Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Practical lesson for statistics

Author: Omid David Tabibi

Date: 10:56:59 01/01/04

Go up one level in this thread


On December 31, 2003 at 14:47:28, Vincent Diepeveen wrote:

>On December 31, 2003 at 12:25:11, Tord Romstad wrote:
>
>>On December 30, 2003 at 13:59:22, Vincent Diepeveen wrote:
>>
>>>SOS as we can see from your rating list is estimated by you at 2253 this is
>>>wrong it *always* wins from crafty.
>>
>>Are you sure about this?  I have never seen SOS play against Crafty, but what
>>you write above still seems rather unlikely to me.  My own engine scores around
>>45% against SOS (the version that ships with the latest versions of Arena), but
>>only about 20% against Crafty.
>>
>>Tord
>
>SOS kills crafty completely let me assure you that.
>
>You need a bugfree engine to kill crafty, that's all.

I wouldn't go that far. Crafty doesn't use unconventional weapons, but it does
use its conventional arsenal with a good efficiency. That is, although it uses
only normal algorithms, it is very well optimized.

Even though I believe that without some unconventional weapons it is impossible
to challenge the very top engines, it takes far more than a merely bugfree
engine to beat Crafty.

Speaking about SOS, playing against ParSOS in Graz, I noticed that it is using
very interesting pruning methods. In middlegame I never saw it searching less
than 17 plies, with a branching factor of about 2. So, it might be very well the
case that the latest SOS is considerably stronger than Crafty. The public
versions of SOS are improving greatly in each version. Falcon used to beat SOS 2
usually about 6-0, or 5.5-0.5. Against SOS 3, the result is usually about 4-2.
No longer results like 6-0...


>
>SOS is bugfree.



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.