Usually I would take into account the Power to space use ratio but this time the “game” is to determine the maximum power by kite area according to different methods.
The topic on How to calculate the power generation from the given kite size? Approximation. is based on the traditional calculation from M.Loyd’s model, leading to a theoretical equality between flygen and yo-yo methods. For both we have a power of about only 4/9 of the potential of the kite (perhaps even less) as a result of the loss of speed by the propeller drag (flygen) or by the swept area going downwind (yo-yo), assuming the speed becomes 2/3 from the initial kite speed.
The loss with an hydro-turbine could be equivalent to the loss of a flygen, if there is no other drag than the drag from the hydro-turbine.
Concerning yo-yo method the time of production is about 1/2 the total time. So the extracted power would be less than 1/4 the potential of the kite.
Concerning flygen method the extracted power would be less than 1/2 the potential of the kite.
Now other methods such like ground-based crosswind rope drive or straight rail could perhaps extract more as there is no downwind motion (unlike yo-yo) or aero(hydro)dynamic drag to convert wind power (unlike flygen and hydro turbine), the kite motion being almost crosswind. However some losses could occur at each half turn due to the required time to do it. Another option is for the kite to do figure eight downwind like for NTS rail, knowing the loss due to downwind motion would lesser than the loss with yo-yo method (point to be clarified, however).
Payne’s patent US3987987 figure 5 has some similarities and some differences with the immediately preceding methods. I think the simultaneous traction of both pulleys leads to some loss.
Rotating devices can be efficient but don’t use the full potential of a kite due to the limit of the speed, above all the inner side of the kite.
There are other parameters such like the quantity of involved material, favoring yo-yo method (excepted for the generator).
These are rough approximations but this topic could allow to determine a few orders of magnitude.