Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A Blast from the past - Feng Hsu

Author: Robert Hyatt

Date: 18:09:58 04/22/05

Go up one level in this thread


On April 22, 2005 at 17:25:25, Rolf Tueschen wrote:

>On April 22, 2005 at 09:18:38, Robert Hyatt wrote:
>
>>On April 21, 2005 at 17:14:34, Rolf Tueschen wrote:
>>
>>>On April 19, 2005 at 19:16:20, Robert Hyatt wrote:
>>>
>>>
>>>>Nothing in the contract forced IBM to produce anything.  Kasparov forgot about
>>>>that in the negotiations.  Had the match ended amicably I'd guarantee you he
>>>>could have gotten the printouts easily enough.  But I wouldn't give him time of
>>>>day after he went on stage and accused me of cheating.
>>>
>>>
>>>You confuse completely cheating to Kasparov with cheating to science.
>>Cheating.  (verb).  To violate existing rules when participating in a contest,
>>so as to obtain an unfair advantage and influence the outcome.
>>
>>Now exactly how/where did the DB guys "cheat science".  What "contest" was there
>>between the DB team (IBM) and "science"???
>
>Thanks for the question. My English sucks. Cheating science = violating
>methodological fundaments of scientific experiments. Here testing the strength
>of their machine against the best available human chess player. By tearing him
>into psycho war they could confuse Kasparov and the team can say they beat him
>fair, but then they forgot what the whole show was meant for. Again: do you like
>winning ugly? I thought no. Can winning ugly prove in any relevant issue PROVE
>something about the chess play of the machine? -- Nope! Certainly not.



I'll remind you of my "take" on this again.  The first "experiment" was to beat
the world champion in a standard-time-control match.  A good future goal would
be to enter a computer in the FIDE world-championship cycle and see if it could
play its way to the championship game and then beat the world champion.

Nothing wrong with the second experiment, and perhaps one day it will happen.
But until 1997 the first goal had not even happened, and it was the first
baby-step to take before trying the harder goal of "self-exposure" by playing
many public games where everyone could theoretically find your weaknesses and
exploit them.  And to date, every program I have seen certainly has weaknesses.
Of course they also have some terrific strengths to partially (or fully) offset
such weaknesses.



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.