That makes sense. I think the key learnings from the maths are that a 45 degree bank is optimal for height loss per unit angle turned, and that there is no optimum speed -- it should be as low as possible. It then becomes a question of practicality and pilot skill.
When I did some simple flight tests at altitude in a Mooney 201 (probably 15 years ago), I remember some insights:
* with a decent aural stall warner, flying the aircraft at the edge of the stall-warner onset is fairly easy -- relax pressure if you hear it
* while the height loss in the turn is fairly low (as Rogers predicts), you come out of the turn at low speed and at high rate of descent -- that requires you to accelerate to make a normal landing, which takes time and height
* in the analysis of where you end up, a lot depends on the difference between climb gradient and glide ratio -- in a powerful aircraft, it's much easier to get back to the airfield, while in a poorer performer (or at higher weights, density altitude), you'll end up beyond the upwind threshold.