You make a very strong case that the Nazi example does go in the other direction. I withdraw that example. If anything it goes strongly in favor of Deutsch's point.
I'm not convinced by the relevancy of your point about the historical state during the colonization of North America. The point is not whether or not someone eventually transformed, the point is that violent, expansionist groups can win over less expansionist groups.
Deutsch's definition of "the bad guys" is not the most expansionist groups.
He would regard the colonizers as the good guys (well, better guys) because their society was less static, more open to improvement, more tolerant of non-conformist people, more tolerant of new ideas, more free, and so on. There's a reason the natives had worse technology and their culture remained static for so long: they had a society that squashes innovation.
http://vimeo.com/22099396
What do people think of this, from a Bayesian perspective?
It is a talk given to the Oxford Transhumanists. Their previous speaker was Eliezer Yudkowsky. Audio version and past talks here: http://groupspaces.com/oxfordtranshumanists/pages/past-talks