Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: WCCC and some NPS observations

Author: blass uri

Date: 09:26:09 06/23/99

Go up one level in this thread



On June 23, 1999 at 11:11:26, Thorsten Czub wrote:

>On June 23, 1999 at 10:26:31, blass uri wrote:
>>The question is if they still use a root processor or do something more
>>intelligent like a proccesor not of the root.
>>
>>Did you see cases when the score of the position was changed  after trading
>>queens and not because of the fact that Fritz saw deeper?
>>
>>Uri
>
>Fritz had 0.00 and -0.03 and -0.12 and 0.06 etc the whole game !
>over and over draw-scores.
>It seems unable to evaluate in a normal way.
>Like in Paris. senseless try out of moves without having any idea about
>the score ... leads to losses. or draws. rarely to good games.

I think that these scores may be sometimes reasonable.
If the position is unclear then it is reasonable to evaluate it as close to
draw.

The main problem that I find with Fritz is not draw scores but the fact that it
is a root processor.
It lost a game in Paris against stobor because it did not see the results of
trading queens and evaluated it as close to draw before the trading and after
the trading suddenly saw that the evaluation is negative against it and not
because of the fact that it could see deeper.

It could draw this game by repetition instead of losing if it could see the
result of trading queens before trading queens and not after it.

I think also that the opening in this game was not good and did not let it to
play for a win but this is another story.

Uri



This page took 0.01 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.