It's amazing what you can do, at least in theory, with optics on LEDs. I plan to calibrate the light so it is possible to dial in the PAR you want - any combination of the two colors of LEDs. The only problem I see with that is that so much of the light will be reflected off the glass, that my in-air calibration could be 20% or so low. That is still pretty good - aim for 50 mms of PAR and get 60 in water.
As I see it, the biggest problem with lighting a 36 inch high tank is the huge range of PAR values from substrate height to the top of the water. Raising the fixture a foot helps with that, but raising it 3 feet would have helped a lot more. I'm thinking that the next iteration on obtaining the "best" light will be when we can get a light that keeps the PAR within +/- 20% from top of tank to bottom of tank.