Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: what program has the best endgame playing ability?

Author: Omid David Tabibi

Date: 09:19:52 12/15/02

Go up one level in this thread


On December 15, 2002 at 11:59:18, Uri Blass wrote:

>On December 15, 2002 at 11:39:59, Mogens Larsen wrote:
>
>>On December 15, 2002 at 11:04:59, Omid David Tabibi wrote:
>>
>>>Not at all! A program that judges the above position at 0.00at its evaluation
>>>function, might opt for another _real_ winning variation, avoiding a draw; while
>>>a program without such knowledge will enter this blockage thinking that it has a
>>>certain victory in hand.
>>
>>I think you have to be careful with that kind of reasoning, because both
>>scenarios involve "might". Either might choose another variation and might enter
>>a blockade. The position a few moves before the one you posted would be more
>>interesting. Would this or that engine end up in a blocked position voluntarily.
>>There should be some paths to choose from. Looking at a score and trying to
>>reason what it might have done is no good.
>>
>>Regards,
>>Mogens
>
>I can add that the problem with static evaluation is that a program may be a
>root processor so the fact that it knows to evaluate fortress position at the
>root does not mean that it is going to know to evaluate it when the draw
>position is deep in the tree.
>

True, and that is why I incorporate my blockage detection scheme within the
evaluation function, so that it is applied to nodes in the search tree.

>Uri



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.