Author: Chris Carson
Date: 03:26:03 06/02/02
Go up one level in this thread
On June 01, 2002 at 23:10:23, Ricardo Gibert wrote: >On June 01, 2002 at 12:31:22, Chris Carson wrote: > >>The Rating of Chessplayers, Past and Present >>by Arpad E. Elo, 1978 >> >>http://www.amazon.com/exec/obidos/ASIN/0668047216/qid=1022948840/sr=1-1/ref=sr_1_1/103-2617820-3864631 >> >> >>Here is an overview of what is in ELO's book: >> >>"This work was written in 1978 by the Old Man himself, Arpad Elo. Chess players >>remember Elo as the cantankerous professor of physics from Marquette University, >>and as the inventor of the rating systems used by the United States Chess >>Federation (USCF) and the International Chess Federation (FIDE). Who better to >>explain the mathematics behind the rating system which, according to a letter >>Elo wrote to Chess Life several years ago, "is, after all, MY system". The >>explanations are semi-technical, but understandable by anybody with a >>mathematical inclination, regardless of their education in that field. There is >>even a chapter in which ratings are calculated retroactively, for grandmasters >>of bygone days. The book was written B.K. (before Kasparov), so of course Bobby >>Fischer comes out on top, with a rating of about 2780. His closest competitors >>are Lasker, Capablanca, and Botvinnik, each of whom peaked at about 2720. The >>average rating of tournament players in the U.S., by the way, is about 1500, >>several classes below the stars. Interesting reading." >> >>If you will note, Elo himself compared Fisher with Lasker, Capablanca, and >>Botvinnik. > >This is well known about Elo's work. What is also well known is it was and is >faulty. In other words, his comparisons between different time periods does not >have a sound statistical foundation, so quoting "Elo" does not help you make >your case. First, you were quoting Elo and asked for a citation, so I gave your citation, it does not support your view. Yes, I have read his book and understand the flaws, not sure you do. His big flaw was to compare people that never had a rating, he did this by manufacturing a rating based on common opponents from the actual pool. There may be some validity to this, but not stat sound. However, FIDE is a common pool, these limitations do not exists and a ratings comparison is valid, but not the only possible comparison. Other factors can be considered, but ratings are one valid factor. It is clear we disagree, however, I doubt you can agree to disagree.
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.