Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Dynamically Generated Partial Tablebases

Author: Jay Scott

Date: 13:26:25 02/28/01

Go up one level in this thread


On February 27, 2001 at 07:42:21, Steve Schooler wrote:

>I believe that this position can be routinely
>resolved by using the following algorithm to dynamically generate a PARTIAL
>tablebase:

Yes, that will work for some positions. It's a type of exhaustive search,
so it avoids the need for tons o' knowledge. Then if you store the results,
it *becomes* tons o' knowledge. :-)

I have a vague recollection that somebody has done this by manual setup
of tablebase software to solve specific positions of interest, but
I can't remember any details. :-/

The idea can be extended by not requiring all leaf positions to be definite
wins. Suppose you have a position where all the pawns are blocked, but there are
pieces free to move. Take the leaf positions of your universe to be the
positions which are mate *or* where something just got captured. Then do the
tablebase backup mambo, and when you're done you'll know whether it's possible,
from your initial position, for either side to force a capture.

This search may find a mate or prove that the position is a draw. It may only
show that one side can force a capture, which doesn't prove the result either
way--but the program can evaluate the positions after possible captures to try
for more information.

  Jay



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.