Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Shredder 4 - Rebel Century Minimatch

Author: Harald Faber

Date: 12:55:29 12/29/99

Go up one level in this thread


On December 29, 1999 at 09:09:53, Ramón Martínez wrote:

>On December 29, 1999 at 00:11:26, Harald Faber wrote:
>
>>On December 28, 1999 at 20:15:52, Vicente Fernández wrote:
>>
>>>Here I post the games of a minimatch between Rebel Century and Shredder 4, on
>>>two computers, one k6III-400, 128 ram, and the other a K6II-300, 128 ram.  Games
>>>in 30 minutes each, handled manually. Programs played alternating computers.
>>>Rebel Century play with its strongest settings plus antigrandmaster= smart, 60
>>>hash. Shredder played with 90 hash.
>>>	Frankly, I expected much more from Shredder 4. A 3.5-0.5 result for Rebel was
>>>too much for me, but then came the surprise. As far as I can see, Shredder's
>>>book needs improvement: The World Champion starts thinking when Rebel still has
>>>three or more moves on its book. And the evaluation... just see for yourselves:
>>>it just keeps jumping from +40 to - 1.40 and so on, while Rebel's is steady and
>>>reliable.  I don't know what's the matter with Shredder 4 (yes, I turned the
>>>commentary off etc. etc.), but it doesn't look like a World Champ in my
>>>machines. And considering the unsteady evaluations, I won't use it either for
>>>game analysis. Compared to Hiarcs 7.32, Rebel Century, Chessmaster 6000, Fritz 6
>>>(light), Fritz 5.32 or Junior 5, Shredder 4 's evaluation is a joke. Please,
>>>take a look at the games and to the jumpy evaluations. Compare them to the ones
>>>you get from Rebel Century and see the difference.
>>>Vicente F. Herrasti
>>>
>>
>>Apparently you, like some others, underestimate Rebel.
>>And 4 games rapid chess is a) very few and b) not comparable with tournament
>>time control.
>
>I don't underestimate Rebel. Far from it. And yes, it is obvious that game 30 is
>not equal to tournament time control. And more yes, of course I know 4 games are
>not enough. My point was the irregular evaluation I noticed on Shredder 4. As an
>example, look at the first game, move 31: eval: +.33. MOve 32: -2.83. Century
>saw this 3 or 4 moves before and Shredder didn't even notice.


This happens to all programs.


>By the way: ¿Why
>did you understand that I underestimated Rebel? I cannot read anything in my
>post that leads to such an opinion. Neither did I said anything like "4 games
>are enough".


Your tone was a bit disappointed, such as if you expected Shredder4 to win
clearly.




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.