Author: James T. Walker
Date: 05:51:54 03/08/99
Go up one level in this thread
On March 08, 1999 at 04:05:20, Matt wrote: >On March 07, 1999 at 23:18:42, Jeremiah Penery wrote: > >> >>I think, however, that this isn't what Uri was talking about. I believe he was >>referring to the cases where a program says "Mate in 8" (for example), then >>after the opponent's move it announces "Mate in 10", and the mate scores jump >>around for a few moves until finally it gets the correct number. I've seen it >>happen with most programs, and it bugs me too. :) >> >>Jeremiah > >If all the programs do this, I would assume it is something they all have in >common. Would it be possible that alpha beta search finds mate in 8, because it >has not searched all the possible moves, then it later finds that one side can >delay mate for a few moves? Hello Matt, The problem is with programs calling mate in less number than actually exist. No problem calling mate in 10 when mate in 8 exist. Most all programs do that because when extending search they find the longer mate quicker due to checks. But if a shorter mate exist without checks inbetween it's ok. To my way of thinking, if it calls mate in 8 but the absolute shortest mate is mate in 9 then it is a bug and should be fixed. Jim Walker
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.