 4. Bg2 dxc4 5. Qa4+ Nbd7 6. Qxc4 a6 7. Qd3 c5 8. dxc5 Bxc5 9. Nf3 O-O 10. O-O Qe7 11. Nc3 b6 12. Ne4 Nxe4 13. Qxe4 Nf6 14. Qh4 (When a computer offers you two rooks for a queen, it's probably a good idea to decline.)14. ... Bb7 15. Bg5 Rfd8 16. Bxf6 Qxf6 17. Qxf6 gxf6 (Black's activity and slight lead in development will not last for long; therefore, white has the better endgame due to the battered black pawns.) 18. Rfd1 Kf8 19. Ne1 Bxg2 20. Kxg2 f5 21. Rxd8+ Rxd8 22. Nd3 Bd4 23. Rc1 e5 (Moves like 20. ... f5 and 23. ... e5 are typical of a common problems computers have with pawns in endgames: they are often too enthusiastic to move them.) 24. Rc2 Rd5 25. Nb4 Rb5 26. Nxa6 Rxb2 27. Rxb2 Bxb2 28. Nb4 Kg7 29. Nd5 Bd4 30. a4 (Fixing the pawn on b6. But might it have been more exposed on b5 or b4 anyhow? Moves 30 and 31 are critical for determining a winning plan for white - and maybe he got it wrong. E.g., or the analysis here.) 30. ... Bc5 31. h3 f6 32. f3 Kg6 33. e4 h5 34. g4 hxg4 35. hxg4 fxe4 36. fxe4 (Maybe white should have penetrated with his king on the queenside before this simplification, or prepared e4 but without g4.) 36. ... Kg5 37. Kf3 Kg6 38. Ke2 Kg5 39. Kd3 Bg1 40. Kc4 Bf2 41. Kb5 Kxg4 42. Nxf6+ (42. Nxb6 Bxb6 43. Kxb6 f5 is trivially drawn.)42. ... Kf3 43. Kc6 Bh4! (Now the draw is crystal clear.) 44. Nd7 Kxe4 45. Kxb6 Bf2+ 46. Kc6 Be1 47. Nxe5 Game drawn 1/2-1/2


Anonymous said...

Thanks for the nice links :)

You can also use (you can sign up for a 7-day trial account), which will have live radio coverage of all the matches.

Tom Chivers said...

You're welcome - and thanks for that tip too. I'll definitely check it out.

doggy said...

Hi Tom, thanks for mentioning me! I'd like to ask you something - can you email me? doggy {at}

Tom Chivers said...

You're welcome - your post was the best I found!

Tom Chivers said...

Hi everyone

Here are two games from earlier this week where Fritz 10 played Gata Kamsy twice on the ICC, at the time limit of 3+1. Kamsky's ICC handle is 'Talion', Fritz 10 in this case 'BigMomma'.

They are both quite interesting games, in completely different ways. The first sees black (Fritz) crashing through the centre then mating with a neat tactic.

The second shows how clueless computers can be in the endgame - look how many times Kamsky plays Kd7-c7 in a row. One might assume the Fritz operator knows little about chess or otherwise would have accepted a bored Kamsky's draw offers. And one might also assume BigMomma will soon be on many no play lists due to this kind of time-wasting.

[Event "ICC 3 1"]
[Site "Internet Chess Club"]
[Date "2006.11.23"]
[Round "-"]
[White "Talion"]
[Black "BigMomma"]
[Result "0-1"]
[ICCResult "White resigns"]
[WhiteElo "3191"]
[BlackElo "3586"]
[Opening "King's Indian: Torre attack"]
[ECO "A48"]
[NIC "QP.01"]
[Time "18:04:26"]
[TimeControl "180+1"]

1. d4 Nf6 2. Nf3 g6 3. Bg5 Bg7 4. Nbd2 d5 5. e3 O-O 6. c3 Nbd7 7. Be2 Re8 8.
O-O e5 9. Rc1 c6 10. b4 h6 11. Bh4 Qe7 12. b5 e4 13. Ne1 c5 14. Nc2 b6 15.
c4 Bb7 16. Nb3 Rac8 17. dxc5 bxc5 18. a4 Ne5 19. Na3 d4 20. Qd2 Nf3+ 21.
gxf3 exf3 22. Bd3 Qe6 {White resigns} 0-1

[Event "ICC 3 1"]
[Site "Internet Chess Club"]
[Date "2006.11.23"]
[Round "-"]
[White "BigMomma"]
[Black "Talion"]
[Result "1/2-1/2"]
[ICCResult "Game drawn by the 50 move rule"]
[WhiteElo "3583"]
[BlackElo "3194"]
[Opening "QGD Slav: Süchting variation"]
[ECO "D15"]
[NIC "SL.03"]
[Time "17:56:51"]
[TimeControl "180+1"]

1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 Qb6 5. Qb3 Bg4 6. Qxb6 axb6 7. cxd5 Bxf3
8. gxf3 Nxd5 9. Nxd5 cxd5 10. e4 e6 11. Be3 Nc6 12. a3 Bd6 13. Rc1 f5 14. e5
Be7 15. Bb5 Kd7 16. Ke2 g5 17. f4 g4 18. f3 h5 19. a4 Rhg8 20. Rhg1 Rg7 21.
Bf2 Rag8 22. Ke3 h4 23. fxg4 Rxg4 24. Rxg4 Rxg4 25. h3 Rg8 26. Kf3 Rg7 27.
Ke2 Rg2 28. Rh1 Kd8 29. Kf3 Rg6 30. Rc1 Kd7 31. Rc3 Kc7 32. Bd3 Rg8 33. b3
Kd7 34. Bb5 Kc7 35. Rc2 Kd7 36. Rc1 Kc7 37. Ke3 Kd7 38. Ke2 Kc7 39. Rc2 Kd7
40. Rc3 Kc7 41. Ke3 Kd7 42. Rd3 Kc7 43. Rd1 Kd7 44. Kf3 Kc7 45. Ra1 Rg7 46.
Ra2 Rg8 47. Rb2 Rg7 48. Rd2 Rg8 49. Ke2 Rg2 50. Kf1 Rg8 51. Rd1 Kd7 52. Ke2
Kc7 53. Bd3 Kd7 54. Re1 Kc7 55. Kf3 Kd7 56. Rc1 Kc7 57. Be2 Kd7 58. Rd1 Kc7
59. Rf1 Kd7 60. Rb1 Kc7 61. Bf1 Kd7 62. Bb5 Kc7 63. Bd3 Kd7 64. Rd1 Kc7 65.
Ra1 Kd7 66. Bf1 Kc7 67. Rc1 Kd7 68. Ke3 Kc7 69. Rc3 Kd7 70. Rc2 Kc7 71. Bb5
Kd7 72. Ra2 Kc7 73. Rd2 Kd7 74. Rb2 Kc7 75. Bf1 Kd7 76. Bd3 Kc7 77. Rc2 Kd7
78. Be2 Kc7 79. Rc1 Kd7 80. Rd1 Kc7 81. Bf1 Kd7 82. Kf3 Kc7 83. Rc1 Kd7
{Game drawn by the 50 move rule}

Fritz 10 on 1/2 of a Woodcrest @ 3ghz, Fritz 10.ctg, 600mb HT.

Via this site: