Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Verified Null-Move Pruning, ICGA 25(3)

Author: Tony Werten

Date: 10:40:05 11/20/02

Go up one level in this thread


On November 20, 2002 at 13:22:30, Omid David Tabibi wrote:

>On November 20, 2002 at 12:57:24, Tony Werten wrote:
>
>>On November 20, 2002 at 12:31:43, Omid David Tabibi wrote:
>>
>>>On November 20, 2002 at 12:05:09, Tony Werten wrote:
>>>
>>>>On November 20, 2002 at 11:52:49, Gian-Carlo Pascutto wrote:
>>>>
>>>>>My very first thought after looking through this was:
>>>>>
>>>>>'You note that Heinz R=2/3 appears to be superior to
>>>>>R=2 and R=3, but you don't include it in the comparison.'
>>>>
>>>>From the data, I'd guess that both are about equal. I'll have a test after
>>>>diner.
>>>>
>>>
>>>The important difference is that verified null-move pruning on average yields
>>>greater tactical strength than standard R = 2. The smaller your quiescence
>>>search is, the better will be the performance of verified null-move pruning.
>>
>>Unfortunately, XiniX has a very big quiescence search. But the first
>>testpositions are not negative. Though the tree seems to be getting bigger,
>>moves are found earlier sometimes.
>>
>
>The size of the quiescence search has a direct impact on the efficiency of
>verified null-move pruning. If you have a huge quiescence search, then the size
>of the tree constructed by verified null-move pruning might exceed that of the
>standard version.

Yes, first tests do seem to indicate this. I'll toy around with it some more.

Tony

>
>Omid.
>
>>
>>Tony
>>
>>>
>>>Omid.
>>>
>>>
>>>>Tony
>>>>
>>>>>
>>>>>--
>>>>>GCP



This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.