Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Quastion: is rybka a clone-engine or principial a own code ? (n/t)

Author: Ryan B.

Date: 21:30:57 12/04/05

Go up one level in this thread


On December 05, 2005 at 00:03:42, Uri Blass wrote:

>On December 04, 2005 at 23:25:54, Timothy J. Frohlick wrote:
>
>>Eduard,
>>
>>vrajlich@alum.mit.edu
>>
>>Ask Mr. Rajlich. He might even speak german. Try english and ask him to answer
>>you in german if possible.
>>
>>PS How could a clone do this well? The author is an international chess master.
>
>I claim nothing about Rybka but basically a program can do very well even when
>99% of the code is from another program.
>
>TogaII1.1 or GambitFruit are based on fruit2.1 and are better than fruit2.1
>
>Uri

I do not think that Rybka is a clone and I agree 100% with what you say.  Gambit
Fruit is a clone and has almost no tuning time invested yet and the version I
just sent to Alex to put up on uciengines.com (Beta 4bx) scored just over 60% vs
Toga 1.1 head to head after 50 games (small sample set so worth little I know).
Of course I expect with a set of well rounded opponents Toga is still stronger
overall as Gambit Fruit has a lot of untuned new chess knowledge. Toga itself is
a great example that a clone can be stronger than its base also.  I think a
clone can be as strong as Rybka is showing but if it was a cone the author would
have said so I am sure.  I hope the great chess engine becomes public in some
way soon.

Ryan



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.