Brian;

That speed is impressive but you were answering the wrong question.

The probability should be about the same if you used the same number of predictions over the same time span and same real quakes.

The question was how many hits you get from random predictions with a 3 day window centered on the predicted date. That will vary from zero to as many predictions as you make.

I expected a normal distribution centered on the probability but instead I found a skewed distribution, slowly rising to a peak then rapidly falling.

I'm not enough of a mathematician to know what to do about it.

Roger