I believe friction does have to do with surface area, and in a big way.
I'll put this in ironworker terms, because thats what I am. Take a 15/16" Barrel Pin for instance (pointed at both ends, but straight 15/16" in the midsection) and drive it through a 15/16th" hole in a 2" plate. The pointed part goes in easy, and as you get into the area where the hole is the size of the pin, it gets tighter. The more the pin goes into the plate, the tighter it gets, until the straight section is out the other side of the plate. Thats because as you drive it in, you have more and more of the surface of the pin contacting the hole, until you reach the other side. At this point you have total contact, and there is no way more surface area contact can occur. As you continue to drive it through you continue at the same tension, until the rear taper is now entering the hole, now with every drive you have less and less surface area contact and the pin gets easier to drive. This is because the constant friction is now applied to less surface area.
The idea of ball bearings, roller bearings, etc. is the same. The ball only contacts a surface at the tangent point, so friction is only applied at that point, a bushing used in a similar application however has much more surface contact and would build much more friction and heat.
So cutting down on the surface area by cutting flutes in a bolt reduces the amount of surface contact in a bolt action. thus reducing friction.
I wonder though, how much the diameter of a bolt that is not bead blasted would increase if it were bead blasted without turning it down some first?
Paul