> One thing that troubles me in the paper is that the researchers appear to have gone looking for precursor patterns in an ad hoc way ... just trying different binning techniques and delays until they got a signal.

That's just p-hacking.