Just call me Pete. You could test reliability, that is the ability to get consistent measurements (or ratings), although in this discussion the term is used more as the common language term which would be something like "useful and accurate." Validity along with reliability are the statistical measures that would be appropriate for that. You could test the reliability of lets say the measurement of Hart's defense. You could use an independent video set-up, in particular one that has a constant view and compare against other fielders in RF in Miller Park. You could also reconstruct the performances using the videos and measuring time and distance of ball and fielders. Another almost essential test which might be done, but I wouldn't be surprised if it wasn't, is testing reliability between evaluators. Just have different guys look at the same videos and see how consistent they are. A player who gets high ratings on defensive metrics but who I see often make really bad mistakes is Colby Rasmus. If he is being judged by some inexperienced young person who is also a Cards fan or who is "in to" the Fangraphs crowd where he is adored, the results may be worthless. You are left with the ultimate question of what is good which is what you seemed to be saying. All I was really doing was challenging the notion that UZR is reliable.