While its great that Graham has said all this, and it shows a lot of professionalism, it was probably not really necessary. After all, its a book, not a research paper. The point is, it gets more expensive to fix a bug later on. We aren't going to shoot down the book because some numbers, only tangentially related to the subject (of learning TDD for iOS), were off. Still, kudos for blogging it when he came across new information.