Glen: "Regarding eartones, why is it that an eartone for one subduction region be the same as another? The Benioff regions are different, and the processes are clearly defferent."

It would be helpful if you stated in what ways they are different, bearing in mind eartone issues.

"Point being, that if somebody is getting eartones for earthquakes, they would have eartones all the time... Why pick everything from the Kurils to Yakataga to make a prediction instead of China to the Kermandec?"

Glen, you are starting to sound like me, at least to me anyway, that's how I sound. In your initial post, you have hit on an interesting point by implication. A good illustration of "qualitative" evaluation distinguished from purely "quantitative" (which is why I pointedly asked you about the pure statistical quantitative analysis which by admission does not apply to XYZ categories of predictions yet is applied anyway to XYZ predictions anyway because "nothing else is availabe", Roger is bored, and so on):

If an eartone someone could, for example, divide a 1200 mile stretch into 3 or 4 segments, and consistently predict which segment would next have a >5 quake, based on differing geology or subduction process, that would be interesting. There would be a conceptual underpinning.

Roger: "I have explained the Monte Carlo method more than once already."

Goodnes, Roger, you are starting to sound like Shan. How ironic. But you forgot to add "to the extent possible" like Shan does. Live and learn!

Even John didn't quite get what you were talking about with Monte Carlo. For your sake, I sure hope it allows you to evaluate predictions that have varying probabilities, unlike your old Jonesian Model. Otherwise, you had better find a new hobby fast.