Author: pavel
Date: 18:14:10 05/22/02
Go up one level in this thread
On May 22, 2002 at 17:42:38, Jon Dart wrote: >On May 22, 2002 at 15:56:06, pavel wrote: > >>[D]1r4k1/pr4pp/4pp2/3pP1PP/3P1P2/2N5/1PR5/2K5 w - - 0 1 >> >> >>having a relaxed game with arasan, after going through a drill with >>chessmaster8000, and learning from Josh how to use rook and knight more >>efficiently. >> >>Arasan was forced to use 5second per move on its own GUI. On the above position >>Arasan resigns, any perticular reason? > >I have recently been looking at a similar situation that occurred in a computer >-computer match I was running at home. As here, it resigned when in fact ti >wasn't losing. When I looked at the arasan.log file, the score at the point it >resigned was -Mate1 (mate in 1). > >If I reloaded the game into the Arasan GUI, moved to the end, and made Arasan >search there, the problem did not recur (I also saw this with the game you >posted). > >There is some code that causes Arasan to resign if it determines that it is >subject to a relatively shallow forced mate. I suspect this can be erroneously >triggered, possibly due to some bad handling of mate scores in the hash table. >But I have so far not been able to reproduce this problem, which makes it hard >to fix. > >In my experience this occurs quite rarely. If you do find it is a problem, you >can disable resignation. > >If you have any further information that could help track this problem down, I'd >appreciate hearing about it. At this moment I dont have any idea why this could have happened, or any suggestion as to how to fix it, but I will make sure I will let you know as soon as I find out. regards pavs
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.