Hi all;

Here's a question for the experts in here.

I'm going to write a "do everything" program. It will take a list of predictions, compare each one to all quakes in a file, select the best fitting quake and calculate the probability of a quake of that "goodness of fit".

The question concerns the relative importance of the 3 parameters (time, distance, size).

It seems to me size is most important, preference given to largest. We don't want to accept a small nearby quake when a mag 7 is just a little farther away, do we?

All 3 need practical limits. Can't wait forever or look too far away after all.

This program would work the way Lowell has been doing it and be useful for training. My hit or miss method is best suited to predictions. This new one is for forecasts or for refining a method being developed.

Anyone have any ideas?

Roger