A question for the arm-chair analysts in this thread: which one has a better history of not being wrong, excluding this data point?