If the H loses GPS but still maintains link with the ST-16 you maintain a "dynamic" home location. If both GPS in the H is lost, and link with the -16, there is no possibility of the H returning to any specific location, pre-set, dynamic or otherwise. If it does not know where it is, or where anything else is (loss of GPS position reference data and loss of RC signal) there's not a **** thing anyone can do. It's one of the prices of relying on technology instead of maintaining full manual position control at all times.
Something that may be of interest is a difference between military and hobby equipment. If military stuff loses both GPS and control link for a previously established period of time the military stuff is often programmed to self enter a "terminal mode" and fly itself into the ground at a high rate of speed. Hobby grade stuff lacks that sophistication and leaves a small chance of the owner recovering the unit if they marked it with contact information. A slim chance, but still a chance.
My point here is that if the H is the device that loses GPS, and typically that is the case in GPS loss events, you could have programmed a location accurate to the inch and it would not matter. It would have no current location data to reference back from, and no means of navigating to a place that could not exist without an active GPS unit. The only hope of a self return is through the remaining link with the -16, be it stationary or moving. We cannot totally rely on having a continuous and effective GPS data reference for the aircraft to track or return to. GPS is not perfect.
As one that has experienced a fly away, in my case the -16 knew where the H was for the entire event, and the link that was maintained with the aircraft permitted a reference location that permitted locating and recovery. In my case things happened fast and LOS was blocked by trees, which prevented application of effective manual control that could have brought the H back instead of finding a perch in a tree. A programmed home point would not have worked in that event but the link with the -16 might have been used to save the day. Don't know for sure that it would have but anything relying on a GPS location had no chance. Another action that might have been effective would have the -16 placed in Angle mode with no further inputs in hopes the aircraft ceased directional flight. Lessons learned...
In effect, we do have somewhat redundant navigation/RTH functions in the H. GPS of course, but when that fails we have the dynamic link with the H. Associated with that dynamic link is the directional arrow on the video screen. Something to consider is what if a home point had been programmed into the H and the owner was on a boat performing a follow. What would happen if the H lost link when the boat was a couple miles from the take off point? If the boat location at time of launch was the take off point, where would the H go? Back to where the boat was at the time of launch of course. Did it have enough battery to get there Was the return path over water the entire distance? It certainly would not go to where the boat currently was, unless it had a dynamic link with the controller.
What's interesting is that regardless of what we are provided there's always something people want that wasn't provided. If there's a thousand users there's probably a thousand, or more, different wants, and that's just not possible to cover for less than a few tens of thousands of $$. I have them as well. I wish the H was programmed to auto land, in place, any time link is lost for >3 seconds. Perhaps permit the users to adjust that lost link period to something they found more suitable for their situations. But if it had that capability there would be people that would complain it had it. Damned if you do, damned if you don't.