I agree that would be a more convincing experiment - but with a real world rather than sealed enclosure). But much harder to conduct unless you have a second rider willing to go on all the same rides and then switching off who has the dirty chains so the effect of any difference in riding style/weight/etc. can be evaluated. That's hard to arrange over the 5000 or more miles to do the experiment so I'm not hopeful anyone will volunteer to conduct it.

While the experiment that was done suffers from the cleaned chain picking up grit from the cog and chain wheel teeth it still seems to me that it should show some of the benefit from cleaning. The fact that no benefit at all was shown and in fact there was faster wear is pretty convincing to me until I'm shown some different results from either an equivalent or better controlled experiment.