First, I think you are assuming too much. Normal returns mix the water pretty well and while the pump is running, I doubt there is much of a temperature difference between the top and bottom of the pool if you took the time to measure it. And if there is, you can just point the returns downward and solve that.


total expenditure of about only 1/3 of what it costs doing it "normally"...
How do you get that?



It occurs to me that you are trying to pump against a column of water 4, 5, 6 or 7 feet high, and each cubic foot of water is 65lbs, so you're moving at least 260lbs of water (yeah, I know it's not a square foot, but you get the drift).
After leaving the pump, the water starts near water level, travels downward to the main drain and then back up so you really aren't pumping against a column of water because there is head gain on the way down and head loss on the way up so net static head is 0.