Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The engines of the year 2004

Author: Uri Blass

Date: 14:52:04 12/27/04

Go up one level in this thread


On December 27, 2004 at 15:44:40, Peter Berger wrote:

>I was very much impressed by ChessTiger2004 recently, that performs very
>strongly on my computers, to an extent that I am much surprised by some bad
>results by others, especially in the SSDF matches. But of course I haven't
>played 1000s of games with it yet :).
>
>When it is about general news the engine of the year should maybe be Diep that
>finished 3rd in the WCCC and won the following tournament in Leiden too.
>
>Talking about huge improvements Movei also comes to mind. I constantly have to
>change its sparring partners - engines like MChess and ChessSystemTal are no
>match anymore, when they used to spank little Movei only half a year ago. I
>think Movei has reached ChessGenius level by now ( the chessprogram I mean).

Thanks but I think that no engine was close to the improvement of fruit.

Fruit was non existent before march of the year and suddeenly it is competing
for the title of the best free engine.

Part of the improvement in movei is thanks to the fact that I cared to look at
the code of fruit but there are still many things that Fabian does better than
me including complicated evaluation function that fabian for some reason
describe as simple.

Even Fruit1.0 evaluates some details about pawn structure that movei simply does
not evaluate.

Note that I tried to evaluate backward pawns but it was not productive based on
my tests so I left it but another idea that is used by fruit to do average
between opening and endgame evaluation based on the stage of the game seems to
work in movei based on results.

There are many ideas in fruit that I simply did not try.

Uri



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.