It is not wise to let battery run empty, so I won't do it just for test. Also it is just old wives tales that you need to run battery empty to get best perfomance. It should not apply to modern li-ion batteries. It is more than enough to run few normal cycles if you want to make sure something like battery meter accuracy. Also you must understand that battery capacity and battery charge level are two completely different things. Capacity is what the battery nominal capacity is, and charge level has no effect to it what so ever (but you can of course cause battery capacity to drop if you keep using it wrong, ie. running empty and store it empty, or keep it full all the time and so on). For example battery capacity in this iPad is 8827 mAh and when you test capacity it should show that what ever the actual charge level of the battery is at that time. So capacity should show 8827/8827 (apps may round it to 8820) and charge level can be what ever. In my unit that is not the case, it shows capacity 8400/8820 mAh and charge level when full 8400/8400 mAh. I hope this helps understand what the issue is.
What margin of error? You think battery health programs (apps, Coconut Battery, in-store diagnostics) work in some random way giving what ever the results from unit to unit? Not. If you read for example link I provided in my previous posts, you can see that they all give the comparable results.
I agree, as long as we are not even talking about the same things.
If I understood correctly warranty terms say over 80% battery condition is considered good. If it is agreeable to buyer, it is up to each and everyone. To put this in perspective, I guess some peoples would accept that brand new car which have estimated life say 500000 miles, has already 100000 miles driven already and still pay the full price. I not one of them.