Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Concerning GM Strength in Top Programs

Author: Micheal Cummings

Date: 20:22:37 10/03/99

Go up one level in this thread


On October 03, 1999 at 15:44:11, Fernando Villegas wrote:

>
>
>Once again, due to Rebel result against a GM, some discussion has arisen respect
>real strength of top programs. Are they equal or near equal GM players? They are
>not, by far? What?
>Arguments against Top programs having reached GM level tends to remark that too
>few games at tournament rithm has been played  to do possible to count wins or
>draws of top programs against human opposition, but then the same is truth for
>the opposite: too few games has been played to take into account lost games as a
>decisive proof  of non-GM level in top programs.
>Impressions? There are many. What Thorsten say, that he is capable of measuring
>the level of a program just looking at it playing a game or even looking some
>moves made by it in the game, could sound too extreme for many people here, but
>that is exactly what we do if we qualify a program without enough results to
>tell.
>Beside, there are some conceptual problems: what is GM level? That of Kasparov,
>Karpov and a dozen guys on the top twelve rank?   Or that of “weak”, normal, GM?
>GM is a too wide category for the facts that enclose. 50 points of difference in
>this range is far more significative than in the expert or just Fide master
>level.  There are worlds of difference in strategical and technical
>understanding between Kasparov or Karpov and any GM player 50 points less
>strong, but in expert or master level there are only some degrees of better
>handling of conventional tactics and technics.
>And then we have the very old discussion about tactics and strategic capacities
>and the rol they perform. “Programs does not know nothing of this compared with
>real GM, they just are good tacticians, even relentless and unsurpassable in
>that area”, but then, how much strategic or tactical is chess in itself?  Even
>in GM playing, how many games are won and lost because tiny strategic
>advantages? I bet than many, many GM games are won and lost for the same kind of
>tactical reasons we, amateurs, win or lose, but only with more intricate
>patterns in more hidden, sophisticated variation lines.
>But then, forgetting that, each time we see a game won by a computer against a
>very strong human player on the ground of tactic blows, we tend to devaluate it
>because “was not won” with a flashy show of strategies or technics. And if the
>contrary happens, it is said, then, that “even” in tactics GM are superior to
>machines.
>So, we have many sides to support our biases, but the only one thing we lack to
>finish the argument is  just enough games, many  classes of games. Until then we
>cannot say too much except to talk loud  about our tastes and impressions.
>My impression? Top programs are in the area of IM masters, maybe also biting the
>line of very weak GM. And that is enough to give them the chance to win to a
>strong  GM from time to time. But it seems they lack the ultimate degree of
>subtility and understanding a big GM have. Of course this is only an impression
>Fernando

I think they are of GM strength, With a big enough opening books and the engine
continually updated and fine tuned it will always stop a GM from creating killer
lines and plays against a program.

At the moment with a few games being played to either deny or prove the point,
we will never know soon.

As for comments from GM about strength of programs, I take very little notice, I
look at their obsevations but most of the time they say they are not up to
standard of Human GM's, but interestigly enough, this is usually after they lose
to a program.

Find one GM who will support a machine being stronger than a human brain which
programs this machine.

Plus as for the personality and conduct of some of these GM's making these
comments, well who knows. Some GM's are less than nice people. Genius's in their
own mind.



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.