You make a very good point in para 1; the likelihood is that a typical watch tech will see a greater proportion of cheaper watches containing cheaper movements because in all probability there will be more of these floating about.We should remember that Sellita has sold their cheapest version of the SW200 to the cheapest brands, while the 2824 has been sold in much higher percentages in its more upscale versions to bigger companies with more extensive quality control. It wasn't until five or six years ago that luxury (non-Swatch Group) brands started buying Sellita movements in bulk.
My Sellita-equipped watches have required less repair than my ETA-equipped watches, but I don't own more than one or two watches with standard-grade movements. If a watchmaker has substantive data, we should be able to evaluate what has been going wrong, the relationship to watch-company price point and QC, and any trends over time.
Example: I have heard many American car mechanics of deep experience complain that Toyota and Honda use bolt sizes one size too small for any given task, compared to traditional American carmakers. I happen to hold that view myself, based on (professional) experience from so long ago that it is probably no longer relevant. But Toyotas are still undoubtedly more reliable than those other brands.
And some even deeply experienced mechanics will also have deep-seated opinions about Ford versus Chevy, based as much as anything on what their pappy liked, and will color their own experience with that bias.
There are lots of ETA movements that passed through the Sellita factory, when Sellita was doing mostly contract work for ETA. That doesn't make data any easier to interpret.
Rick "extrapolation causes a lot of Internet FUD" Denney
However, whilst I don't know a huge amount about movements, but I was under the impression the only differences between lower and higher grade movements is specific components (shock protection; balance wheels and hairsprings etc), and that the rest of it was bog-standard, assembled in the same way to the same tolerances, thus unless failure is in these areas (possible with shock protection I suppose), then there shouldn't be that much difference in reliability, only accuracy.
I didn't have time to explore in detail the areas where Sellitas fail, so I can't comment. If I get the chance I'll do so.