Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Statistical data about draws and rating differences

Author: Mike S.

Date: 19:12:54 02/04/01

Go up one level in this thread


On February 04, 2001 at 18:54:47, Ralf Elvsén wrote:

>>http://math.bu.edu/INDIVIDUAL/mg/ratings.html
>
>There is a link to a graph showing the draw frequency for players
>within 100 rating points as a function of their rating. Interesting...

I have done a similar analysis, using the Big Database '99, to see the
percentage of draws when BOTH players were within a certain rating range of 100
elo each. The games were from 1990...1998:

rating range #games %draws
--------------------------
     - 1800 151000   27
1801 - 1900    603   35
1901 - 2000   1319   38
2001 - 2100   4607   35
2101 - 2200  12477   36
2201 - 2300  28635   39
2301 - 2400  29498   48
2401 - 2500  25970   54
2501 - 2600  10567   56
2601 -        2277   53

Looks like the "many draws" start at around 2300 elo.

What may be more interesting is, that the number of black wins is "shrinking"
(so to say) more, than the number of white wins:

-1800 : 1-0 40%, 0-1 33%
2601+ : 1-0 30% (-10), 0-1 17% (-16)

Additionally, I took a look at (a smaller number of) games, where "rating
groups" - across 200 points this time - played against each other, i.e. players
within 1900...2100 against players with 2300...2500 elo. The results were not
surprising: The draw percentage was highest within the same ratings group, and
lower against stronger or weaker players, for all groups.

Regards,
M.Scheidl



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.