Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: more examples for search-based stupidity

Author: Hristo

Date: 20:05:00 06/13/98

Go up one level in this thread


I understand, perfectly, what you are talking about!

>>Can you name one program(or computer) that can predict(calculate) every
>>possible move from the very beginig of the game unitl the very last move
>
>yes given the time, most programs can, if they don't have a ply cut off
>point that is to low. :)

I've tried Rebel, Fritz5 and crafty and all of them have very similar
problems.
I've done test where even after 48 hours of "thinking" the programs were
unable to find the best possible move. Since none of the programs
performs a full search on the possible postitions there is a cut-off
reason to not evaluate a branch ... cut-off a branch here and there and
your tree doesn't give you enough information about the position! If one
evaluates ALL possible moves and builds a full tree with no cut-offs
then it will work. Just how long shall we wait until this happens??

About your comparison of Fritz and CST ... what is your point? One of
them works different than the other!? Very well, thats great! Somebody
has got enough "balls" to not follow the "gray creek" no matter how
appealing it might seem.

I've never seen CST so I don't know how well it plays!


Hristo












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.