this third to last paragraph offers a summary:

NASA’s ‘Remarkably Unremarkable’ Success Rate

On the basis of the analysis presented above, the success rate referred to in the NASA announcement is perhaps better characterized as ‘remarkably unremarkable’ rather than amazing. The simple analysis presented here suggests that the RT method does not predict the locations of future earthquakes any better than merely assuming that future earthquakes occur near past earthquakes. There may be something more significant to the RT method than just proximity to past earthquakes, and this analysis
does not demonstrate that there is not. However, the amazing success of predicting 14 of 15 of the earthquakes listed in the scorecard is not a very rigorous test of their method.