Lloyd is entitled to prefer what he does, but that does not change the facts and the laws of optics. And the facts are that it is simply easier to design slower and less compact lenses.
It may have occurred to you that both the Otus and the Sigma Art are big lenses. The Otus gets diffraction limited somewhere between f/4 and f/5.6. Both lenses use a generous amount SD glass and also aspherical elements. Obviously much smaller designs would be possible if near optimal sharpness at f/1.4 was not required.
The Sony 55/1.8 EF lens is a good example of a very sharp lens with slightly smaller maximum aperture. That is a relatively small lens.
If your subject implies large depth of field, indeed larger sensors are at a disadvantage. But this has nothing to do with the intrinsic quality of lenses. Besides, smaller sensors reach diffraction faster with their smaller pixels.
Obviously, a MF camera is rarely required if all output will be 2 mpix max. OTOH, fine art inkjet papers are available in 150 cm (60") wide versions. This allows slightly bigger prints than you seem to consider.
Well, the reason I discuss A2 size prints is that many folks have 17" wide printers and those produce about A2 size prints, and the reason I looked into it was pretty much because a poster asked about visible advantage with MFD. What I see here is that there is small, or possibly none difference in that size between 24 MP full frame using upper end zooms and Hasselblad V-series primes in A2-size prints viewed at 40-50cm distance with normal vision. With 36 MP and high end primes the difference would be less.
In larger sizes the differences are more visible. I do print larger, but not larger than 70x100 cm as I lack wall space. Anyway, it is a fact that screens dont't display more than around 2MP with present technology, 8MP with 4K and 36 MP with 8K. I would love to have an 4K projector, but the few that are around are very expensive. Of course, if you afford MFD you can also afford 4K projection, but it will still only use around 66% of the capability of a 24 MP DSLR. What I say is simply that pixel peeping at actual pixels doesn't have relevance for any normal way to present images. Unless you print very large images at 100 PPI (as computer screens normally have around 100PPI).
Also if you have a large image, you probably need to reposition the observer when comparing images. Small images can be compared with one image on he top of the other, much more demanding comparison. You can of course do that with small crops of large images.
Regardinf diffraction, as an engineer you should be aware that diffraction is not dependent of pixel size, it is just a function of aperture. Of course with a sharper lens or a more high resolving sensor you are going to loose more.
Very clearly, if you are shooting flat objects, use optimal apertures with perfect focusing and near perfect lenses a larger sensor will yield better results.
Also, larger sensors and truly excellent lenses like the HR lenses from Rodenstock and Schneider may offer more visual advantage.
Another point, quite relevant here, is that this discussion is not about MFD vs. DSLRs but about the 645Z. The 645Z is a small sensor MFD device, it is just 44x33mm. Also, very few of the Pentax lenses are modern designs. Lloyd has tested a dozen or so of them on the Pentax 645D.
These two videos may offer some insight in the relevant parameters, even if they are about motion pictures and not stills:
https://www.youtube.com/watch?feature=player_detailpage&v=iBKDjLeNlsQhttps://www.youtube.com/watch?feature=player_detailpage&v=v96yhEr-DWMThe message is quite clear, the low frequency MTF is mouch more relevant for perceived sharpness than high frequency detail, and the high frequency detail is what we see when we pixel peep. I am pretty sure this also applies to stills.
Best regards
Erik