Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: book learning

Author: Robert Hyatt

Date: 21:11:51 02/17/01

Go up one level in this thread


On February 17, 2001 at 14:14:20, Ernst A. Heinz wrote:

>Hi Bob,
>
>>On February 16, 2001 at 15:16:57, Ernst A. Heinz wrote:
>>
>>>Hi Martin,
>>>
>>>>i would like to implement a book learning feature for my checkers program.
>>>>unfortunately, i have no idea how :-)
>>>>so: is there a 'most relevant webpage' which describes how to do this or could
>>>>someone give a short description on how to do it?
>>>
>>>Michael Buro implemented book learning for his champion
>>>Othello program "Logistello" and described the schemes
>>>in "Toward Opening Book Learning", ICCA Journal 22(2),
>>>Sept. 1999. A preprint of the article is available online
>>>at http://www.neci.nj.nec.com/homepages/mic/ps/book.ps.gz.
>>>
>>>=Ernst=
>>
>>That isn't so hot for chess.  The way the learning works is based on a game
>>tree that is pretty narrow, which chess is not while othello is.  Frey also
>>did book learning with othello (I think it was in the old chess skill book).
>>But again, I don't think that approach is very applicable to chess either.
>
>I agree with you regarding the possible applicability of
>Michael Buro's approach for chess -- but Martin Fierz is
>working on *checkers* (not on chess) which features much
>narrower search trees than chess in general. Hence, I
>still think the approach is worthwhile to look at for
>Martin's particular purpose.
>
>=Ernst=


For checkers I would probably agree... although jonathan's approach of
searching from the beginning to the endgame tables is the way the game will
be solved before long.



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.