Author: David B Weller
Date: 06:38:13 10/23/04
Go up one level in this thread
I think your right. Since, IF you are actually 'do'-ing the move in question, then the ONLY thing you save by detirmining that the score will probably fail-high in quies() [and therefore prune it], is basically the call to quies()! Because you basically had to do your whole eval to detirmine if you should prune. [ie., material[side]-material[xside]+FUTILE_MARGIN] So in the case of a very simple eval[esp. material+pst], I think you are correct. The way I understand futility prunning [and please, someone correct me if i am wrong] is that you are doing a kind of pre-quies()-lazy-eval which asserts that if a simple[eg., material only] eval PLUS margin cannot reach alpha, quies() is going to fail-high on the full-evaluation-stand-pat-score and return - so, there is no point in making that call. it is futile
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.