If you have a polished tungsten carbide ball accelerated by uniformly compressed air out of a polished barrel into vacuum, I'm pretty sure the stabilized and unstabilized balls perform identically. In a real world scenario you are having a lot of imperfections and non-uniformities to deal with, and you need to average out these effects by rotation.
As for the golf ball dimples, the reason a golf ball performs better with dimples than without (by about a factor of 2 - 3 distance wise) is due to the speed regime. Usually, a sphere will have nice laminar flow around it right up to midpoint, where the laminar flow leaves the surface, and the resulting turbulence generated drag. The little dimples on the golf ball produce a small layer of turbulent flow on the surface (little more drag) but allow the laminar flow to stay close to the ball for most of its length (much less drag). Unfortunately that only works in the 150 - 250 ft/s speed range, too slow for bullets to be useful.
F 135 - the right choice