Author: Gian-Carlo Pascutto
Date: 06:16:50 06/11/01
Go up one level in this thread
On June 11, 2001 at 06:08:58, Peter Fendrich wrote: >I've made a program like that but it's based on the trinomial distribution >which should be more accurate for chess results. The draw case makes a >difference. Yes. I cross-checked with your utility and in case of no draws the numbers are correct (well, the last 2 add up to 202 games of course :) When draws come into play the 'isbetter' probability is always higher (empirically tested), so the numbers I gave are ok to use as a 'minimum'. >The purpose of the program is to be able of say something about match results >with a small number of games if something is possible to say. Small numbers are >small numbers and there are no magic formulas to overcome that! Games with only >one opponent is biased against only that opponent. No magic formulas for that >either! This is just life... Yes. My main interest is to demonstrate that most match results posted here mean nothing. >Anyone interested can get the program from me with a "popular" explanation >about the used theories. Thanks. I will study it as soon as I have time. -- GCP
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.