Don Hitchcock provided me with the following prediction by Jim Berkland (I can't seem to reach Berkland's web site):

For the October 13-20, 2000 Seismic Window I predict with 75% confidence: (1) a 3.5-5.5M quake within 140 miles of San Francisco; (2) a 3.5-6.0M quake within 140 miles of Los Angeles; a(3) a 3.0-5.0M quake in Washington and/or Oregon; and (4) a major world quake (probably in the Pacific Ring of Fire, where 80% of the 7+ magnitude quakes occur.)

I evaluate these four predictions to have the following probability of happening by chance based on catalogs:

1) 23.4%, 2) 34.5%, 3) 14.3%, 4) 25.3% for a combined probability of 68% which is about the same as Berkland claims. In other words, as far as I can tell, he is giving himself a 70% chance of succeeding when that is the probability of succeeding by chance. To use a term from probability, his prediction is the same as one made with "no skill." To make matters worse, his prediction of a 3.0 or greater in Oregon or Washington has a much higher probability of succeeding than the 14% I give it since my catalogs for Washington and Oregon only go down to events of magnitude 4.0. I just don't have the data to estimate a WA or OR event this small.

In any case, let's wait and see what happens when the window closes on the 20th.