Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: NULL move, general principle -- possible alternative

Author: Bruce Moreland

Date: 17:49:23 09/29/99

Go up one level in this thread


On September 29, 1999 at 19:54:01, Dann Corbit wrote:

>For NULL move, we have two kinds of situations:
>
>0.  A move is so bad I wish I did not have to make it because I would rather
>skip a turn.
>
>1.  An opponent's move is so bad they would not want to make it because they
>would rather skip a turn.
>
>In the first case, it seems we do not need to consider this move unless there
>are no alternatives.
>
>In the second case, similarly unless the *opponent* has no alternatives.
>
>What I am wondering is "why is the loss of a tempo the magic dividing line?"
>
>It seems the dividing line could be put anywhere we want.  Has anyone
>experimented with a floating [don't bother looking here] line that can be
>redefined during run time?  I know this sounds a lot like ordinary alpha/beta
>but what I mean is to change the threshold for null-move to something other than
>the normal value.

I haven't tried to deal with the first part of this, but I can handle your
question a paragraph from the end.

The reason you skip a move is in order to determine if the opponent has a
threat.  If their is bad and they don't even have a threat you have to counter,
you are probably doing great.

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.