I seem to have annoyed some statistics fundamentalists with the last statistics post. So let me spell it out from first principles...