igor comments on Expecting Short Inferential Distances