Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Doesn't appear to work for me neither.

Author: Vincent Diepeveen

Date: 18:40:27 11/21/02

Go up one level in this thread


On November 21, 2002 at 17:30:27, Antonio Dieguez wrote:

You are comparing search times with each other. That is
correct way to do it. He isn't comparing search times with
each other.

Also he claims that his near to fullwidth search is just needing
only 50% more nodes than with only nullmove. God knows how he
managed that. For me fullwidth versus a search with nullmove
is plies difference. Not 50% in nodes.

So his own results, seen from time viewpoint also show clearly
that nullmove is superior to his verification search.

Bad new name for something already existing.

>
>I get more nodes per ply than with r=2 even.
>And less test positions solved, that is.
>
>I did the tests at few seconds, and the searches to fixed depth to average depth
>of 8. I could try longer searches but it doesn't smell good.
>
>I get very similar results with and without the research for zugzwang.
>
>I have seen papers that say that something is the best, or better than other
>thing, for example that MTD is superior to other searches and jokes like that (I
>mean the strong claim as a joke. It could be true but just because it worked for
>him/them is nothing.), I though this time was going to be different... but it
>looks it won't... and look at the start of the Conclusion phew! you could have
>put some "In Genesis,..."
>
>
>
>Still, I will keep trying to make something like this work (as many other
>things), thanks for sharing.



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.