Author: Tord Romstad
Date: 08:08:38 04/26/04
Go up one level in this thread
On April 26, 2004 at 07:20:33, Daniel Shawul wrote: >Hi Tord Hi Daniel! > >very interseting exeriment indeed! >I am doing A currently and sticking to it just because the effectiveness is >93-94% usually. >About B , I think the criteria should be > eval - our hanging piece - their hanging piece > beta > do null move. > If we do null move, we lose our hanging piece AND ALSO we chose not to capture >the the opponent piece so we lost that too. I think it should be minus both >cases. Minus in both cases might be worth a try, but it seems illogical to me. The whole point of variant A is to make an extremely crude estimate of the result of the null move search before starting it. You would expect the result of the search to be somewhere around eval - our hanging piece. Variant B is a (very poor) attempt to make a more exact guess. The assumption is that the null move search will result in the line null move - opponent captures our piece - we capture opponent's piece The expected score after this line is eval - our piece + their piece. There are, of course, a million ways this primitive way of estimating the result of the null move search could go wrong, for instance the very common case when the opponent's capture saves his hanging piece. I believe more in variant A. Tord
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.