Computer Chess Club Archives




Subject: Re: significant math

Author: Uri Blass

Date: 22:50:38 11/19/02

Go up one level in this thread

On November 20, 2002 at 01:06:34, Robert Hyatt wrote:

>On November 19, 2002 at 23:54:40, Dan Andersson wrote:
>>I don't see how you can show that by comparing NPS of two wildly different
>>programs. Suppose one of them has a more efficient evaluation function. There is
>>no basis for comparision. Other than that they both play chess :)
>>MvH Dan Andersson
>Actually, at least for the case of Crafty and Yace, there is a good basis for
>comparison.  They are relatively equal in playing strength.  So any advantage in
>bitboards over arrays or vice-versa would seem to be _very_ minimal, which is
>what I have generally claimed for 32 bit machines...
>Not a great comparison, of course, but it at least suggests that on the PC,
>isn't much difference.  Or take fritz vs a commercial program.  Fritz supposedly
>is now a bitmap program.  Other commercial programs are not.  Yet there seems to
>be little difference in overall strength.
>Which lends more evidence to my "bitmaps are a break-even affair on 32 bit
>machines" conclusion...
>Of course, a "few" won't let such real data prevent them from producing heaps
>of disinformation...

I do not think that playing strength of different programs give evidence which
way is better because there are other differences.

The only possible evidence is results of the same programmer who does the
1)try both ways.
2)make his work public.
3)ask people for suggestion how to improve in both ways.


This page took 0.01 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.