Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Proof number & conspiracy search

Author: Georg v. Zimmermann

Date: 09:06:48 08/29/02

Go up one level in this thread


On August 29, 2002 at 09:56:03, Andrzej Nagorko wrote:

>On August 29, 2002 at 09:23:22, Eli Liang wrote:
>
>>Is Sjeng the only chess program that has implemented proof number search?  Why
>>should that be?
>
>  Proof number search is important in games where there long, forced mating
>sequences are common. This is important only in some chess variants like
>crazyhouse and suicide chess and not in regular chess. Sjeng plays both
>variants, that is why it uses pn-search.

You are right about most here. But PN search does not work well in crazyhouse
chess, because like in normal chess, most positions end with an evaluation. And
PN only works well for yes/no decissions (e.g. "can I mate here?"). If the
answer to that is :"no" , you still have to find a good move ...

Georg



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.