Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Using GNU Chess to generate sample data for Behaviour Cloning

Author: Luis Smith

Date: 00:16:19 11/10/03

Go up one level in this thread


On November 10, 2003 at 02:51:59, Alex Holland wrote:

>I'm doing a Third Year Computer Science degree project on Behaviour Cloning for
>Chess - using Inductive Logic Programming (Aleph in particular) to copy the
>chess playing behaviour of an existing person or machine.
>
>One of the sources I'm considering cloning is GNU Chess, as I can easily have a
>look at it and see how it works, and it's fairly well documented. To this end,
>I'd like to set it up to play against itself for a few days, keeping all the
>game moves saved in a big text file. I can then use this file as positive
>examples for the ILP algorithms.
>
>However, I can't find any way to do this. I know I can save games into PGN
>format, but I can't find any way to make it play against itself, short of typing
>"go" over and over again

I'm not much of a programmer, but I would reccomend TSCP "Tom Kerrigan's Simple
Chess Program".  All you have to do is compile it, then recompile it with a
different name.  The source code is clean and a lot of programmers use it as
their basis for creating their own.

http://home.comcast.net/~tckerrigan/

>
>Ideally what I want is for GNU chess to play against itself until endgame, then
>save a PGN, then repeat. Is there any facility for this, or am I going to have
>to exert my shaky C++ skills on the sourcecode?
>
>For reference, I'm running GNU Chess 5.06 on Gentoo GNU/Linux, and I'll be
>starting by playing KRK games.
>
>Secondly, does anyone know of any strictly deterministic chess programs, i.e.
>ones that make the same moves every time in a given situation? I realise such a
>program would be fairly useless to play against, but it would be pedagogically
>interesting to have a noise-free source to attempt to clone.
>
>Lastly, anyone generally got any documents they could point me at that they
>think might be relevant? I've found Jay Scott's machine learning page, which is
>a big help.
>
>Thanks in advance.
>
>Alex



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.