@JS_Brouillon replying to one of my comments: “The “triangle” arrangement of the kite-to-kite tether also prevents the speed imbalances from making a mess.”
In some way, @Rodread 's Daisy achieved connections between the blades. So an alternate approach would be to supply a pitch control mean. I remember Daisy had a very good power coefficient (Cp) of 0.15 or 0.18, perhaps 0.2, and with a potential for improvement. An important advantage is the consistency, with the blades rotating at the same speed during each rotation, unlike crosswind kites which lose a lot in accelerations and decelerations, not to mention the risks of mess involved.
A TRPT could work like a solid structure up to limit heights.