Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Artificial Intelligence in Computer Chess

Author: Sune Fischer

Date: 15:14:59 03/28/04

Go up one level in this thread


On March 28, 2004 at 17:37:55, Artem Pyatakov wrote:

>Where are you getting the information that humans use Alpha-Beta in every game?
>There is some evidence of forward search in GM analysis, but the majority of the
>decisions are made on the pattern-recognition level, not with search.

I don't see the two as mutually exclusive, you need some kind of search to
support the pattern-recognition.

You can probably also say that humans often think depth-first, for instance

[D]r5k1/5p1p/pb4p1/1p1q4/8/3B1P2/PPP3P1/1K1Q3R w - - 0 1

The pattern here would be the unprotected black queen.
First you note the white bishop can't give checks.
Then you note that _if_ the black king was on a different square, say f7 or h7
the bishop could take on g6 and afterwards win the queen.
So how do we get the black king to h7? We can take on h7 with the rook!

So the variant is mentally constructed backwards and the sacrifice seems quite
natural.

Before comitting to the sac you must examine black's options after Rxh7, he is
not forced to take. For instance he might have a check which saves his queen and
refutes the line ie. he fails high on the check.

All in all I think that is somewhat similar to what goes on in alpha-beta.

-S.



This page took 0.11 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.