Even in the finished challenges code for chapter 13, the two avatar views don’t really seem to behave correctly in a proper rebounding motion. If you look closely, the two avatar views are morphing from the very beginning, long before the two views actually make contact with one another.
It seems more logical that the two views should 1) make contact, 2) deform, 3) spring back into normal dimensions, 4) and finally return to their original locations. The effect would also look better if their internal image views also deformed in the same manner as their corresponding shape layers and mask layers.