Roger:

Nicely done & nice of you to chase that down.

With that answer, it seems more likely that maybe Lowell had data suggesting the probability might go up after a large flare for maybe larger sized quakes like M6.5+ or maybe M7+...over and above what would normally be expected.

I think the way it would be done is to do just what you did for say M7+ quakes and then separate the cases when there was a big flare (5 day windows after a flare) from those cases where there was no flare to see if the rates of occurence (based on observed rate) is any different.

I think flare data only goes back to '97 or '98 (I'm guessing) so the data you use should only go back that far as well (and not back to '90 or whatever since it would be unknown whether the time window periods are "flare windows" or "Not flare window cases" (apples 'n oranges mixed).

Anyway, I believe Lowell must have already done something like this or may know someone who has.