Author: Don Dailey
Date: 18:34:57 07/20/98
Go up one level in this thread
On July 20, 1998 at 18:00:51, blass uri wrote: > >On July 20, 1998 at 17:20:11, Robert Hyatt wrote: > >>On July 16, 1998 at 04:33:06, Amir Ban wrote: >> >>>On July 15, 1998 at 16:50:09, Robert Hyatt wrote: >>> >>>>On July 15, 1998 at 11:03:10, Danniel Corbit wrote: >>>> >>> >>>>I agree with you 100%. However, if you look at the Fredkin prize award >>>>information, DT was clearly playing "at GM strength, based on a >2550 >>>>rating for 25 consecutive games, computed using normal rating procedures." >>>> >>>>But, as you pointed out, it wasn't a "GM" in the FIDE list. It might well >>>>have been one in the USCF listing, there I don't know. There are multiple >>>>federations that award GM titles of course... Only FIDE awards the IGM >>>>title. >>> >>>Playing 25 games at performance of 2550 doesn't get you a 2550 rating. If you >>>started out at 2400, for example, you will advance to only about 2450. >>> >>>Amir >> >>I believe that I gave one wrong impression and one wrong piece of data, based >>on re-reading some old literature I have here. >> >>1. The Fredkin prize required a >2500 performance rating over 25 consecutive >>games. >> >>2. Deep Thought produced a performance rating over 2650 for 25 consecutive >>games. >> >>The rating was, (if my old email from Hans was/is still valid) computed as the >>usual sum(wins+400, draws, losses-400)/N.. >> >>Which means that you had to produce a performance rating of 2500+ and *maintain* >>it for 25 games so that you couldn't have a short "spike" and get over the hump >>easily. But it was a performance rating, which means it was only applied to any >>25 consecutive games they played. >> >>I had overlooked the >2650 rating they produced however (this was deep thought >>2 IIRC) which was far slower than DB or DB-2. But >2650 is still quite an >>accomplishment... regardless of how you look at it... and it couldn't be >>blamed on "computer shock" either as there were plenty of games circulating >>around for opponents to study. > >I think humans did not know to play against computers When DB had the >2650 >result like they know now. >I have a game of Deep Thought in 1988 against a commercial machine >Mephisto >Deep thought did not win convincingly and Mephisto missed a win in this game. >the game is in the book how to beat your chess computer by David Levy > >I believe the version of Deep Blue that played in the computer championship on a >slow machine was better than DeepThought (otherwise DeepThought was playing and >not deep blue) and this version did a draw against wchess and >lost to fritz3. > >When I see these results I think DeepThought is not better than >today's programs > >Uri Although I don't have the same respect for Deep Thought that Bob Hyatt does, I still have a lot of respect for it. It has beaten my program twice, and both times probably prevented my program from becoming world champion. Two games is a small sample to be sure, but common sense makes me believe this was no fluke. I don't think Deep Thought (or Deep Blue) is invincible and you would see many wins and draws against it but I have no doubt it would still dominate the micro's. Murray estimated his odds of winning that championship at about 50/50. He recognized that Deep Blue would be a heavy favorite in any single game but that dodging 5 bullets is a harder task. I was more optimistic than even they were, thinking Deep Thought had about a 70% chance of winning. As it turned out they drew a game and lost a game and this put them out of it. To me this was not a surprising result, only an unlucky one for them. If you played 100 tournaments with the same exact players, I believe they would win over 70% of these tournaments, but there would be several they did not win. Also I believe that MOST of the field would have won at least one of these 100 touraments. I believe the difference between the very best micro and Deep Thought at this tournament was much less than the difference between the best micro and the worst 10 percent. I think there is a large range of strengths in the programs that show up at these tournaments. - Don
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.