Author: Art Basham
Date: 08:27:40 06/21/02
Go up one level in this thread
Hi.. For the most part, I think they are the best kind of test positions available to us! Example, this mate in 10... [D]8/8/4k3/2K1N1B1/5Pp1/8/2B2P1p/6n1 w - - After all, checkmate is what chess is all about, is it not? Any program that cannot find mate in a composed problem really has serious "problems"...I think. Just ask John Merlino....:-) What makes ChessMaster so great is this ability to find mates in such a short time...even in a composed problem. regards! Art ================================================ On June 21, 2002 at 08:42:46, Robert Henry Durrett wrote: > >Just wondering. > >Whether or not a test position is useful seems to depend on what it would/could >show. Different algorithms each have their own strengths and weaknesses and, it >would seem, test positions useful for checking out the performance of particular >algorithms might not be useful for checking out other algorithms. At least, >this is my current perception of the situation. > >Composed problems are usually associated with endgames, although not >necessarily. > >The performance of various endgame subroutines [which implement specific >algorithms], in particular, might be subject to test using specialized test >positions. > >Would composed endgame problems be well suited for this purpose? > > >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.