[Skywise]

It's too bad that the article is behind a paywall. It would be interesting to see what statistical methods they used to tease out the pattern, not that I'd be ale to replicate it very well. My statistics knowledge sucks. Although I wonder if some sort of Fourier analysis could be applied. I don't know the math on that either but I am very familiar with it's use in audio and radio frequency spectral analysis.

But I tend to agree with Roger here. Given the short catalog and therefore low numbers of this size quake, it's hard to see how a pattern could be discerned. I mean, there's not even four full cycles in the catalog (2017-1899 = 118 years divided by 32 = 3.6875). Drawing from my astronomical knowledge, I think that's barely the minimum number of orbits required for planetary transits of stars to reliably detect a planet from the dimming of the star as the planet blocks some of the light.

Brian