Yes he did. I'm just trying to be more proactive to generate more concrete evidence, or lack thereof, of prediction credibility, and to learn about prediction credibility myself of course.

Windowing is basically what I did. I am, however, concerned with the comment that Alan was using windowing on one hand to reduce the impact of aftershocks on the statistics (I'm not sure how), but on the other hand, I assume at the same time, using the Jones/Reasonburgh formulae to work aftershocks into the historical model. My guess is there is more to it, and it would be interesting to hear his comments.

Maybe I should assemble requests for Alan to be proactive, and let him do the calculations? Is anybody interested in that sort of analysis of their predictions?