That paper keeps getting trotted out as proof of something that the paper never tested nor demonstrated.
They took a bunch of used oils, some of which were utterly shagged, had thickened excessivly and had poor TBN.
Then they use a contacting surface tribometer and measured the establishment of the tribofilms with new and progressively used oil. The used oil formed tribofilms on fresh metal surfaces quicker than fresh.
This is entirely to be expected, as the first part of the laying down of tribofilms involves partial destruction of the ZDDP/Mo into more reactive species...some additive companies put a lot of different elemental "variaties" in so that there are different points of activation.
So the oil with the most already partially reacted species in it produced the best tribofilms the earliest...exactly as expected.
The Used oil, excessively thick, wiht no TBN to speak of was not necessarily the best oil to have in the engine. Extrapolating the limits (surface tribometer) of this study INTO that realm is really stretching the bow a bit (well a lot).
And at OCI, the quote that you have provided assumes that ALL of the tribofilm is removed to fresh metal (not true, never seen anything on that).
As to the thought process...
https://bobistheoilguy.com/forums/ubbthreads.php/topics/4060600/1