Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rebel's anti-GM option

Author: Don Dailey

Date: 20:09:39 06/23/98

Go up one level in this thread


On June 23, 1998 at 18:57:32, Ed Schröder wrote:

>>Posted by Robert Hyatt on June 23, 1998 at 14:58:42:
>
>>Let me pass some friendly advice along to Ed:
>
>>Ed, if it is your intent to be "speculative" against Anand, I speculate
>>that you are not going to win a single game.  You *cannot* get away with
>>that sort of stuff against him.  You need to visit ICC, where I can put
>>you in touch with Yasser, who is a horrific calculator.  He will show
>>you why "speculation" is a long-term alias for resigning...  These guys
>>can calculate, and they can calculate just as good as the best of the
>>programs around can.  You have to out-play them, *not* try to out-
>>speculate them.
>
>Thanks for good advice. Rebel played once or twice against Yasser at
>Aegon and Yasser is indeed a computer killer, he simply knows how to...
>
>I have experienced in all these games Rebel played at Aegon against GM's
>a pattern how they do that. I was there behind the board and could see
>it happen with my own eyes.
>
>It's my opinion a clean Rebel (like 9.0 or even 10.0) without any extra
>is without a chance against a player like Anand. Since I know based on my
>Aegon experience how they (in general) do the all the computer killing
>I wrote a piece of software that tries to de-stabilize THEIR system.
>
>If they play anti-COMPUTER style I fight back with anti-GM style. What
>I have now is a first serious try.
>
>>Yasser is a particular handful, and he's not as much of a problem as is
>>Anand.  I've played Anand some games, but promised to not reveal the games
>>nor where they were played.  He is an amazing calculator.  Don't sell him
>>short and expect complicated positions to confuse him...
>
>And Anand is in TOP condition too looking at all his latest tournament
>victories :(
>
>>There is no "anti-GM" strategy, other than simply playing good, solid
>>chess.  Anything else will backfire...
>
>I do not agree with you. There is an anti-GM strategy and it is just as
>real as there is an anti-COMPUTER strategy. Only for us (the programmers)
>it is perhaps(?) a new area where we have to put in time and effort
>before we see the first results. That's normal with all new developments.
>
>I surely believe my anti-GM option will play better against Anand than
>playing without it. And I am pretty sure Hsu and co have something
>similar in Deep Blue BECAUSE YOU NEED SOMETHING LIKE THAT if you want
>to win against these giants.
>
>- Ed -


In my humble opinion, there are anti-GM strategies available.  What
Bob says is pretty valid though,  it's sort of like playing with fire,
but you can do this if you are a trained fire handler!

In many positions there is a choice of good moves, all of them
objectively equal in the game theoretic sense, but not equal in
your opponents ability to deal with them.  The trick is to lean
toward the right type of moves without weakening the program.  I
see no reason why this cannot be done without compromising the
program.    I think that is the only real question, can you do
some things that might affect the programs style of play, and
yet does not weaken it?  I don't see why not!

If you try to get speculative or deliberately play provocative
moves that may be weaker objectively in order to confuse the
opponent, then perhaps this is a mistake.

Also you want to get a tactical position, but not just any
tactical position.  There are certain tactics that humans are
still much better at as Bob says.  You want the tactics to be
such that MANY lines must be explored.  Really long forced
sequences will favor the strong Grandmaster who will no doubt
search much deeper than the computer.  We really need a highly
tactical position with many plausible moves at several levels.

I don't know if this is easy to achieve but I am convinced that
although Bob is right in general about the best humans being
better calculators, there are types of positions where computers
will outperform them.  I do not know how easy it is to encourage
these kinds of positions (with or) without weakening the program.

Ed, can you share any of these anti-GM ideas with us or is it
a commercial thing you are not at liberty to discuss?  If so I
can respect that, but I would like to hear them.

- Don




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.