I got this lecture from JohnnyL8 (Vacuum Facts) after he saw my video (above).
He is totally out of his mind and better to avoid and just forget.
"Nevertheless, good effort on your part to at least attempt to show an honest test of the very point which was initially dismissed. It nicely shows everything I said here about how Kirby's airflow drops like a stone on carpet relative to the open hose, is 100% true. Unfortunately, it’s not possible to draw too meaningful a conclusion about the quantified relative cleaning performance from your test, because measurements are inaccurate and suffer from many testing flaws which affect results. Firstly, you only make a single pass forwards and backwards, which doesn’t account for statistical variability of particle removal from carpets; I've discussed this many times on my channel. Secondly, you don’t accurately represent that real-world users would repeat passes (at least twice) for obviously exaggerated messes like you had as an initial condition, and in such a way that unswept belt lines are addressed. Thirdly, you used the Kirby after an initial clean from both units to attempt to quantify the residual remaining. This limits the residual mass extraction to the performance of the Kirby in both cases which is another source of error. This mixing of machines can’t be done in performance testing, as it mixes variables and renders it an unfair test.
You'll see this week some highly accurate, milligram accuracy, fully quantified testing of relative performance of a Kirby with a more representative dust type containing a range of particle sizes, and which accounts for statistical variability in deep-cleaning. Unlike most amateur tests, the results are fairly conclusive and devoid of sources of common error, and are thus less inaccurate. I'll also add that there is no good evidence that US (or EU) homes have one type of carpet over another. This is a made up myth. I heard that most US homes favour hard flooring, in fact."