Author: Robert Hyatt
Date: 12:12:53 02/27/05
Go up one level in this thread
On February 27, 2005 at 12:47:00, Peter Skinner wrote:
>On February 27, 2005 at 11:51:29, Uri Blass wrote:
>
>>On February 27, 2005 at 11:19:35, Peter Skinner wrote:
>>
>>>On February 27, 2005 at 11:03:42, Christos Gitsis wrote:
>>>
>>>>So you think Crafty 19.17 is stronger under Linux? Is Crafty stronger under
>>>>Linux than it is under WinBoard in general?
>>>
>>>No.
>>>
>>>The Windows version is about 5% faster due to compilers.
>>>
>>>Other than that, they are the _exact_ same source code, under virtually the
>>>_same_ interface.
>>>
>>>I find there no difference what so ever.
>>>
>>>Peter
>>
>>Maybe the problem was that bob does not use Crafty under xboard on ICC.
>
>From Crafty's "variables" on ICC:
>interface="xboard 4.2.6 + crafty"
>
>So yes he _does_ use xboard on ICC.
>
>>I am not sure what was exactly the problem but I remember that there was some
>>analysis that Bob hyatt could not reproduce with crafty19.17 under different
>>conditions.
>>
>>Uri
>
>I find very often that Crafty plays odd moves that I can never reproduce when
>later playing a game.
>
>This happens in _all_ versions since 19.15. 19.15 doesn't seem to have this
>problem at all.
>
>Something that was changed or introduced since that version has messed something
>up.
>
>Just the other day while watching Crafty on ICC, I saw it lose a game where it
>was in a completely won position. I thought I had saved it for later analysis,
>but for some odd reason games that I have been saving with Blitzen are saved
>incomplete. I finally had to re-install the program to get it to work right.
>
>I noticed Bob is using 19.20 on ICC now, but it's rating has dropped big time.
>It is sitting at 2653. That is pretty low for it, when just 2 weeks ago it was
>rated 2841. See here:
>
>[Event "ICS rated blitz match"]
>[Site "chessclub.com"]
>[Date "2005.02.15"]
>[Round "-"]
>[White "Somnus"]
>[Black "crafty"]
>[Result "1-0"]
>[WhiteElo "2780"]
>[BlackElo "2841"]
>[TimeControl "300+3"]
>
>1. d4 d5 2. c4 e6 3. Nf3 Nf6 4. Nc3 c6 5. e3 Nbd7 6. Qc2 Bd6 7. Be2 O-O 8.
>O-O dxc4 9. Bxc4 b5 10. Be2 Bb7 11. Rd1 Qc7 12. e4 e5 13. g3 Rfe8 14. a3 a5
>15. Bg5 exd4 16. Nxd4 b4 17. Nf5 Bc5 18. Bf4 Qc8 19. Na4 Bf8 20. axb4 axb4
>21. f3 Ba6 22. Bxa6 Rxa6 23. b3 Ne5 24. Qe2 Nfd7 25. Be3 Ra8 26. f4 Ng6 27.
>e5 Ra6 28. g4 Ra8 29. Ra2 Qc7 30. Rad2 Nb8 31. Qc4 Qa5 32. h4 Qc7 33. h5
>Nh8 34. Bb6 Qb7 35. Rd8 Na6 36. R1d7 Qxb6+ 37. Nxb6 Raxd8 38. Rxd8 Rxd8 39.
>Qxa6 g6 40. Nd6 gxh5 41. gxh5 Bxd6 42. exd6 Rxd6 43. Qc8+ Kg7 44. Nd7 Rd1+
>45. Kf2 Rxd7 46. Qxd7 c5 47. Qe7 Ng6 48. hxg6 hxg6 49. Qxc5 g5 50. fxg5 Kg6
>51. Qxb4 Kf5 52. Qe7 Kf4 53. Qxf7+ Ke4 54. g6 Kd4 55. g7 Kc3 56. g8=Q Kb4
>57. Qf6
>{crafty resigns} 1-0
>
>Peter
If you run the _exact_ same version for 6 months, using the exact same hardware,
you will find 300 point rating swings. It is just part of the normal standard
deviation of rating played at fairly fast time controls...
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.