I have dedicated runs of 12ga and 10 ga. Romex. I also have a run of 10ga. stranded THHN in FMC (Flexible Metallic Conduit). Soon I will run a sheathed cable of 6ga (is this tri-plex? I cannot recall what we used to call it. Twisted three conductor 2+gr?).
The runs are all 35' homeruns.
There are measurable differences in the impedance and voltage drop among the various cable sizes (I have not measures them here) over 35'. That aspect is not theoretical; however how that maps to SQ is a little "squishier."
I think the THHN line in FMC sounds fantastic. I also think it improved a bit by adding a TT7 box to the mix.
The line in conduit is very "fast" IMO. As I know adding a larger gauge run will improve the physical characteristics of current delivery, I am going to run the new #6 cable.
I have read papers and comments noting the parallel conductor structure of Romex creates an undesirable, dirty ground.
I have also read that cable in a conduit should definitely be twisted uniformly to avoid impedance variations along the run. Some maintain only the hot and neutral should be twisted, and the ground run along-side in PVC conduit.
I have read metallic conduit (and outlet boxes) should be avoided, but I have both in some places.
I think there are compelling reasons for using various solutions, including cost and difficulty of a given solution in a particular use-case. Some of those needs will supersede theoretical aspects as they relate to SQ. IMO.