Author: David Dory
Date: 16:55:02 03/25/02
Go up one level in this thread
On March 25, 2002 at 17:08:23, Terry McCracken wrote:
>On March 25, 2002 at 14:40:39, Ed Schröder wrote:
>
>>On March 25, 2002 at 14:30:01, ALI MIRAFZALI wrote:
>>
>>>Actually the Computers are stronger than the strongest Human players.Human chess
>>>has too many weaknesses.But the funny thing is that every time a GM loses to a
>>>Machine ;there is an excuse.Sometimes even other people angry at the machine win
>>>come up with excuses for the GM.For example GM Rhode believes that Kasparovs
>>>loss to Deeper Blue in 1997 was because he was having a bad day.
>>>If a chess player comes up with an excuse everytime he loses a game he is
>>>indirectly claiming perfection.Computers still dont play perfect chess;but
>>>Human chess is FAR FAR FAR from perfect.
>>
>>You got it, I applaude.
>>
>>Ed
>
>I'm afraid I don't concur. I do think that computers are better than most human
>players and have reached the GM level marginally, (anti-computer play
>notwithstanding), but have not become better than the very elite.
>
>When Kramnik defeats Deep Fritz in their upcoming match, then what will people
>say?
>
>People can and sometimes do have legitimate excuses, computers OTOH don't unless
>the system goes down:o)
>
>I'll also add that strong masters, not even grandmasters who play alot of
>computer chess, will fair better than strong grandmasters who don't play
>computers often.
>
>Why, you may ask? Well computers and humans are very different in their approach
>to chess, and the human must adapt to the computer. When the human master or
>grandmaster becomes more and more familiar with computer play and different
>programmes his/her score against them will rise.
>
>If Kramnik plays a considerable amount of computer chess than the odds of him
>losing to a computer diminishes, as he now understands the methods to employ
>against them.
>
>People still learn, adapt better than any so-called A.I. system yet created.
>In a decade A.I. may help machines defeat humans in chess, maybe entirely?
>In 50 years I almost certain they will! But today the edge and "intelligence" is
>with the human player not the machine.
>
>Terry
Just as the human player has excuses for losing. So too does a computer program.
After all, the program is written entirely by humans!
Good excuses go both ways, how about:
1. Too little time to prepare (the person, or a new version) for a match, being
surely THE BIG EXCUSE for both GM's and software writers AND testers.
I don't understand your comment about people still learning and adapting better
than any A.I system, Terry.
OF COURSE, Rather obvious since WE people are writing the A.I. systems. AI has
done so poorly in chess I couldn't even tell you one competitive programs from
the AI class.
In chess, it isn't just a matter of "intelligence" or "edge in knowledge".
There's a BASKETBALL factor here (well, it is March madness). Regardless of how
much you know, what depth of theory you may have:
YOU HAVE TO PUT THE BALL THROUGH THE HOOP !!
You want to score the point in chess,
YOU HAVE TO MAKE THE RIGHT MOVES OTB !!!
Computers are still kinda dumb in some aspects of chess - but they're slicker
than snot in the execution of __anything__ they DO know. Look at Tablebases,
hah!! How slick is that??
There's no doubt that humans will continue to learn better things about chess.
And there's no doubt we'll teach some/most of that to the programs that play
chess.
The humans keep playing better, the programs keep playing better. In my opinion
the real kicker right now is the hardware improvement. That's a real nail in the
human's chess coffin.
If the software development was to stop, the computers would fade quickly
because their opening books would be "cracked" and flaws heretofore unknown,
would be in the spotlight. GM's would quickly learn how to exploit those
"cracks" and have a field day on all the top computer programs.
But that ain't gonna happen. The programs will be updated (and indeed the
computers will contribute heavily to new chess theory).
I think the future matches of top GM's and top programs with fast hardware
will be __very__ interesting. More knowledge to the GM, but fantastic execution
of what it knows, to the chess program.
(if you remember the movie "Escape from New York")
"Kramnik, come on out and Play-ay"
Dave
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.