Author: Bruce Moreland
Date: 00:05:59 08/14/01
Go up one level in this thread
On August 13, 2001 at 07:53:22, Harald Faber wrote: >On August 13, 2001 at 07:14:48, Gian-Carlo Pascutto wrote: > >>On August 13, 2001 at 04:45:37, Harald Faber wrote: >> >>>And what makes me wonder is that how can Shredder say +0.2 at d=10/11/12 when >>>directly after Rxb3 Shredder only needs d=8 to see the disaster. >> >>Just simple nullmove I think. >> >>-- >>GCP > >I thought nullmove only fails in Zugzwang positions?! Null move screws up if you give up material and then take a while to quietly smash the opponent. Here's an example: Set up the typical case of a black pawn on f3, white pawns on f2, g3, h2, king on g1. White to move, 4-ply search: <some dumb move> Qh3 <some other dumb move> Qg2# No problem. This is detected and possibly avoided by substituting "<some smart move>" for "<some dumb move>". With null move R=2, white passes the first move, and gives black one ply (4-2-1) to generate a threat. So: <pass> Qh3 No problem, white is happy and cuts off because, say, black sacrificed a pawn to achieve this situation, and since the pawn isn't recovered we're certainly below beta. White is nowhere near to finding this, actually; even an extra ply doesn't help. I tested the Rxb3 problem on my program with R=3, and it took about ten times longer to find it than with R=2. bruce
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.