Author: Dann Corbit
Date: 14:05:45 07/15/02
Go up one level in this thread
On July 15, 2002 at 17:02:42, Dann Corbit wrote: >On July 15, 2002 at 16:43:39, Gian-Carlo Pascutto wrote: > >>On July 15, 2002 at 16:19:32, Dann Corbit wrote: >> >>>On July 15, 2002 at 16:09:57, Gian-Carlo Pascutto wrote: >>>[snip] >>>>Have a foolproof method for identiefying zugzwang? >>> >>>Sure. >>> >>>Never bother analyzing the root node. Analyze all forward nodes from the root >>>instead at ply zero [from current]. >>> >>>If every evaluation is worse than the current evaluation before analysis, then >>>you are zugzwang. >>> >>>It's also handy for move ordering. >> >>This doesn't work :) >> >>(Explanation left as exercise to the reader) > >It leaves false positives. It also assumes an exhaustive ply 0 search (eval). Can you show a case where you are zugzwang and one of the forward moves looks better for the side to move? At any rate, it is a simple place to turn on double null move instead of regular null move. In the case of a fail low for the current position, it's a good idea to widen the search anyway.
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.