Peter, great to see you collecting this kind of data and getting some usable insights for us all. I think what a lot of people forget is that the lifting ability of the frame plotted against weight is not a linear curve; when you are well under 1000g, you can add grams a lot more easily with very minimal effect on flight time because you are still well within the ability of the motors + props to deliver lift. Once you get past about 1100g, though, every additional gram carries a steeper penalty -- 1 second, 1.2 seconds, 1.5 seconds -- until you get to that very last gram at which the frame can no longer sustain flight at all. So the "last gram" you might say carries an infinite loss factor.
The other Great Debate that we should be having is, how far down can you *really* fly your packs without damaging them -- or perhaps better said, where the detriment to the pack is balanced against the benefit of additional flight time?
For example -- would you fly a $25 pack down to 8.9V if it meant you could have 11 full minutes of flight
with a gimbal and VTX, but totally ruined the pack? Probably not. On the other hand, would you do it if it shortened the life of the pack to only 25 flights? At that point you just might, because the extra utility only "costs" you $1 per flight.
I am discovering that some of the very well made packs from Thunder Power, MaxAmps, etc., are much more "tolerant" of being flown to far lower voltages than the recommended 10.7V (under load) and higher depletion than the "common knowledge" of 80 percent. As a test I have flown a Thunder Power 2700mAh pack down to 9.2V loaded (recovery voltage of 10.7) about 10 times and so far see zero impact on capacity -- but of course I need about 30 more flights before I can make any claims.
However, just being willing to take that slightly larger pack well below the Phantom-default voltage levels took me from 4:45 (with stock battery and stock voltage settings) to 9:05 (stock props, Zenmuse, OSD, Immersion 600mw TX). Just using the larger battery with stock voltage settings only netted me an additional 30 seconds or so. But being willing to push the battery harder got me to nine minutes. Now, am I killing the pack? Don't know yet -- but even if I am, I'll pay $2 a flight to fly almost twice as long.
Point is, I think even with the stock batteries we could all be flying significantly longer with minimal imapact through more aggressive voltage management.