It seems simple deduction to me that, regardless of the pattern of quake prone regions, that there will be areas which will be a specific number of degrees from other quake prone regions.

Try doing the same test, but at various other distances. For example, 42 degrees, 127, and 154. (just pulled those out of thin air)

If it's not too much work, graph out how many correlations there are at all degree ranges 1-180. I'd expect there to be a big peak at low values (too close) but the distribution across the graph might be interesting. I'd be curious at what distances the most correlations are, and whether 103 is the optimal distance for such correlations.

And finally, does it mean anything?

Brian