He didn't get a false negative, i.e., predicted no earthquake for Haiti. That's very different from no prediction at all. Moreover, if he really does do prediction you'd expect to actually focus on specific areas, and Haiti isn't likely to have be an area of investigation.

No cites of published work. No reference to methodology or research. No error bars given for incredibly precise predictions (54% this, 64% that). The bogo meter is going off loudly here.