Author: Uri Blass
Date: 08:39:31 04/08/05
Go up one level in this thread
On April 08, 2005 at 11:12:42, Günther Simon wrote:
>On April 08, 2005 at 10:57:02, Uri Blass wrote:
>
>>Movei in AEGT does not use tablebases except in the root.
>>
>>I find that this helped it to save a draw against Little Goliath and here is the
>>game
>>
>>[Event "AEGT 4"]
>>[Site "Queen Class H1"]
>>[Date "2005.04.03"]
>>[Round "1.5"]
>>[White "Little Goliath Revival"]
>>[Black "Movei CCT7"]
>>[Result "1/2-1/2"]
>>[ECO "A87"]
>>[PlyCount "117"]
>>[EventDate "2005.04.01"]
>>[EventType "tourn"]
>>[TimeControl "40/1080:0"]
>...
>>51. Bxf7
>>{[eval 857,16] [emt 0:00:11]} Kxf7 {[eval 189,15] [emt 0:00:18]} 52. f4 {
>>(Kg3) [eval 32740,17] [emt 0:00:08]} gxf4 {(Ke6) [eval 193,18] [emt 0:00:22]}
>>53. Kf3 {[eval 32750,4] [emt 0:00:00]} Kg7 {(Kf6) [eval 285,18] [emt 0:00:36]}
>>54. e4 {(Kxf4) [eval 32754,4] [emt 0:00:00]} fxe3 {[eval 0,22] [emt 0:00:22]}
>>55. Kxe3 {[eval 0,1] [emt 0:00:01]} Kf6 {[eval 0,1] [emt 0:00:00]} 56. Kf4 {
>>[eval 0,1] [emt 0:00:00]} Kg6 {(Ke6) [eval 0,1] [emt 0:00:00]} 57. Ke3 {
>>(Kg3) [eval 0,1] [emt 0:00:00]} Kg5 {(h5) [eval 0,1] [emt 0:00:00]} 58. Kf3 {
>>[eval 0,1] [emt 0:00:00]} h5 {[eval 0,1] [emt 0:00:00]} 59. gxh5 {
>>[eval 0,1] [emt 0:00:00]} 1/2-1/2
>>
>>Movei with tablebases not at the root cannot find the move 52...gxf4 and prefers
>>to resign that does not give practical chances unlike gxf4
>>
>>Uri
>
>This looks very much like a huge e.p. bug in LittleGoliath,
>especially, as it already shows mate score the moves before
>and all moves except 54...fxe3 are still mate in some moves.
>
>54.e4??
>[D]8/6k1/7p/8/5pP1/5K2/4P3/8 w - - 0 54
I agree
The bug happens near tablebase positions and I guess that using tablebases not
at the root could prevent Movei to get the relevant position when Little Goliath
blundered.
Uri
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.