Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Move ordering (and other ways to improve a program)

Author: Nicolas GUIBERT

Date: 15:22:04 06/23/99

Go up one level in this thread



>That was behind Flits in Paris recently?

Yes that was Buggy.

>
>My draughts programs name: Napoleon
>programming: Vincent Diepeveen
>knowledge and book: Marcel Monteba
>
>>>At this time, I am looking for articles about about move ordering.
>
>it's indeed tough in draughts!
>
>I usual search like 2-4 ply deeper in the middlegame/opening
>than any opponent, and in the endgame i search like 20 ply deeper...
>
>Draughts is an ideal testground to measure how well one knows things
>from algorithms, also what's happening in draughts might be a
>cool way of seeing how things might get for chess, as after like a
>night of searching in middlegame/opening Napoleon gets up to 30
>full width search depth!
>

I heard about your program being faster than the others.
But I am still wondering how you can achieve such results.

Even with perfect move ordering and "incredibly lucky" perfect choice of the
smallest parts of the tree, I can't see this possible without cutting a lot the
trees in the variants where 1, 2 or 3 pieces are already lost !

Then, what do you mean by full width search depth ?





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.