I din't check the video, cause i'm not very interested, but i've read all previous posts.
I agree that you should be suspicious if the lens you're testing shows sub-par performance, but let's be fair. The only tests done on multiple samples are LensRental's.
Most respected testing sites don't warn about sample variation, nor re-test optics that should deserve a second chance.
Very likely we also never see negative tests of brands that actually finance the site.
I think that some comments are a bit harsh. Some people seem to take it personally
IF (i repeat, if) the issue has nothing to do with lens calibration, and the lens showed no visible damage, i'd say that it's just a bona fide mistake. It happens.
If you want to get an idea of what sample variation means, you don't need to read the very informative articles on lensrental.com, just a quick peek at:
Measured Resolution Numbers
Film era tests done with resolution targets, with "user lenses", some purchased new and other acquired second-hand years before.
The numbers are impressive. The much praised 77mm Limited is not a proper lemon, but is actually outperformed by humbler lenses.
I remember reading the methodology, long time ago. The tests were well made, and the testers quite trustworthy.
The context s different, but it's still a good example that shows how sample variation can affect tests in a major way.
Telling the difference between "damaged sample" and "bottom of the range" isn't as easy as it might seem...
I guess everything should be taken with a grain of salt. There is plain incompetence, there are biased reviews, and there is sample variation... most of the times we can't tell. Using common sense always helps, though
cheers
P