Roger:

Yes...you do have lots (if not most) of the pieces.

I think what you have might need some modifications to process a list of predictions and accumulate stats on them as described.

Also, for "closest" misses that 10% increase (in time, distance, etc.) should find the best fit in the increased 10% range but then drop back to the best fit (quake) percentage increase to get the odds (and not use the full 10% data since it may exceed the best fit quakes range (in time,distance, & maybe mag)).

Of course, if people are afraid to be evaluated then the program may not be used all that much anyway...though it would quickly separate "the talkers" from the real deals ....

The persons who would want to decline such an evaluation obviously have to ask themselves what their real motives are....

.02