Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How many programs find 14. ...Nxe4

Author: Anthony Cozzie

Date: 06:48:50 12/30/04

Go up one level in this thread


<snip>

>When there are 3 PV's with the same eval... (in a non-forcing position)
>which move will be played?
>
>Don't know do you?

This occurs very rarely.  The one it plays will be the one it searched first (in
other words, it depends on the internal move ordering).  Sorry that isn't a very
good answer.

<snip>

>Please think about this, and comment further if you will.
>Thanks,
>CY

If you are a human, sure it is interesting tool for analysis.  If the computer
has some sacrifice that is "just below the horizon" then maybe it would work out
if you played down it a bit.  Or, perhaps you have a move you want to play, and
you are curious about the score relative to the best move.  No one is disputing
this, although you should be aware that generating multiple best moves slows it
down a LOT.  Shredder would probably have gone two ply deeper in 1-best-move
mode.

However, in terms of the computer "getting" a move - it doesn't count, for the
reasons I outlined in my last post.  You are arguing a point that, while
correct, doesn't affect the real question, or at least are using words that
don't fit with the accepted usage here.

anthony

P.S. The true eval is the eval of the best move (that is the assumption of
alpha-beta, that both sides play the optimum move).



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.