It should be easy to evaluate all of the APD's and find those who's failure rate falls outside of the standard median.

I'm sure they would refrain from pulling shenanigans when they are being observed, but maybe a better debriefing of failed pilots could be done by a third party to figure out why people are failing and if there are things that training could focus on, or to allow pilots to express the perceived unfairness they received without being further penalized.