yes, and I see better now what you're saying. You want to divide the time into unit blocks and overlay the windows on that, redefining the limits each pass. But that still messes up the number of windows or the size of windows at each end just as my drawing shows.

A quake at the starting end will be in too few windows, making the probability too low. The other end will have less of a problem but still won't be correct.

Roger