Author: Uri Blass
Date: 16:15:08 08/28/03
Go up one level in this thread
On August 28, 2003 at 17:35:14, Russell Reagan wrote: >I was wondering how often paradoxes happen in computer chess games. For >instance, does a program without a transposition table ever beat a program with >a transposition table consistently (assuming they are otherwise similar)? Or a >program without any forward pruning beats one that uses null move? > >I'll give an example to show what I mean by "otherwise similar": > >Program A: >basic alpha-beta search >move ordering >qsearch >evaluation function > >Program B: >basic alpha-beta search >move ordering >qsearch >evaluation function >transposition table > >The only main difference is the transposition table, even though the details of >move ordering, qsearch, and evaluation might be different. It doesn't seem like >program A should ever beat program B consistently. It is dependent on the evaluation. I believe that a good evaluation may beat only piece square table program at least in case that the only difference is the tranposition tables. > >Are there any examples of such a paradox occuring? For instance, maybe a program >lacked some major component that another had (transposition table, forward >pruning, etc.) but it was still stronger because of, say, superior positional >evaluation. A lot of things are dependent on evaluation so I see no way to say the search is the same and the only difference is evaluation. The optimal search decisions for piece square table program may be different than the optimal decisions for a better program. 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.