Is it possible that the dynamics of certain sections of a fault generally don't allow all of the built up strain to be released? To clarify, obviously small quakes don't release all the strain. And large quakes may release nearly all of it. But perhaps some faults have quakes that only release a smaller amount.

Then, when a big quake comes at it from one end, basically like a hammer hitting it, it causes what strain is there that on it's own is insufficient to cause a quake to release from sheer momentum of the rupture. I would imagine the hammer has to be big enough, though.

Alas, we just don't know quite enough yet. What starts the quake, and what makes it stop. I think one of the big unanswered questions is what's the difference in the starting of a M2 and M7, if anything.

Brian