I don't see why you think the problem implies 50-50 odds. On the contrary, I think the cleveness of the problem lies in conflation of the 50-50 chance of choosing D or 2D, but in the unknown odds of $100 being D or 2D.