Author: Peter Kappler
Date: 22:08:10 01/11/00
Go up one level in this thread
On January 11, 2000 at 21:41:01, Leon Stancliff wrote: > In my opinion the question as to whether the top microcomputer programs can >play at grandmaster level has been answered. The average rating for the >opponents of Rebel Century in the twenty games actually played was 2510 Elo. The >average Mhz rating of the computers was 498. Rebel won by 11-9 which gave it >55%. I think something is wrong with your list of games. There have been 21 games played at 40/2, with Rebel scoring 5 wins, 10 draws, and 6 losses. Here's a complete list: Player Elo Rebel Hardware Sorin 2524 0.5 K6/3-600 Rohde 2585 0 K6/3-600 Rohde 2585 0.5 K6/3-600 Hoffman 2501 0 K6/3-600 Anand 2795 0.5 K6/2-450 Anand 2795 0 K6/2-450 Ruzele 2515 0.5 K6/3-600 Krivonosov 2432 0.5 P2-400 Zapolskis 2399 1 K6/2-300 Grabliauskas 2466 0.5 P2-350 Sherbakov 2541 1 K6/3-600 Zagorskis 2486 1 K6/3-600 Grabliauskas 2466 0 P3-500 Sakalauskas 2450 0 P2-350 Krivonosov 2432 0.5 K6/2-300 Baburin 2593 0.5 K6/3-600 Sherbakov 2552 0 K6/3-600 Russek 2398 1 K6/3-450 Russek 2398 0.5 K6/3-450 Russek 2398 0.5 K6/3-450 Russek 2398 1 K6/3-450 By my numbers that's a 2491 performance rating. Certainly a strong IM-level performance, but not quite GM strength, and well short of what would be necessary to actually achieve a GM norm. Of course, Rebel wasn't playing on the fastest currently available hardware, but at 40/2 it doesn't make a huge difference. How much better it would score on an Athlon 750 is another debate entirely - but I seriously doubt it would even amount to another 50 ELO points if you're talking about 40/2 games against humans. > The point differential for 55% wins is 35. When we add 35 points to 2510 we >have 2545 Elo. Rebel Century does play at grandmaster level. The question now >will become, "Where does the top micro program rate among the world list of >grandmasters?" > An Elo rating of 2545 would place Rebel Century at the 219th position out of >652 living grandmasters. This places the program in the top third of the GM >list. The only real live question now is "How long will it be before the top >computers will among the top one hundred?" Will it happen within three years? To me, it's more interesting to ask how long before it can crack the top 10, which is roughly 2700 FIDE. I think it might still be a very long time. --Peter
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.