Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: using bayselo versus elostat

Author: Dann Corbit

Date: 20:05:42 02/03/06

Go up one level in this thread


On February 03, 2006 at 22:53:56, Joseph Ciarrochi wrote:

>bayeselo generates numbers that are quite different, in scale, from those
>generated by elostat (e.g, cegt ratings). Is their anyway to make the scale the
>same?

Yes.  There is a help system for that program:

E:\programme\winboard>bayeselo
version 0053, Copyright (C) 1997-2005 Remi Coulom.
compiled Jan 26 2006 12:49:16.
This program comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it
under the terms and conditions of the GNU General Public License.
See http://www.gnu.org/copyleft/gpl.html for details.
ResultSet>?

Syntax for command lines
~~~~~~~~~~~~~~~~~~~~~~~~
_command_ _parameters_ [< _input_] [>[>] _output_] [;_comment_]

Universal commands
~~~~~~~~~~~~~~~~~~
x ............... close this console interface
p ............... go to the parent interface
children ........ list children interfaces
child <n> ....... select a child interface

echo <params> ... repeat <params> to the output stream
! [<command>] ... make a system call
prompt [on|off] . show|hide the command prompt
myself <n> ...... call recursively this interface <n> times

ResultSet commands
~~~~~~~~~~~~~~~~~~
players ......... list players
results ......... list game results
pgn ............. write results in PGN format

reset ........... reset results and players
addplayer <pn> .. add one player with name <pn>
addresult w b r . add result (w = white, b = black, r = result)
                   r = 0 (b wins), 1 (draw), or 2 (w wins)
readpgn <file>... read PGN file
connect [p] [fr]  remove players that are not connected to p

elo ............. open Elo-estimation interface

ResultSet>readpgn book.pgn
26085 game(s) loaded, 10 game(s) with unknown result ignored.
ResultSet>elo
ResultSet-EloRating>?

Syntax for command lines
~~~~~~~~~~~~~~~~~~~~~~~~
_command_ _parameters_ [< _input_] [>[>] _output_] [;_comment_]

Universal commands
~~~~~~~~~~~~~~~~~~
x ............... close this console interface
p ............... go to the parent interface
children ........ list children interfaces
child <n> ....... select a child interface

echo <params> ... repeat <params> to the output stream
! [<command>] ... make a system call
prompt [on|off] . show|hide the command prompt
myself <n> ...... call recursively this interface <n> times

EloRating commands
~~~~~~~~~~~~~~~~~~
advantage [x] ... get[set] advantage of playing first
drawelo [x] ..... get[set] draw Elo
prior [x] ....... get[set] prior (= number of virtual draws)
elo [p] [elo] ... get[set] Elo of player number p
mm [a] [d] ...... compute maximum-likelihood Elos:
                   a: flag to compute advantage (default = 0)
                   d: flag to compute elodraw (default = 0)
elostat ......... compute ratings with ELOstat algorithm

ratings [min] ... list players with at least min games
details ......... detailed list
offset [elo] .... get[set] Elo offset
scale ........... get[set] Elo scale

p <w> <b> <r> ... probability of result <r> with Elos <w> and <b>
likelihood ...... log-likelihood of the current Elo list
prediction ...... open prediction interface

plotres [i] [M] . plot result likelihoods as a function of rating diff:
                   i: number of intervals (default = 21)
                   M: maximum rating difference (default = 500)
plotdraw [n] [d]  plot draw frequency as a function of average rating
                   n: number of games per group (default = 100)
                   d: maximum rating difference (default = 100)

confidence ...... get[set] level of confidence intervals
variance ........ compute intervals with the diagonal of the Hessian
covariance ...... compute intervals with the full Hessian
los [f] [p] [w] . likelihood of superiority (f=first,p=players,w=width)

minelo [x] ...... get[set] minimum Elo
maxelo [x] ...... get[set] maximum Elo
resolution [n] .. get[set] resolution
jointdist [p] ... compute intervals from joint distribution
exactdist [p] ... compute intervals assuming exact opponent Elos
advdist ......... likelihood distribution of advantage
drawdist ........ likelihood distribution of drawelo

ResultSet-EloRating>


>Also, has anybody done a comparison of the two rating systems (bayelo versus
>elostat). Are their big differences?

For the most part the differences can be ignored.  It appears to me that
BayesElo is probably more accurate for the extreme cases.  But (for instance) if
you take WBEC premier division and analyze with both systems, I guess that both
will have the same ranking.  So the differences are usually not important.

You can do more with BayesElo, but it is a little more complicated.



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.