Comment by is74

Actually, it does, since the difference in performance between entry #1 and entry #2 is so huge (25% error vs 15% error!), and since this is by far the hardest computer vision challenge yet!