Ramblings and ruminations on chess in SE Wisconsin, the USA and the World

Arpad Elo

It was near the end of the month of July in 1935 in the Schroeder Hotel in Milwaukee, Wisconsin. The Western Chess Association had recently renamed itself the American Chess Federation, and this was its annual championship tournament. Held every year since 1899, it had been called the Western Open, but the previous year it was given a new name, in honor of the new name of the organization: The American Open. The reason for the new name was that there really was no national chess organization worthy of the name. There was the National Chess Federation, but about the only thing it did was organize the Olympiad team every year. The members of the Western Chess Association felt it was time for an active national body.

This year it would all change. At the business meeting for the ACF this year they would elect new officers, and take on a more national role. The tournament name change reflected the fact that it was the national tournament championship. A few years later they would merge with the NCF to yield one national organizing body for chess. The business meeting was starting, but one of the delegates (the current Wisconsin champion) was late; his game (against Reuben Fine) was running long. Fine had made a mistake early in the game and, with his back to the wall, was playing on in hopes of a perpetual check. Finally, the game was declared drawn when the position repeated for the third time (Fine was down two pawns in a Queen ending) and Arpad Elo made his delayed entrance into the business meeting, to discover that in his absence he had been elected president of the ACF. As ACF president, he started a membership drive which in the end netted about 300 members. (Up until this point, the Western Chess Association was more like a country club than anything else, with a few members meeting every year at a resort to play chess. There was not much of a membership list for either the ACF/WCA or the NCF. Elo wanted more of a city club than a country club, where everyone could join.)

Chicagoan Kirk Holland, who succeeded Elo in 1937, would later (1939) chair the meeting on the merger between the ACF and the NCF, and when he finally banged the gavel at the end of that meeting, it would be to announce the birth of the United States Chess Federation. Elo was one of the charter members of the directorate of the fledgling organization.

Later, in 1959, Jerry Spann (then USCF President) asked Elo to overhaul the USCF rating system. Elo noticed some anomalies in the system used by the USCF to rate players. Among other things it was possible to lose every game while gaining rating points (and, similarly, lose rating points while winning every game). While the system provided for some variability in the performance of a chessplayer, it assumed that all possible levels of a player’s performance were equally probable, that is, that it was equally probable a player of a given rating would play at that level or several hundred points above it. He proposed a different equation, based upon a bell curve, which would more closely model reality. (Ratings trivia: The current ratings go to four digits simply because that was what the earlier formula did and Elo didn’t want to confuse people by suddenly supplying substantially different numbers. The old system was predicated on the idea that a strong club player would be 2000, so Elo built his formula to reflect that. Ratings are not and have never been accurate to four figures.) Later, he went on to calculate ratings for FIDE using his formula by hand on an HP calculator.

To the end of his life he remained a little chagrined by what the chess world made of his formula. (He once said, “The Elo ratings are fought for and strived for until it becomes really amusing. Sometimes I think I’ve created a Frankenstein’s Monster! Young players are really more interested in what they do rating-wise than what they do across the board.”) What he had intended was simply to create a metric to gauge performance. Those who have read Deming realize that once you create a performance metric, there is a great tendency to emphasize the metric over the act itself. (In the world of manufacturing, if you create a parts per minute metric, the tendency for both workers and managers is to shift the focus to the number of parts per minute, rather than keep the focus on creating the parts themselves.) It bothered him that organizers and players both began to treat ratings as more important than playing good chess.

If the rating system was a bittersweet legacy, it still didn’t sour him on chess. He was involved in promoting chess in the Milwaukee County parks system, with attendance at scholastic tournaments numbering in the thousands in its heyday, as well as youth and industrial league competition for teams. He was chairman of the commitee which organized the 1935 American Open, which attracted more players than in the history of that tournament up to that time.