Author: Christophe Theron
Date: 15:44:58 04/09/98
Go up one level in this thread
On April 09, 1998 at 16:10:17, Thorsten Czub wrote:
>>What you describe can be achieved by a "classic" minimax approach. As
>>far as I know, Chris Whittington did not claim he had a new search
>>algorithm, but he did claim that his evaluation was different.
>
>Almost anything in chris chess-engine looks weird.
>I am not sure if you can relate it very easily.
>We should ask chris. One thing is for sure: CSTal has very different
>evaluations, very different main-lines and sacs and also sacs where all
>other programs say: there is no sac possible. And suddenly, after a few
>moves, the others find out about.
That's why I went to a store in Paris right after the last day of the
WMCCC, I bought CSTal.
To my real shame, I didn't find a single minute till then to install and
test Chris W. creature. You know how busy I have been working on my
program. I worked a lot on the endgames recently, and used Genius as a
sparring partner (don't try this at home).
I was planning to play some games against CSTal to see how my program
could resist those wild sacs.
Chess Tiger, up to version 10, was a very slow program. I have been
trying for years to emulate a human way of exploring the tree. So I can
only have the greatest respect for Chris Whittington's work. I didn't
like the attitude he had when he was posting here. Too much provocation
and pedantic sentences. Even if he had written the best program in the
world I wouldn't accept that.
But I regret he doesn't post here any more. It is really a pity if he
stops any work on CSTal.
It is true that even a program with a simple design is very hard to
develop and tune. So maybe his very complex program is so hard to
improve and even debug that he felt the task was impossible?
Anyway, Mr W., if you read this, I would like to say that I feel no joy
when an opponent resigns. Especially a good one, and a very special one,
like CSTal. I really hope you come back here some day.
>If you want we can post a few strange positions only CSTal finds and ask
>the others what happens there. We can post the sendinfos/log-files of
>CSTal.
Yes I would like to see some. I'm sure you have plenty of them,
Thorsten. Don't you?
>I could e.g. let CSTal fight vs. ChessTiger (so that people will not
>say: ough - you cheated because you are against ANTI Fritz) .
>I like ChessTiger AND CSTal and we could see how both programs fight
>each other and you Christophe could maybe comment why Tiger played
>soandso.
>And maybe we get chris to comment why CSTal played soandso.
>If we choose Tiger vs. CSTal people would not THINK this is a campaign
>AGAINST a commercial program.
Maybe Chris can say why his program played this or that, but I'm almost
unable to do the same. I would have to analyze each position for a
really long time, and I'm not sure to find enough time to do that. And
maybe I don't want to reveal my weaknesses! :)
I'm going to try the CSTal I've bought, and see if it could give
interesting results. In the mean time, you could post some interesting
positions here in FEN or EPD format. Positions that really happened in
CSTal games would be the most interesting.
Feel free to post some positions from Fritz5/CSTal games. It would be,
at last, the first drops of concrete data about Fritz, after all that
blah blah.
>Maybe this would clearify enough to
>show where CSTal is different.
Let's begin with some positions along with their log files...
Christophe
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.