EQF, for your graphs to mean something you would need to plot ALL M6.5 earthquakes that appeared around the world on them: you could work with Roger to exclude aftershocks.

The Southern California Earthquake Center has been working for several years with international partners on their Collaboratory for Earthquake Predictability. It's not a secret; abstracts are online and papers get published. We are talking "short-term" prediction. They are only interested in Physics-based schemes that can be evaluated. You, and others here, probably have not heard much from this because (although I could be wrong), there has not been success. Shan's does not make logical sense that it could work. I probably don't know enough about your EM signals, but as you have been hearing from me, I have strong doubts.

Chris