We provide rankings across uncertain outputs generated by agents functioning within the Principal-Agent paradigm. For agents who are identical except for their productivity, a necessary (but not sufficient) condition for an agent to be preferred is that her output dominates that of lower agents...