Rate-and-state theory predicts a nucleation phase lasting weeks or longer. Our precise measurement of the lack of correlation between earthquakes and tides in California supports rate-and-state theories. Therefore our result gives some hope to those who think earthquakes might be predictable days or weeks ahead of time.

Which part of this chain of logic do you not understand? Or are you claiming comparison of observations of the time of earthquakes with the tides at those times to test a specific model of the nucleation of earthquakes is "not a hypothesis, let alone a theory"?

This is the kind of data-theory comparison by which science progresses.