I agree that the formula is wrong. I considered the possibility that it was sarcasm, but either way it's wrong.
Either he's fixed it since the two of you wrote what you wrote, or you're wrong, or I'm terribly confused. It now says (with obvious abbreviations)
P(A|B) = P(B|A)P(A)/P(B)
which is correct.
A fun comic about seashells and Bayes' Theorem. http://xkcd.com/1236/