Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Questions on the Alpha-Beta algorithm and other things.

Author: Sean Empey

Date: 13:23:35 02/17/00

Go up one level in this thread


On February 17, 2000 at 15:30:10, Jeff Anderson wrote:

>Hi, right now I am trying to learn about the Alpha-Beta algorithm, I have a few
>questions that maybe someone can help me with:
>
>
>With sophisticated move ordering, like that found in Crafty, how much time is
>saved (like in actual numbers or %'s) over having no alpha-beta alogirithm at
>all?
>
>With random move ordering how much time is saved over having no alpha-beta
>algorithm at all?
>
>
>Are there any top programs which do not use the alpha-beta algorithm in their
>search?
>
>If one has a program that creates a tree with no evaluations and no effort to
>find the best move, simply creating a brute force tree, how much faster will
>this be than when it creating a tree with evaluations attached (assume no
>pruning or anything to the tree)?
>
>
>
>Thanks in advance,
>Jeff

All programs use some sort of 'Alpha-beta' like searching (unless it is a
Brute-force searcher). If you did not it would take ages to search brute-force.
Your NPS will drop with Move-ordering, but will allow you to search the best
moves first. You can try both and see the huge difference between ordering and
non ordering (if you are using Alpha-Beta). I do not know the actual numbers in
%, but it is a big time savings to have a good move-order. So I guess the answer
would be a tremendous amount of time is saved using Alpha-Beta and move ordering
vs. Brute force. No question about it.



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.