Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Statistical Emulation of Chessplayers

Author: Bob Durrett

Date: 18:14:58 11/02/02

Go up one level in this thread


On November 02, 2002 at 20:59:31, Russell Reagan wrote:

>I'm sure you could do what you're wanting to do if you could get enough sample
>games for the different classes. Throw in a markov chain algorithm, maybe with a
>few improvements for selecting a move when no chain exists, and there you go. A
>markov chain algorithm can produce English text that is readable and makes sense
>(at least grammatically), or music, or whatever you give it to sample. I would
>guess chess moves would work also, although it probably wouldn't be very strong.
>This would probably work for mimicing lower level human players though,
>something which strong engines don't seem to do well.
>
>Russell

That's wonderful!  Please make one available ASAP! : )

Bob D.



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.