Michael: Thanks for responding.

"some slight subjectivity"??? Isn't it a 1 or a 0?

I checked into the catalogs a little. I find it very interesting. First, I took all the events from 24.7N to 23.29N and 121.94E to 120.62E. These are the coordinates that are the minimum and maximum from a list of events reported here in Taiwan between 9/20/99 and 10/1/99 inclusive. I already have the Taiwan list on file so it was easy to search. Here in Taiwan, there are 112 events on this list. The USGS returns 157 events and 20 of them have no magnitude.

The USGS seems to be very accurate for events in the US. Even smaller events. No doubt if I do other various searches, they will show variation also. Sometimes more and sometimes less than reported locally. So now, your subjectivity is also effected by the subjectivity of the USGS to report events.

Not only that, but by using your formula, wouldn't an aftershock to a large event be just as probable as an actual temblor? What I mean by that is: say the day of the M7.6 here I predicted an M6 for 48 hours. There were 6 such events reported here (5 by USGS standards.) Wouldn't that be a much more probable event than if I predicted an M6 one week earlier?

Plus - if you use a 30 year record, wouldn't the probability be different if you use a 40 record? In my opinion, this still makes the whole calculation thing just an estimate.

Considering geologic time, if you are in a location that historically has large events, but none has occured in the last 30 years, the probability will be much lower than if one had occured. The 157 events from the USGS list at the beginning of this post are all from a little over a weeks time. If the M7.6 had not ocurred, there may be typically between 2 and 10 events for the same time frame. A pretty big difference in probability even over the 30 years.