This sort of thing is exactly why I don't think the argument in question is correct and why I'm comfortable with my own Pr(God) being orders of magnitude smaller than the fraction of theists in the educated population.

However, simplicio is right that by taking this sort of view one becomes more vulnerable to closed-mindedness. The price of being more confident when right is being less persuadable when wrong. I think simplicio's second paragraph has it pretty much exactly right: in cases where you're disagreeing starkly with a lot of smart people, don't adjust your probabilities, adjust your behaviour and give the improbable hypothesis more consideration and more time than your current estimate of its probability would justify on its own.