Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tell me about computers learning....

Author: Dann Corbit

Date: 12:30:12 10/27/98

Go up one level in this thread


On October 27, 1998 at 09:09:46, Gunnar Andersson wrote:
[snip]
>I have been working on a computer Othello program for the last year (it's called
>Zebra and of decent strength, it has won all games against human players since
>January) and in this field the following scheme is used:
>  For each book position, evaluate the best move not having been played in
>  the game database using a midgame search. Then do a negamax over the entire
>  game tree where won games are scored as +INFTY, lost as -INFTY and drawn as 0.
>  In internal nodes the negamax is the best of the child nodes in the game tree
>  and the best deviation move.
>This approach is used by all the top Othello programs. For it to make sense,
>the evaluation function must be quite good, but I don't think that is a problem
>for today's chess programs.
This is very interesting.  I think that there is still a problem with applying
this to chess, but I would be interested to hear your thoughts.  The biggest
problem with chess search/evaluation functions (as I see it) is that the
information is almost *never* perfect.  There are so many combinations and
possibilites that you are always relying on an estimate.  And you can never
carry out the calculations to the end to find out either.
How would this technique work when the information is known to be imperfect
(sometimes completely wrong!)?



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.