He uses a sample size of one from the 70's to claim theories "really work." Further, his examples of economic theories that work come almost all come from mathematicians/physicists that moved into the field.