I should elaborate on the last item. Efficiency is nice, and you can use JavaProp or the Adkins-Liebeck paper and airfoil data to find the most efficient prop, but there are other considerations. We have found that the CA props (for example) make nicer corners than the APCs, and it is well worth adding another cell to the battery just for the corner. The APC props also make more noise in corners than the CA or Igor props. Some of this seems to be because of the APC's flexibility, but a carbon APC copy still makes noise. This can ding your score in the wind if a judge listens for time between corners, rather than looking at the shape of a maneuver. Some of the noise may come from the changing prop moment of inertia as you're turning a corner. 3-blade props have a lot less variation in moment of inertia in a corner. The 3-bladers also seem to be easier on motor bearings, but bearing life seems to depend on even weirder phenomena. Then there are the issues of yaw caused by turning corners and by operating the prop axis at an angle relative to the oncoming air.
If you do consider prop efficiency, where do you want it? It might be better to have more efficiency in maneuvers than in level flight. Also, given that you can calculate and maximize efficiency, do you want to maximize efficiency itself or the rate of change of efficiency with airspeed?