Author: Torstein Hall
Date: 15:11:25 07/23/01
Go up one level in this thread
On July 23, 2001 at 16:31:55, Mogens Larsen wrote: >On July 23, 2001 at 12:37:08, Torstein Hall wrote: > >>That an IM has the capability to defeat one of the top programs on fast hardware >>under match conditions has not been seen for quite a while. So if no one has >>data to prove me wrong, I think the demand for proof now is on the "Computer is >>not GM" side. We have seen so many comps fighting equal or beating GM's, so now >>I want to see an IM beating the computers! > >It seems like you've missed some of the content of my message. One of the main >reasons why the "Are computer programs of GM strength or not" question is >ridiculous, is the lack of actual matches. Therefore the actual knowledge about >computer programs as opponents from a human perspective is very low, because it >isn't necessary and worthwhile to prepare for that option. So there's no actual >competition and since the question is based on that presumption, it is >meaningless by default. I'm not against computer programs being awarded an >imaginary GM title. Especially if it can save the mental health of a few >persons. But I'll wait until an imaginary GM title seems less silly and makes >sense. I actually agrees on that part. I find it ridiculous to award Computer the title GM under any conditions! (Its a title for humans having taken part in the right tournaments with a high enough score.) But I interpret it in a different way, as meening: Are computer as strong as a (mediocre) GM. I think they at least are getting pretty close. And as you know, the computers rarly tires etc. etc. And now there are some evidence for that strenght. We had Van der Wiel - Rebel match, Hübner - Fritz and Tiger wiping the floor with a pretty strong bunch of GM's in Argentina. (I must surly have forgotten some other matches as well.) So I think the latest results clearly points in that direction. >So the statement that "That an IM has the capability to defeat one of the top >programs on fast hardware under match conditions has not been seen for quite a >while" is true, but it doesn't reflect that computer programs are proven >superior. My hypothesis is that it's still possible for a computer knowledgable >IM to play an even game against any computer program on any hardware. Mainly >because Huebner performance wouldn't scare the pants of any IM IMO and it is >copyable. > >>PS I still belive Kramnik will make Chop Sui ( How do I spell that Chinese >>dish?) of Fritz. > >Maybe, maybe not. It would be the first _real_ encounter by my standards between >a GM and a computer program in a long time. Anything better than 4.5 points for >the computer would mean that they've reached something comparable to GM strength >given their current configuration. Anything less than 4.5 points that includes a >win would also be significant. Is it going to be a 9 game match? Then I predict 3 wins for Kramnik with 6 draws. Kramnik can draw anyone, or in this case rather anything, when he likes to do that. Best regards Torstein > >Regards, >Mogens
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.