Interesting analysis! Though I wonder how easy it is to go fishing for correlation when you can choose the cutoff time. Are the correlations still comparably st

Interesting analysis! Though I wonder how easy it is to go fishing for correlation when you can choose the cutoff time. Are the correlations still comparably strong when the analysis is done with two days or four days? Also I'm not sure how "relative risk" translates to absolute risk. Would a place with already high risk (say one in twenty years) see the doubling of absolute risk (one in ten years) in the days following an earthquake on the other side of the globe? (I know, the answer is probably that there is disagreement over this.) > Since the seismic waves converge at the antipode, this means that regions directly opposite a large mainshock should be on high alert for at least three days, after which the relative risk reduces. A doubling of risk does not change the danger in ways that should let us change our behavior. If it's dangerous at 2 times the rate it's dangerous at the normal rate too. The three days to be spent on "high alert" stand against just six days on "normal alert" with the same risk of a strike.