This strikes me as a variant of the Texas Sharpshooter fallacy. Shoot blindly at a wall a bunch of times, then go down and draw targets around your bulletholes. What Watts is doing here is similar: make a whole bunch of predictions well away from mainstream, ignore all the ones that failed, then trumpet from the rooftops when we hit an outlier year that happens to match one. In actual fact, he has a long, (externally) documented history of being spectacularly wrong with the vast majority of his predictions.
Incidentally, the makers of that graph were being very diplomatic when they chose the term "heuristic" for his method.
