I think this topic went off on how to measure the test signal levels or something along those lines. What that has to do with Q accuracy is beyond me as it is a ratio as you stated and as those familiar with this know. Maybe the Q was covered but I can't easily see how. I am a Nanovna user now that I am retired and don't have access to the big $ instruments where I used to work.
In regard to your question about measuring the accuracy of the Q returned by the Nanovna, I want to mention that this is a tough measurement to make accurately and get consistent results between instruments....even HP instruments until improvements were made in how the measurement is done. This became apparent when I worked for a company that had a lot of custom inductors made by outside suppliers and Q was not a good spec to use unless it was a one-sided spec stating that the Q had to be greater than some quantity and the Q had to be fairly low. The problem surfaced when the inductors were specified against measurements made using the old HP 4342 Q Meter and even the earlier Boonton Radio Corp Q meters from the 50's. The 4342 and the BRC-160 for instance would report the Q to be much higher than the more modern instruments. These older Q meters used a tuned-circuit method where the Q was calculated (using meters) to be the ratio of the voltage across the inductor to the applied voltage. The newer instruments used reflection methods (couplers to measure S11) or used 4 wire measurements. The problem is that the resistive part of the impedance is so low compared to the reactance that these instruments didn't have the resolution or the ability to determine R in the presence of a high reactance or could not resolve the phase angle accurately enough and consistent enough. ( a small change in phase makes a big shift in the small R)
Now for the Nanovna, I found the accuracy of high impedances to be questionable above a few hundred ohms making it difficult to determine the Q accurately. For instance, the last time I did a comparison, my old BRC-160 determines the Q of a particular inductor to be about 130 and the Nano came back with something in the 40's. To get around the problem with accuracy at Hi Z's and even low Z's, I use the S21 Series-Thru and the S21 Shunt-Thru methods. Accuracy has been good enough for my impedance matching and comparison things that I find my self doing.
