I think about this a bit. I had the temerity to try and write to Dawkins about this some time ago. I think I got a sense of "yes, given the chance, I think use of more statistical language would have been beneficial, but none of us realized how strongly the intentionality would be misunderstood" in reply, but I can't find the emails so I suspect I may have mis-remembered.
Really I'd like to ask it face to face. For the experience.