Rather than worry about “efficiency”, I think we need to fall back and look at the graph in my original post. The measured and the CFD data show that the lift produced by the sharp leading edge configuration is very very linear between +/- 8-9 degrees of angle of attack. I initially thought for sure that the razor sharp LE would not work at all. I thought that it might have a positive effect on grooving or killing any hunting, but intuition said, like David I think believes, that it would be a disaster once I started to maneuver. That’s not the case.
All along the problem with the tail end of a stunt ship has been getting a repeatable linear effect from the stab/elevator. Sealing the hinge line helps greatly with repeatability. The shape of the stab appeared to be very important also. Many different sections have been tried through the years, some better than others. The “pointy leading edge” is the latest attempt that has shown good results. I was looking for published measured data that supported the empirical results, and it’s there.
Since flying one of these “razor tails” work and we don’t feel a stall of the tail, I think we can conclude that the tails don’t actually see a very high angle of attack like we might suspect. Surely something less than +/- ten degrees. This is a question that would be fun to look at, “what is the incidence of the flow at the tail during a tight square maneuver?”. The dynamics of the flow at the tail in a tight turn is effected by several factors; the downwash from wing and flaps, the induced up wash of the elevator, the length of the tail swinging through the arc of the turn, and the rate of the turn. We need a simple simulation to look at this.
So. look again at the data and observe the very linear relation between the angle of attack and the lift for the reversed flow condition. Don’t try to look at a big picture of a backwards flying airfoil at large angles of flow.