Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question on EGTB generation theory

Author: Heiner Marxen

Date: 14:51:34 07/03/02

Go up one level in this thread


On July 03, 2002 at 13:32:34, Alvaro Jose Povoa Cardoso wrote:

>I red some papers on EGTB generation.
>They mention that once a position's value is final (it's number of unknown
>children is zero) it (the value) can be propagated to the position's parents
>(plus 1).
>And if some or all of it's parent's value are also final (the number of their
>children with an unknown value is zero) we recursively propagate it's value the
>their parents and so on.
>My question is if this step is absolutely necessary for the algorithm or on the
>other hand is just a speed up.

As long as you guarantee to consider all (not yet final) parents sometime
later it is not necessary to consider them immediately and now.
I'm not even sure whether this would be a speed-up.

>Thanks in advance,
>Alvaro Cardoso

Cheers,
Heiner



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.