Author: Dann Corbit
Date: 15:57:28 02/04/99
Go up one level in this thread
On February 04, 1999 at 18:30:00, Reynolds Takata wrote: >Say that Hiarcs is actually ELO 2550 strength, and he plays GM 2585 ELO. The >2585 defeats H7 in a 6 game match by 3.5 to 2.5. or even less. Well firstly i >would hope that no one would start claiming that the prog isn't GM strength >because it lost. If it's only 2550 it's supposed to lose. If H7 wins, well >that speaks for itself :). I believe also if the comp achieved merely the score >above it would have a relative provisional of 2518, and perhaps that would speak >a little something as well. Winning or losing a single match is not a sure indicator of GM strength. If it achieves a single win and some draws, that would be an indicator that the program is probably very strong. To know if it is of GM strength, it should be subject to the exact same mathematical procedures and scrutiny that a regular GM would be. In other words, we apply the results of these games to the ELO formula. After one serious match, we could give it a provisional rating. After a large number of matches we would be relatively sure how strong the program is. In my opinion, the matches should be against different opponents for the data to be reliable.
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.