Tim,

Your claim that there would be a 6+ somewhere in the state on a day on which a 6 DID hit is quite unlikely at random. The fact remains it is hard to explain why your prediction should work, although others here apply similar techniques. If it works there must be a way it can work, which so far is a problem.

A small correction - the recent prediction was for a 6.4+ quake in a period ENDING Sept 5, not just one on that day. In any case, it missed. Another somewhat-off prediction was the USGS one for the Parkfield event to hit by 1990. Oops. So perhaps noone is outperforming you around here.

John