Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Anand comment about Deep Blue

Author: Robert Hyatt

Date: 13:32:45 01/13/00

Go up one level in this thread


On January 13, 2000 at 14:37:13, Dave Gomboc wrote:

>On January 13, 2000 at 10:17:30, Robert Hyatt wrote:
>
>>On January 13, 2000 at 06:05:23, Tom Kerrigan wrote:
>>
>>>On January 12, 2000 at 14:45:13, Rajen Gupta wrote:
>>>
>>>>from the atrocious results of rebel vs grandmasters it is pretty clear that once
>>>>humans get the feel of playing against computers or even a particular computer,
>>>>they (ie humans)would get the better of computers.It is a fact that kasparov had
>>>>no opportunity at all to evaluate deep blue while the reverse was not true.In
>>>>all fairness kasparov should have been allowed a couple of weeks with deep
>>>>blue-he would have identified the holes and blasted through them-just as the
>>>>humans are now doing to rebel
>>>>
>>>>rajen gupta
>>>
>>>GK should not have been given a few weeks to play with DB prior to the match.
>>>That's silly and obviously unfair. However, I think forcing GK to go into the
>>>match cold was also unfair.
>>>
>>>GK prepares for matches by studying how his opponents played in previous
>>>tournaments. IMHO, DB should have played in a few tournaments before taking on
>>>the champ.
>>>
>>>But it's all a moot point now...
>>>
>>>-Tom
>>
>>
>>Remember that DB Jr _did_ play in lots of matches prior to the Kasparov
>>match.  Of course, many of those were DB-1 processors.  And the rest had
>>DB-2 processors, but only one SP processor.  And then the problem would be
>>that they were modifying the software right up to the match, which would
>>make it play differently.  And they modified it during the match, which
>>could make a hole found in game one disappear by game 2.
>
>Mods were done after game 2, not after game 1.
>
>Dave


Wasn't really saying they modified it period.  I said they "could" modify it
after any game where a hole was found, which would render old 'weaknesses'
information obsolete...

The point is that games played by a program 6 months ago probably don't say
much about games played by that program today.



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.