Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Infinite Analysis problems, different results, different methods

Author: Kurt Utzinger

Date: 00:09:03 09/12/04

Go up one level in this thread


On September 11, 2004 at 16:13:38, Derek Paquette wrote:

>This happens to occur with junior 8.0.0.2 and Shredder 8, and it REALLY REALY
>makes me angry.
>
>Say I have a set position, doesn't matter what position.
>
>I load shredder 8, set 64hash lets say, or 256...doesn't matter.
>
>I run infinite analysis, it gives me the best move, nevermind the series of
>moves, i just mean the best next move.  Same goes for Junior 8, NOW... here is
>the kicker...
>
>I load the same position, go to game, go to game levels, and put it on a long
>game, or short game, or any time control, click on MOVE NOW and of course it
>starts calculating the best position, until it decides to make a move depending
>on what sort of time control i have specified.
>
>To my suprise, rather my horror, the moves that the program calculates as it
>thinks are DIFFERENT than those in the infinite analysis,
>
>so i can't even use that? or am i not seeing this properly????


      Hi Derek
      This all has been discussed earlier I think. To produce
      identical moves is not so easy and some things must be
      considered for all tests:
      1) ponder=off
      2) same amount of hash size
      3) same playing level (thinking process may otherwise differ)
      4) always clear hash tables before starting analysis
      5) no other or at least identical open tasks on the PC
      6) rebooting PC before each teast
      Even if you follow the above instructions it will of
      course rarely be possible to reproduce all moves played
      in a game just because conditions used can't be repeated.
      Kurt [http://www.utzingerk.com]



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.