Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Computer haters?: No, you are realistic!

Author: Jeroen Noomen

Date: 10:51:58 07/17/00

Go up one level in this thread


On July 17, 2000 at 12:17:32, blass uri wrote:

Hi Uri,

>I read in chesskasparov site that Junior could avoid the problem against piket
>by 8.h4 and if 8...h6 9.e6

Yes, it is 'could'. But it wasn't played, so it remains speculation.

>The difference between the scores of 8.h4 and the move that was played is
>small(at least for Junior5.9) and it is possible that Junior could find h4 with
>better hardware.

It is not a hardware problem. THE problem is that most computer chess programs
do not understand the formulating of a plan when having a blocked position. This
is not only about one move, e.g. 8. h4. Even without 8. h4 it had quite a good
position, but simply had no idea what was going on. Believe me Uri, I play chess
for 25 years now and YOU CAN'T SOLVE THIS PROBLEM BY STRONGER HARDWARE.

>piket could not be sure that he will win before the game because he could not >be sure that Junior will not find 8.h4.

Nothing is sure in life. Piket's strategy was brilliant and if I didn't know who
was the White player in this game, I seriously would think it was a game played
in a simultaneous exhibition.

>If the definition of GM strength includes to play weaker than 1800 players in
>some positins than I guess that computer without GM strengh can get 3000 rating
>in the future because even if they understand the stonewall and do not
>understand fortress positions they will not be GM strength by your definition.

I don't understand what you mean. IMO a super GM can play all types of positions
and he has the UNDERSTANDING of several thousands of plans. Furthermore, a super
GM is able to recognize patterns and is able to adjust himself to the opponent.
As long as a computer program doesn't know 'blocked positions and king's attacks
are my weak point, I should avoid them', or the programmers stop using all their
efforts on NPS and NOT solving these types of position, GM's will do one simple
thing: They have LEARNED from this tournament. They now know how to play against
the machine. And as long as nothing is done to prevent disasters like against
Piket, in my opinion a chess computer program has no GM knowledge.

But this is repeating my view, since I already said: GM RESULTS yes, GM
UNDERSTANDING no. One last simple line to support my view:

In open positions DJ scored 4 out of 6, which is a TPR way above 2800.
In closed positions DJ scored 0,5 out of 3, which is rather bad.

Jeroen



This page took 0.03 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.