With respect to theoretical height loss from apex to completion of a loop, the minimum radius will be achieved by maintaining the maximum usable angle of attack. Therefore, what is needed is knowledge of the lift coefficient that can be achieved. The lift curve slope of the wing will vary as a function of both Mach number and Reynold's number although below about 0.5M the predominant effect will be from Reynold's number. Therefore, the maximum angle of attack and, de facto, lift coefficient will probably vary throughout the manoeuvre. Unless the aircraft is above corner speed, the normal acceleration achieved is irrelevant although its value will be used for any height loss calculation. Drag will indeed affect the speed profile through the second half of a loop but thrust will also have an effect; it is thrust - drag that will determine the actual speed profile and thus the radius.
When pulling through at maximum angle of attack with a maximum speed that is below 0.5M and is below corner speed, the height loss variation as a function of the speed profile is actually less than many may intuitively think.