Computer Chess Club Archives




Subject: Re: A story about NPS

Author: Ulrich Tuerke

Date: 14:36:24 01/26/01

Go up one level in this thread

On January 26, 2001 at 12:51:31, Severi Salminen wrote:

>>If you "advance" to a pure minmax search, you'll again get a higher nps.
>I don't think so because you'd have to generate all the moves in every node.
>But, hey, I have an idea! A minimax search algorith which would produce only one
>random move at every node! That would be _really_ fast!!

I think, the reason is that there are no cuts in minmax, whereas in alpha-beta
as a consequence from the alpha-beta cuts, you will generate many moves in vain.
So, the ratio #executed moves/#generated moves is far higher. In this respect,
pure minmax is more efficient than alpha beta.
You'll observe the analogous effect when you have bad move ordering: less cuts
==> higher nps

In fact, thÝs is a rather drastic effect. Try it.


This page took 0.03 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.