Author: Robert Hyatt
Date: 10:44:56 01/18/04
Go up one level in this thread
On January 18, 2004 at 13:25:48, K. Burcham wrote:
>On January 18, 2004 at 11:20:44, Robert Hyatt wrote:
>
>>On January 18, 2004 at 08:37:19, Harald Faber wrote:
>>
>>>In my few idle time I am testing Shredder 8 on tournament level versus Junior 8
>>>(because I will operate Junior 8 in the computerchess tournament in Bernburg in
>>>March this year). While playing I became witness of this interesting game:
>>>
>>
>>I have seen engines lose from +5 and draw from +8 on ICC. This includes
>>both commercial programs as well as mine. Draws are particularly easy to fall
>>into as with queens on the board, it might be 30 moves before the first
>>repetition occurs and programs simply can't see that every time...
>>
>>Losing from +5 can be caused by several things, from a king attack to mis-
>>understanding a pawn ending.
>>
>>That's why I mentioned my take on programs resigning in important games,
>>there is no benefit to resigning against a computer, and there is a potential
>>problem in that you might _still_ win. See Jonny vs Shredder for example,
>>although there are plenty of other examples floating around in past chess
>>games.
>
>Robert when a program has a three move repetition problem in the code, is this
>different than when a program that cannot see that +4 should be -2 because
>position is too deep to see true score?
>
>kburcham
>
No, you are taking my comment too literally. The repetition bug is one
issue. Evaluation bugs also lose games. Search bugs lose games. Poor
evaluation loses games. Poor extensions lose games. There are thousands
of ways to turn +5 into -Mate...
>>
>>
>>>[Event "Junior8 Vorbereitung, 120'/40+0'/0+60'"]
>>>[Site "CELERON1400"]
>>>[Date "2004.01.18"]
>>>[Round "2"]
>>>[White "Shredder 8"]
>>>[Black "Junior 8"]
>>>[Result "0-1"]
>>>[ECO "B66"]
>>>[Annotator "-0.12;-0.18"]
>>>[PlyCount "96"]
>>>[TimeControl "40/7200:0/0:3600"]
>>>
>>>{W=18.5 ply; 343kN/s B=18.1 ply; 1.137kN/s; 6.709 TBAs} 1. e4 {0} c5 {0} 2.
>>>Nf3 {0} d6 {0} 3. d4 {0} cxd4 {0} 4. Nxd4 {0} Nf6 {0} 5. Nc3 {0} Nc6 {0} 6. Bg5
>>>{0} e6 {0} 7. Qd2 {0} a6 {0} 8. O-O-O {0} h6 {0} 9. Be3 {0} Bd7 {0} 10. f4 {0}
>>>Qc7 {0} 11. Bd3 {0} b5 {0} 12. Kb1 {0} Na5 {0} 13. Qe1 {0} Nc4 {0} 14. Bc1 {0}
>>>b4 {0} 15. Nce2 {0} a5 {0} 16. h3 {Beide letzter Buchzug 0} a4 {-0.18/18 448}
>>>17. g4 {-0.12/17 0} a3 {0.00/17 367} 18. b3 {-0.12/19 0} Nb2 {0.03/17 186} 19.
>>>Bxb2 {0.37/20 806} axb2 {-0.02/18 2} 20. Qxb4 {0.48/21 1474} d5 {
>>>(Le7) 0.28/19 2} 21. Qc3 {0.60/19 317} Qb6 {(Da7) 0.21/18 352} 22. e5 {
>>>0.79/19 413} Ne4 {(Lb4) 0.25/18 2} 23. Bxe4 {0.72/19 487} dxe4 {0.55/17 213}
>>>24. f5 {0.92/19 470} Ba3 {(Da6) 0.55/17 171} 25. fxe6 {1.53/15 47} fxe6 {
>>>(Lxe6) 0.85/17 354} 26. Rhf1 {1.71/17 238} Rf8 {0.96/17 319} 27. Rxf8+ {
>>>(De3) 1.89/17 0} Kxf8 {(Lxf8) 0.94/16 268} 28. Qe3 {2.27/18 235} Kg8 {
>>>1.12/18 731} 29. Qxe4 {2.29/20 88} Rd8 {(Da5) 1.42/18 660} 30. g5 {
>>>(Sc3) 2.98/19 287} hxg5 {1.59/17 503} 31. Qg6 {(Sc3) 2.74/19 21} Bb5 {
>>>1.15/17 423} 32. c4 {(Dxg5) 2.32/19 13} Bc6 {0.79/17 277} 33. Qxe6+ {2.08/18 71
>>>} Kh8 {0.63/19 464} 34. Qg6 {(Df5) 1.45/19 326} Qc5 {0.23/15 60} 35. Qf5 {
>>>0.83/18 516} Bb7 {0.00/20 2} 36. h4 {0.58/18 199} Rf8 {(gxh4) 0.00/17 133} 37.
>>>Qxg5 {1.89/16 80} Be4+ {-0.04/17 92} 38. Nc2 {-0.42/19 558} Qf2 {-0.07/18 2}
>>>39. Ned4 {-0.42/19 313} Kg8 {-0.37/19 2} 40. Qg1 {-0.18/16 72} Qxh4 {
>>>-0.35/18 365} 41. e6 {-0.19/18 0} Rf2 {-0.94/18 194} 42. Qe1 {-0.34/18 0} Qf4 {
>>>(Lc5) -0.81/17 196} 43. b4 {-1.11/15 157} Kf8 {-2.26/17 85} 44. b5 {-2.60/13 51
>>>} Ke7 {-3.08/16 78} 45. Nc6+ {(c5) -4.27/14 153} Kxe6 {-4.58/16 244} 46. N6d4+
>>>{(c5) -6.86/16 846} Kf7 {-4.62/16 35} 47. c5 {-7.30/15 85} Bxc5 {-4.16/15 1}
>>>48. Kxb2 {-7.05/15 156} Bxd4+ {-4.05/16 2} 0-1
>>>
>>>Shredder 8 on a P4-2.53GHz and Junior 8 on my Celeron@1539MHz, both 256MB hash.
>>>A remarkable result, considerung that Shredder showed almost +3 in move 30 with
>>>search depth of 19.
>>>
>>>Position before 30.g5 +2.98/19:
>>>[D]3r2k1/3b2p1/1q2p2p/4P3/3NQ1P1/bP5P/PpP1N3/1K1R4 w - - 0 30
>>>
>>>Current standing is surprisingly 4-2 in favour of Junior.
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.