Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Anyone disagree with this?

Author: Gerd Isenberg

Date: 09:16:35 10/11/05

Go up one level in this thread


On October 11, 2005 at 10:37:39, chandler yergin wrote:

>Thanks again Richard for your input & views
>PLease let me know if you agree with this Link on how computers work, and we can
>continue the discussion.
>
>http://articles.roshd.ir/articles_folder/mohandesiScience/computer/Howstuffworks%20How%20Chess%20Computers%20Work.htm

Hi Chandler,

The article you mentioned seems a bit over-simplistic and even wrong.

"the algorithm can run about twice as fast and takes a lot less memory"

Better read bruce pages on mini/max and alpha/beta.

http://www.seanet.com/~brucemo/topics/minmax.htm
http://www.seanet.com/~brucemo/topics/alphabeta.htm

To get the idea of alpha beta:

Let say you (or a program) analyze a position and you found a plausible main
variation for one of your moves resulting in an equal position. Then you start
to look for your other moves to possible find a better one, considering
opponents replies as well of course.

You try a "weak" move and your opponent wins with the first considered move let
say a pawn. That is already deterrent enough for you, to don't play that "weak"
move. You don't need to look for possible stronger refutations of your opponent
- may be he even wins a piece or mates you - don't care.

This is called a beta-cut.

Gerd



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.