> Hi Roger. I must admit that I really and truly don’t understand how your program evaluates a prediction. I know of at least 70 of my predictions were all met in regards to the window of time, location and radius and the magnitude range.

Were they in this group of 226? If so, point them out so I can see what's up.

> I thought that in order for the prediction to be a good prediction all 3 of the parameters (time, location, magnitude) had to be met which they were.

That's correct which is why I'm concerned.

> The other thing that is curious is that I got perfect hits on the more difficult predictions as opposed to the ones that weren’t difficult. Why is that?

Beats me but that's why I said you did well.

> I know from talking to various seismologists that the location of any quake is always suspected. The location can be off as much as 5 miles and magnitude off by 0.3. It’s not perfect so why does the person making the prediction have to be perfect? I’m not even sure I understand what perfect is in regards to your program.

"Perfect" means fitting whatever you said in all respects.

> Today’s M=1.5 near Rohnert Park even though it is a miss in all respects told me a lot about the data I use. Kind of interesting in what one can learn from failure.

It's usually a good teacher!

Roger