I think that thread summed it up well refrigerant (DX) versus water/glycol loop. I think the soil side thermal conductivity has to be limiting in most cases (except in aquifers that are slowly 'flowng'). So the higher eff claims need to be based on side by side field reports. I could believe that there is less pump work if the refrigerant is lower viscosity. Personally, the heart of all these systems is the compressor, and the key to longevity is keeping foreign matter out of the compressor wear surfaces and keeping the oil in the motor. It seems that many of the problems (at least with ASHP) are from sloppy installers letting contaminants into the refrigerant loop lineset. So, I would be much happier with my refirgerant loop being contained in a factory sealed unit, rather than running through 100s of feet if field installed tubing.