Interesting question. From time to time people wonder whether widely-spaced earthquakes are more clumped in time than random.

The general pattern is that with human intelligence, a "pattern" is spotted, which proves highly significant, that is, unlikely to happen at random (which may be why the human eye picked it out in the first place). Then a new or broader dataset is analyzed, and the pattern disappears.

I think that is the case for moderately large, widely-spaced earthquakes. A decade ago, scientists thought permanent stresses were necessary to cause earthquakes, now shaking alone without permanent stress is also recognized as a trigger, but it is very inefficient.

All known earthquake triggers also are proportional (or non-linearly proportional) to the stress or shaking, so any effect that does not uniformly diminish with distance is suspect. An M8 might slightly raise the rate of M6s around it, but M6 and M7s probably have little effect beyond their immediate neighborhood.

The effects are so weak that only a few cases of even regional triggering have been detected - 2002 Denali triggering west coast seismicity and tremor, 1992 Landers triggering Rocky Mountain quakes, maybe another one or two.