Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Garry still singing the same Deep Blue blues...

Author: Hans Gerber

Date: 10:25:03 05/05/00

Go up one level in this thread


On May 05, 2000 at 10:12:34, Robert Hyatt wrote:

>On May 05, 2000 at 06:40:51, Hans Gerber wrote:
>
>>
>>Just a few thoughts.
>>
>>There are at least two possible levels of discussion. There is one with quick
>>and smart reactions, sort of short-cut winning points, that is the - let's call
>>it the - non-historical one. The other one is the scientific/ historic level,
>>the one with always integrating at least all the known collected facts and
>>already made arguments.
>>
>>The point here with these alleged "print-outs" at the IBM site is simply that it
>>was already mentioned long ago (in 1997!) that if IBM had wanted to fake these
>>print-outs, they could have done it in _less_ than an hour. In other words, they
>>could never be a proof. It was R. Hyatt BTW who explained that.
>>
>>Back to the main question.
>>
>>The ball does not lie in K.'s court. It is in the court of those who promissed a
>>detailed publication of the output of DB in a scientifically sound manner. What
>>does this mean?? Well, that the output is analysed _and_ (most important) proven
>>as the real/ original output of the machine. Does anybody believe that this
>>proof has been presented somewhere? Even more: Could this be done -
>>theoretically?
>
>How does the ball lie in IBM's court.  They released the files.  Now they have
>to do the impossible and prove that the logs were from DB?  How?  Couldn't they
>(by now) have modified the DB software so that it _would_ produce those exact
>moves and evals?  This is impossible to prove.

Exactly. But what does this mean? That the non-scientist Kasparov of course had
overlooked to give the IBM scientists the _duty_ to guarantee the laws of
science (also the moral) in that scientifical experience. He didn't do that
because he was sure (unfortunately!) that Hsu et al. would feel obliged to
follow those laws/ rules.

A specific example. During the "match" K. had the impression that something was
wrong and he asked for the prints. If the IBM team had given him the prints at
that moment, the whole irritation afterwards would have been omitted. After the
match, after two more years, the whole data doesn't prove anything. I agree.

But whose obligation was it to prove that the machine gave that original output?
Did Kasparov forget another rule for the match? I don't think so.
He couldn't expect that the scientists suddenly would deny him the objective
data.

This alone however is enough for doubts in the scientifical integrity of the
scientists around Hsu. The denial of the prints was more a psychological weapon
which led to K.'s understandable paranoia. Scientists in a friendly exhibition
match, called experience, called Man vs Machine by the massmedia, are not
expected to make use of such instruments. For moral reasons and genuinely
scientific reasons. Because they could never more prove the authenticity of
their data. This however is the least one could expect from a scientific
experience.

Baseline. Perhaps it's not the ball lying in their court, but Hsu et al. lost
their scientific image.

>
>
>>
>>IMO in a certain view, Kasparov's points of critic are still open, unanswered
>>and unsolved.
>
>What is open.  The fact that he is a whiner and sore loser?  Those are pretty
>evident.  The fact that he might be the best chess player of all time?  There
>is plenty of evidence to support that too.  The fact that he is the worst sport
>of all time?  Plenty of evidence there too.  But there is no evidence to suggest
>DB didn't play all of the moves.  Most of them were reproduced by
>microcomputers.  The logs look sane and support the games completely.

However no proof.


>
>
>
>
>>
>>The only ball in K's court, that is the one of the impossibility of explaining
>>the difficulties of such 'questions' to average people. So, it might _look_ as
>>if he, G. Kasparov, is the unsportive poor loser. As if he had the obligation to
>>solve the difficult problem which was and still is genuinely the one of the IBM
>>researchers around Hsu ...
>>
>
>It might "look like his is a poor loser"?  I don't think there is one
>scintilla of doubt.
>
>But even worse, he _continues_ the nonsense.

'Nonsense' because he can't prove that something wasn't quite kosher?
Interesting view ... That Hsu et al. didn't behave like scientists, Kasparov is
to blame for? As I said, of course he looks like a sore loser. _But_ he isn't.
Hsu et al. did something, scientists are forbidden to do. They look worse than
just sore losers.

The question of that exhibition match was what the machine could achieve
_chessically_ against Kasparov or vice versa. The question Hsu et al. did answer
is, could Kasparov be irritated by some clever psychowar tricks (by the human
beings of the team, not by the machine!). The answer is "Yes, he can be
confused". And of course he can, because he's a human being. But was that really
the _question_ of the match? I don't think so.



This page took 0.06 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.