Supposed to be optimal. To me should mean narrower or wider than calculated should lose torque, and it's generally accepted wider will, so to me loss in that direction isn't much proof of anything, tighter than calculated should also be shown to be a loss if the formula finds optimal.