Chess Stack Exchange is a question and answer site for serious players and enthusiasts of chess. It only takes a minute to sign up.
Sign up to join this community
I am looking for a historical summary on how the first Elo ratings came about. Apparently in the first official rating list from 1971 there are already ratings in the range of 2200 to 2760. What are these based on? Did they actually run rated tournaments before this date, or did they calculate ratings based on the results in previous tournaments (which tournaments?)?
Also, initially did all players start with the same rating (which)?
In Robert Howard's article "A complete database of international chess players and chess performance ratings for varied longitudinal studies" (Behavior Research Methods 2006, 38 (4), pp. 698-703) he writes:
The first FIDE rating list is dated December 1970 and was based on games in international tournaments in approximately the previous 2 years.
I guess that Arpad Elo will describe this procedure in more detail in his book "The Rating of Chessplayers, Past and Present" that, unfortunately, I do not own.
USCF had a similar rating system in the 1950s. But it was all manual and rating lists came out infrequently. After that a number of people have tried to improve on it. USCF has even changed their version a couple of times.
Arpad Elo was a big factor in improving all the systems and I supect that the FIDE version now is closer to his than USCF as USCF also has to handle all ratings down to 100 which beginning grade school kids may have while FIDE has the easier job of just ranking the very top players whose ability is not rapidly changing.
One early change USCF added was to adjust for kids who were improving fast and older people reluctant to play them because they would lose more points than they could gain if the ratings had correctly matched their abilities.
All rating systems have the problem of correctly estimating the first rating they give a player as well as allowing for new players learning and improving fast.
One thing none of them seem to have done explicitly that I recall is to readjust the average rating every year. I think USCF may have done that once or twice possibly while making changes to how their ratings are computed. The USCF tournament average used to be about 1500. Tournaments tend to remove the very weak beginners from distorting ratings so if they included everybody the average setting might not work as well. But for the top players FIDE could do it IF they had a problem but their ratings for top players tends to stay more stable overall.
