Author: martin fierz
Date: 07:18:18 08/28/01
Go up one level in this thread
On August 28, 2001 at 02:10:56, Steve Maughan wrote: >How do you calculate the probability distribution of 'A' beating 'B' given their >ELOs? I assume that it's some form of cumulative normal distribution but what >would be the standard deviation? > >All help appreciated! > >Steve actually, you cant calculate the requested probability, only some kind of average result. like 120 elo more is roughly 66% winning probability, but you dont yet know if this will translate into 2 wins and one loss or 1 win and two draws, so you dont have an answer to the exact question you posed. however, if the above is good enough, yes, it is a cumulative normal distribution, with a standard deviation of 200 if i recall that right. at least thats the way the swiss rating list is calculated, but i suppose that number should be that same for all rating lists :-) cheers martin
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.