Author: Rick Bischoff
Date: 15:48:34 07/22/03
Go up one level in this thread
On July 22, 2003 at 18:34:11, Dieter Buerssner wrote:
>My guesses: If you use HTs - an error there. If not, and if you use null move,
>perhaps an error there. For example, doing a null move when in check. Or
>allowing 2 null moves in a row (which might be good, when prepared to it). The 2
>null moves in a row might easily happen by HT cutoffs, when you store fail highs
>after nullmove. You might probe the table and get a cutoff (which was caused by
>nullmove when you stored the score), but in this situation null move should not
>be allowed.
>
>Probably one can imagine many more causes. When you don't give more details,
>much wll be guess work.
>
>Good luck,
>Dieter
Hi,
I use hash tables, yes. No null moves. Nothing fancy in the alpha/beta search
except for a in-check extension. The qsearch searches all captures and if one
of those results in a check, calls alpha/beta with a depth of 1. Here is a
sample game where Ziggy2 blundered as black:
[Event "ICC 5 0"]
[Site "Internet Chess Club"]
[Date "2003.07.22"]
[Round "-"]
[White "*"]
[Black "Ziggy2"]
[Result "1-0"]
[ICCResult "Black checkmated"]
[WhiteElo "1744"]
[BlackElo "1710"]
[Opening "English: four knights, Capablanca variation"]
[ECO "A28"]
[NIC "EO.07"]
[Time "18:21:53"]
[TimeControl "300+0"]
1. Nf3 Nc6 2. c4 e5 3. Nc3 Nf6 4. d3 Bb4 5. Bd2 d5 6. cxd5 Bxc3 7. Bxc3 Qxd5
8. g3 e4 9. dxe4 Qxe4 10. Bg2 Bg4 11. O-O Rd8 12. Qb3 O-O 13. Bxf6 gxf6 14.
Qxb7 Rb8 15. Qa6 Rb6 16. Qa3 Qxe2 17. Rac1 Be6 18. b3 Ra6 19. Qc5 Qxa2 20.
Rfe1 Ra5 21. Qxc6 Qxb3 22. Nd4 Qb6 23. Qxb6 cxb6 24. Nxe6 fxe6 25. Rxe6 Re5
26. Rxe5 fxe5 27. Rc7 Rd8 28. Rxa7 Rd1+ 29. Bf1 Rd4 30. Rb7 Rd6 31. Kg2 Kf8
32. h4 h6 33. g4 e4 34. Kg3 Rd4 35. Bg2 Rb4 36. g5 h5 37. g6 Kg8 38. Re7 e3
39. fxe3 Rg4+ 40. Kh3 Rxg6 41. Bf3 Rd6 42. e4 Rd3 43. Kg3 Ra3 44. e5 Rd3 45.
Kf4 Rd4+ 46. Be4 Rd2 47. Rb7 Rb2 48. e6 Rf2+ 49. Ke5 b5 50. e7 Kf7 51. Bg6+
Kg7 52. e8=Q+ Rf7 53. Qxf7+ Kh6 54. Qh7# {Black checkmated} 1-0
On move 20. ... Ra5 hangs the knight. The analysis when the blunder happened:
10765078 <first : *** timeleft: 20200 basetime=30000 tpm = 500
10765078 <first : depth: 1 e=180 e6b3 nodes: 10
10765078 <first : depth: 2 e=130 a2b3 f3d4 nodes: 1464
10765229 <first : depth: 3 e=112 a2b3 f3g5 c6e5 nodes: 6082
10766156 <first : depth: 4 e=100 a6a5 c5e3 e6d5 f3d4 nodes: 29330
10770075 <first : *** making move a6a5 c5e3 e6d5 f3d4
10770077 <first : move a6a5
And then when I use the same engine to analyze the position after the game was
over:
depth: 1 e=180 e6b3 nodes: 10 checktime_called: 0
depth: 2 e=130 e6b3 f3d4 nodes: 1546
depth: 3 e=112 e6b3 f3g5 c6e5 nodes: 5744
depth: 4 e=112 e6b3 f3g5 c6e5 g5h7 nodes: 14911
depth: 5 e=112 e6b3 g2f1 a6b6 c5f8 g8f8 nodes: 62826
I am just scratching my head at this one.
Rick
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.