USCF ID 12590577 my rating and 6 points!?

[nolan]

There are two considerations for an article in the magazine.

One is the editorial schedule. The April issue was blocked out weeks ago and pretty much locked down two weeks ago.

The second is that someone who can write well should do the article, not me. I’m too technical and too close to the changes being made to write about it without either being defensive or biased.

[rfeditor]

@Smythe Dakota

With the new rating software and procedures, I think everyone has to get over the notion that they have a God-given right to their X number of rating points just because they saw that rating once on a crosstable.

Re-rating is an extremely GOOD idea, but it is necessary for everyone to accept all of the consequences.

There has always been, and will always be, considerable upward political pressure on rating points. It is absolutely essential for USCF to resist this pressure, and to worry about accuracy instead. Otherwise, ratings will become so inflated that each individual rating will become meaningless.

Caving in to pressure from players who want higher ratings is exactly what caused drastic rating inflation in the mid-1970s. By now, this inflation has largely been corrected, although USCF ratings are still about 50 points higher than FIDE ratings.

Bill Smythe

Asserting that re-rates are a good idea is not going to make it so. As far as I can tell, no one has made any real effort to justify the project (“because we can” is not a good enough reason).

One of the implied arguments seems to be that the ratings would thus be more “accurate.” This is silly. There is a big difference between “precision” and “accuracy.” A change of ~10 rating points, which is probably the practical max, has no predictive value.

Another argument seems to be that the DB will be “cleaner” after the re-rates. I’m sure this is very interesting to the programmers. It is of zero interest to players, who are concerned only with the output.

My opinion? Re-rates should be confined exclusively to cases of serious error, and even then there should be a time limit. Each rating supplement should be treated as a restore point, and the ratings worked from there. (The reply to this will no doubt be to cite fraudulent tournaments. My answer: This should be dealt by individual corrections, not with the sledgehammer of mass re-rates.)