I have seen the comparison test in this thread, but wouldn't call it scientific. There is a lot of anecdotal evidence in the thread that the capacity reduction is real, but that doesn't prove it is (or isn't, for that matter). My point is, there is still no conclusive evidence if the degradation is real or a software issue. There are arguments both ways.
I had wanted to try a discharge test over a fixed resistor to prove one way or the other, but it looks like the battery is more complex than I thought and no power is applied to the battery terminals unless a command is given to it. So for now, thats not an option.
With regards to the app readings, I agree it's just reporting the vales from the hardware. When I was referring to potential software issues, I was referring to the battery (BMS) firmware.
I just find it strange that everyone seems to see the exact same degradation characteristics.