Author: Howard Exner
Date: 18:37:30 04/26/98
Go up one level in this thread
>We just completed rounds 1-4 of this event. In rounds 1/2 Crafty played
>GM Roman Dzhindi. As white, Crafty won, as black, it drew. In rounds
three
>and four, against GM Larry Christiansen, it again won with white, and drew
>with black, although it had good winning chances in the black game...
>
>It is currently tied with Yasser with 3.0/4.0 each...
Beating Yasser Seirewan even at this 10 10 time control will be no small
accomplishment. Even thogh these are fun events (like Aegon) I do think
chess GM's are a breed of player that despise losing in any event, fun
or serious. I've never read one accountant of a GM losing in a fun event
saying " oh well, there was no money at stake so I wasn't trying my
best".
On the contrary I recall how bothered Yasser was last year at Aegon
(based
on his comments) when he drew a program (Cilk Chess I believe).
By the way, what do those who attended Aegon think about how hard the
players try to win (Ed,Thorsten?)
>
>before you start on the "I thought you said we had no computer GM
>players" remember that this is 10 10 games... 10 minutes on the clock,
>10 seconds added after each move... pretty fast... although fast enough
>to let crafty hit 10 plies or more most of the time...
But still a very impressive performance by Crafty against very
computer-savy
opponents. Here's a question for those familiar with computer ratings on
the
servers. What are the average time controls for blitz ratings vs
standard?
Is it
something like 5 min per game vs 15 min per game? What are the ratings
of
the top 10 computers on the various servers, both blitz and standard?
I'm interested in seeing of course how these two time controls compare
(how
many rating points on the average do computers lose going from blitz to
standard using the servers as a database). Someone may already have this
data.
Now if only we had a large pool of 40/2 then we could know better how
strong these programs are becoming at that time control.
Bob and I know each others present
stance on this (at least I know your's, Bob). I think they are playing
around
2500 ELO and Bob says "not even close to GM strength yet" (is that an
accurate summary of your position?). This is in 40/2 - regular chess.
I left this one game in with an observation on move 45
>[Event "mmeics rated standard game"]
>[Site "mmeics, Boston, MA USA"]
>[Date "1998.04.25"]
>[Time "16:48:16"]
>[Round "4"]
>[White "Christiansen"]
>[Black "crafty"]
>[WhiteElo "2581"]
>[BlackElo "2413"]
>[TimeControl "600+10"]
>[Mode "ICS"]
>[Result "1/2-1/2"]
>
>1. Nf3 d5 2. a3 Nf6 3. b4 a5 4. b5 Bf5 5. Bb2 e6 6. e3 Bd6 7. d4 c6
>8. Nbd2 O-O 9. c4 dxc4 10. Bxc4 Qb6 11. a4 Rc8 12. O-O Nbd7 13. Rc1 cxb5
>14. Bxb5 Rxc1 15. Qxc1 Qd8 16. Qa1 Rc8 17. Rc1 Bb4 18. h3 Nb6 19. Nc4
>Nxc4 20. Rxc4 Nd5 21. Qc1 Rxc4 22. Qxc4 Nc7 23. Ne5 Be4 24. Qe2 Qg5 25. f3
>Nxb5 26. axb5 Bd5 27. e4 Bb3 28. Qd3 a4 29. b6 Qd2 30. Qxd2 Bxd2 31. Nd3
>Ba5 32. Nc5 Bxb6 33. Nxb3 axb3 34. Kf1 Ba5 35. Ke2 Bb4 36. Kd3 b5 37. g4
>g5 38. e5 h6 39. d5 exd5 40. Kd4 Bd2 41. Kxd5 Kf8 42. Kc5 b4 43. Kc4 Ke7
>44. Kxb3 Ke6 45. Bd4 Be1
Right here might have been the time for Bc3, rather than one move later
as
Crafty played. Given a bit more time I'll wager Crafty would have played
Bc3
on this move. The position of the white king makes a difference.
46. Kc4 Bc3 47. Bf2 Kxe5 48. Bg3+ Ke6 49. f4 f5
>50. fxg5 hxg5 51. gxf5+ Kxf5 52. Bd6 Be1 53. Bc7 Kg6 54. Bd8 Bc3 55. Be7
>Kh6 56. Kd3 Kh5 57. Kc4 Kh4 58. Kd3 Be1 59. Kc4 Bd2 60. Kd3 Bc3 61. Kc4
>Be1 62. Kd3 b3 63. Bf6 Ba5 64. Be7 b2 65. Kc2 Bc3 66. Bd8 Be5 67. Be7 Bh8
>68. Bd8 Bd4 69. Be7 Kh5 70. Bd8 Kg6 71. Ba5 Kf5 72. Bd2 Be5 73. Bb4 Bg7
>74. Bd2 Bd4 75. Kb1 Kf6 76. Kc2 Be5 77. Ba5 Kg6 78. Bd8 Bf6 79. Bb6 Kh5
>80. Bf2 Bg7 81. Kb1 Bc3 82. Kc2 Be5 83. Kb1 Bh8 84. Kc2 Bc3 85. Kb1 Kh6
>86. Kc2 Kg6 87. Bb6 Kf6 88. Be3 Kf5 89. Kb1 Be5 90. Kc2
>{Game drawn by mutual agreement} 1/2-1/2
The win against Roman depicts a nice example of a strength computers
have.
Namely there tenacious ability to slowly convert small advantages. In
this
game
it is not easy to convert the extra doubled a pawns. Roman was hoping
Crafty
would swap rooks (as programs of old would have blindly done) but little
by
little programs seem to be incorporating this kind of knowledge (ie:
avoiding
opposite colored bishop endings when up in pawns)
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.