Author: Uri Blass
Date: 00:48:51 12/20/00
Go up one level in this thread
On December 20, 2000 at 02:54:17, David Rasmussen wrote: >On December 19, 2000 at 12:08:16, Christophe Theron wrote: > >> >>---------------------------------------------- >>Reliability of chess matches (confidence: 80%) >> >> 10 games: 14.0% (105 pts) >> 20 games: 11.0% ( 77 pts) >> 30 games: 9.0% ( 63 pts) >> 40 games: 8.0% ( 56 pts) >> 50 games: 7.0% ( 49 pts) >>100 games: 5.0% ( 35 pts) >>200 games: 3.5% ( 25 pts) >>400 games: 2.5% ( 18 pts) >>600 games: 2.2% ( 15 pts) >>---------------------------------------------- >> > >If this has to be of any use, you have to have at least a 95% level of >significance. 80% is not very useful, IMO. > >The Insane Dane 80% is useful. If you want to know which program to buy getting information that program A is better than B with 80% confidence is better than no information. The main problem with all this information is that there is no clear definition if program A is better than B. Suppose A get 60% against B B gets 60% against C and C gets 60% against A Which program is better A,B or C? If a lot of programs are similiar to C then B will be the best in the ssdf list and if a lot of programs are similiar to B then A will be the best in the ssdf list. Uri
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.