I have never understood adding the percentage of happening to a prediction. There are times when my symptoms are so severe that I am almost certain about my prediction. There are also times when my symptoms are vague....it is the later times when I am not sure whether to post the prediction and the prior when I am certain to do it...I suppose that is my way of expressing 'percentage' chance of happening.
As to your experiment. It reminds me of the gang of physicists who were dtermined to add every possibility to a roulette wheel...ie the speed of the ball thrown, the speed of the wheel...the point of release of the ball etc...As I remember they felt they could predict 100% were they able to figure every variable but they became bogged down by humidity, friction...and the countless numbers of actions such as wind cause by someone passing the table...the list was endless and the group gave up after a bit...although they did have enough stats to call the lay of the ball over 50% of the time.
If you start out with the stats on your database, then its better than having the opposites team playbook. You already know the odds of certain quakes. In fact you could make our predictions based on physical symptoms look feable. I have to admit that this mever occurred to me...
I think when using physical symptoms, one takes a much greater risk in posting a prediction. Now because I have moved out of an active area, I not only use physical symptoms but also read the Garlock list to "see" if there is a pattern developing.
My guess is that your experiment is going to work VERY well....as a matter of fact I'm 80 percent sure it will!

Bob