The book should be replete with examples of various historical fallacies.
And bogus data. His first book had a graph that had to be fake - no noise in the signal. Real data doesn't do that.