So, my first issue with this list is the blanket inclusion of multiple capacity/model for a given series all being lumped together. IMO, that can't happen.
Yes, I realize that it's a lot more work to differentiate between the 550, 650, 750, 850 models, etc., for each series, (Which is why MANY people have begun then quit, working up a list of their own. Including you Jon) but I think that's something you have to do especially since we already KNOW that in many if not most cases, there can be major differences from model to model within the same series. There are VERY good examples of this. I do see that you have SOME done this way, but ALL need to be done this way. Can't do some series one way and some series the other. If there is no evidence that exists for a given model within a series, then, IDK, it doesn't get listed or something. No evidence or hard review should mean no placement on the list. How can you place something in a given spot if you have no evidence to back that placement up? At the very least, verification that a given model has the same platform as a model that DOES have a review, should be prerequisite, and that's being lenient.
Seasonic M12II for example. The 520w and 620w models are group regulated and are an entirely different platform than the 750w and 850w models, which are DC-DC and couldn't reasonably be included on the same tier or considered to be the same type of product as those group regulated parts. EVGA as well. Several of the B3 units Aris tested blew up in testing. One capacity model did not and actually tested out well. There are a variety of examples out there of MANY series from different companies where entirely different platforms are used within the same series, so I think any series that has multiple models at different capacities needs to be considered a unique entry per capacity if any sort of accuracy is to be maintained.
Immediately the E series NZXT units in the very first section of the list come to mind. Aside from the fact that in the review by Aris for the 850w model he clearly states that it is noisy, even in silent mode, has poor transient response on the minor rails, and has some inaccuracies with the software readings. I'm willing to bet that not every capacity model of the NZXT E series turns out the exact same result, they might, but having some evidence of that should be a requirement before simply assuming that they would all be the same because VERY OFTEN they are not.
The fact that the only difference between this unit and the Seasonic SSR-850x it is based off of, is the addition of a digital controller, which by the way Aris says does not improve performance, while the Seasonic SSR-850x itself is entirely missing from the Tier A, suggests the bias against Seasonic that has been pretty evident on this list since it's inception. Might want to address that as well as the blanket lumping of models into one listing as though different capacities didn't perform differently, have some minor different implementations or entirely different platforms in some cases.