Yep. Took a stab at it here based on data conditions a few months ago. The posting is here. Data conditions changed shortly after I posted (EQ caused a strain release), but the current conditions are greater than they were when I posted. Also in the last 30 days, the X461 creepmeter recorded movement not related to an EQ (at least that USGS hasn't posted).
I have been observing geomagnetic spiking occurring appx. 12 hours (+/- 2 hours generally) prior to an event. I first saw this in the Montana EQ. The same was seen prior to the Willits EQ, and several california and Alaskan EQ's. Due to the fact that I read USGS data from 4 large regional sites, narrowing things down to a small specific area is difficult.
Generally, the spike looks like a nail in an otherwise smooth (per se) chart.

As for the other physical basis, the tensor strainmeters, both the 30 day andthe 7 day at Parkfield are showing an increase in strain larger than it was in October.
So, I am patiently waiting, with my mouth closed (figuratively speaking), and I am going to see where this goes.
If I get a sufficient spike out of Fresno without effecting the other Geomag sites, followed by a decent sized EQ (3.5+), then the theory is confirmed.

I am saying that I agree that EQ's come from all sorts of sources and mechanisms. Each technique will detect a certain type of motion.
Unfortunately, there is a quest for one single method to do this, but like the fasteners of a machine, one size does not fit all.

As for Shan. I respect his results and I respect his methods. Does he have 100% accuracy, no. Does he detect certain types of data that accurately corrolate with certain results, absolutely.

We all have one piece of the puzzle, and one single method is not going to work 100% of the time. I am convinced that overlapping methods from different viewpoints will produce results.
If cooperation and the free exchange of information results in lives being saved, and damage being prevented, I'm all for it.