Author: Stephen A. Boak
Date: 11:39:28 03/07/04
Go up one level in this thread
On March 06, 2004 at 15:02:22, Jorge Pichard wrote: >On March 06, 2004 at 13:53:00, Stephen A. Boak wrote: > >>On March 06, 2004 at 12:46:04, Jorge Pichard wrote: >> >>>if a program gets the same percentage, the rating will be lower if the opponents >>>have lower ratings. >>>Jorge >> >>Of course. This is not a testing problem; it is a normal and desirable result >>under the circumstances. >> >>If you score the *same* percentage as another person, but your opponents are >>*weaker* than those who played the other person, why should your performance >>rating be the *same* as the other person, who performed equal percentage >>*against stronger competition*? >> >>Example: You play 6 club players and score 4W - 2L - 0D. Kasparov plays 6 >>world top ten players and scores 4W - 2L - 0D. Who has the better performance >>rating? Whose rating should be higher, yours or Kasparov's? >> > > >Well let me give you my example: If you are rated 2000 and play against 6 >players in Tournament X, all your opponents are rated lower than you by at least >50 points and you score 3W-1L-2D. Then 1 month later your enter another >tournament but your opponents are all rated higher than you by at least 50 >points even if you score this time W3-L2-1D you still will gain points, simply >because of the average rating of your opponents. Incorrect! You never gain points 'simply because of the average rating of your opponents.' You gain points only when your performance is better than expected, based on the relative initial ELOs of the opponents. Your above comments ignore that your performance itself is much better in the second scenario. --Steve > >Jorge
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.