Author: Sune Fischer
Date: 10:19:18 01/17/03
Go up one level in this thread
On January 17, 2003 at 12:32:50, Uri Blass wrote: >>Yes but humans also retire and stop playing at some point, that loses rating >>from the pool so the average on a large scale is pretty consistent. > >it is dependent on the case. >You cannot assume that it is constant. >There are cases when it is obviously go up. > >If you take a team of beginners then you can expect their rating to go up. >rating can go down usually only with old players or with players that do not >play for a long time and the information about the age of the player and about >the time that they did not play can be also used for defining better rating >system. Yes, and that is why I said "large scale", like FIDE or ICC. In a small group the assumptions of statistical means are going to be dangerous. If they all play eachother in a small group, they might all get better but ratings don't change. I don't see that as a fault of the Elo system, if there is no correlation to the outside world you can only measure their relative strength, obviously. >>>You can reduce rating from everyone and keep the average the same. >> >>This is what is done, the winner goes up, loser goes down. If that means the >>loser has now become underrated he can regain the rating by playing other >>players and thus diffuse the loss to the rest of the pool. > >Not exactly. >The loser lose and the winner win but the rating of other is not changed and I >gave the idea to reduce the rating of other players by a very small number so >the average will be the same. I'm not sure I understand then, you add Elo to the winner and subtract from _all_ others? Why not subtract from the loser and add to all the others? >I think that I am going to stop to respond today because if I continue to >respond I will have no chance to implement king safety in Movei for the CCT5. Me too, lots of last minute untested changes. Can't go wrong :) I'm just as worried about my book, what kind of insane suicide openings it will spew out... Oh well, life goes on on monday :) -S. >I am also not optimitic about it even if I stop to respond. > >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.