Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Human vs. human "anti-chess" ?

Author: Rolf Tueschen

Date: 15:01:35 05/22/05

Go up one level in this thread


On May 22, 2005 at 16:45:24, Robert Hollay wrote:

>  I'm not a strong chess player myself and cannot find a satisfactory answer
>to this question.
>  I was thought that so called "anti-chess or anticomp-chess" could be prevented
>with just making some trifling changes in engines.
>  But some engine-programmers expressed in this forum that if they would tune
>their engines against "anti-chess" players, then the engines would be weaker in
>playing "normal chess". Consequently, it's not possible to write an engine
>which could play the strongest possible "normal chess" and at the same time
>efficiently prevent the opponent to play "anti-chess".
>My question is:
>  Theoretically, is this applicable (to a certain extent) against human players?
>If relatively weak players can draw against the top engines, why couldn't an IM
>achieve always a draw against a GM? (using the so called "stonewall" technique)
>
>Robert


Easy one. A GM is not as dumb as a machine. That's the reason why
Anti-computer-chess does exist at all. Think about it for 10 seconds. If you
play Anti-Computer-chess you reduce intentiously your chess talents. You
concentrate on the machine and its alleged weaknesses. I can't see how you could
"tune" your chess for play against human GM. As Kurt said a GM is better anyway
and he sees what you are doing.

Let me make a Rule for a special chessplay where you intentiously play a system,
a very solid piece positioning without intending any active play. Such systems
are known under names like Hedgehog. You can also find in many games of the
german BL the so-called Saacke-System. It's a Caro-Kann with g6 and then Bc8 to
g4, then e7-e6 and later c6-c5 with a very solid position for Black. First time
that was played by no other than Petrosjan. A German expert player of the
category A created with the help of his team members a complete system with
really sophisticated key moves. If White doesn't know the details, Black has an
easy game. But - and this is now the backside of the medal. In the hands of a GM
who also takes care of the important White move h2-h3 against Bg4 - will smash
Black because the position simply is too passive for Black. So, if a GM as White
who knows the details will win against a weaker Black even if Black knows all
the tricks. - So, by definition a GM doesn't play that system with Black.

In other words, a GM knows such strong key moves for all sorts of systems. And
because such systems always are beyond modern chess with its deep ressources,
it's enough easy for a GM to win the game.

Take a machine and compare it with such GM then you directly understand why
machines can well be surprised with such tricks even if the chess is second
best. - It's also clear that all the programmers want to avoid all such systems
because they know how badly the machines do play in such positions. This is what
I call "tuning" the machine against a player like Eduard. In dozens of published
games Eduard has shown where the difficulties are for a machine. All that can be
avoided if it came to a public match between Eduard and a machine. Such events
don't make sense, simply because even if the machine is winning with 100% it
wouldn't be big news!



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.