Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What seems to be forgotten in the Fritz - Kramnik match

Author: Robert Hyatt

Date: 12:27:57 06/06/01

Go up one level in this thread


On June 06, 2001 at 15:00:15, Mike S. wrote:

>On June 06, 2001 at 14:05:14, Robert Hyatt wrote:
>
>>On June 06, 2001 at 12:47:56, Mike S. wrote:
>>>(...)
>>>It wasn't only about the strength level, but also about the "mentality" of chess
>>>programs. I think, a strong micro was the best approximation available.
>
>>I don't agree.  He made game decisions based on how a program (Fritz) would
>>play.  Based on what Fritz could see tactically.  That is about like taking
>>your favorite naval ship and challenging a brand new ship that has never been
>>seen before.  Perhaps they will sink you before you get within 500 miles.
>>
>>It was just the wrong plan...  because to try to play the opponent based on what
>>it will likely do, when you have a very faulty idea about what it can do, is not
>>the way to win the war.  (...)
>
>I understand what you mean... although, on the other hand I'm sure that Kasparov
>of course was aware of that DB could see much deeper, or more respectively.

That is where you are wrong.  Hsu mentioned once that at a press conference,
Kasparov was asked how he thought DB2 would do against a program like Fritz.
He said it would probably win more than it loses, but not by much.  Hsu had
the results of his matches against Fritz already, and he _knew_ that Kasparov
was going to be in serious trouble, because Kasparov _really_ believed that
Fritz was a near-equal.

The rest is history.  The advice given to him was horrible, and he suffered
because of it.  If he had asked the _right_ people, at least he would have
known he was facing something terribly unlike Fritz of 1997.

> But
>it's true I think, that he was too busy thinking (and doubting) what the machine
>might be able to do and what not. The doubts were disturbing him.


No doubt about it.  And his advisors were helping this misconception take
form in his mind by telling him things that were 100% wrong, because they
had _no idea_ what they were talking about.  Self-appointed experts that
really didn't understand DB at all.




>
>Do you think it is much easier now for Kramnik? The Fritz version he will be
>able to prepare with, will be not, or not much different from the match
>version... only the faster hardware won't be available for him AFAIK. Maybe this
>doesn't matter that much.

It won't matter at all.  Fritz that plays Kramnik will have most of the
weaknesses of the current Fritz.  It might have a few fixed, probably a couple
added in unknowingly...



>
>What result do you expect from the match (8 games)?
>
>Thanks,
>M.Scheidl


If it is even close, it would be a great result for Fritz.  I would be
happy (if my program were in this) with even a 2-6 result.  Should I get
more than two points (probably 4 draws) I would be quite happy.  I'm not
sure what to predict since humans are so unpredictable.  2-6 would be
reasonable, but then Fritz could surprise us if Kramnik doesn't pay close
attention and it could do better.  Perhaps significantly better.  I can't
see it having any significantchance of winning the match, however.





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.