I would love to see a difference between a FF 20 1.4 vs a 13 f2.8 on a cropper (granted that is 2 stops which can be huge DOF wise)... what scenarios produce a result on a FF that simply is spectacular and very different than anything on a cropper. I keep seeing these comments, but never see back to back comparisons.
As to a 200 f2, Sigma offers a 50-100 f1.8 for a cropper that would get very close to the results. For a 400 2.8, just use a 300 2.8 on the cropper, actually giving you 480mm for an equivalent view.
I think the only arguments can be made around the wide angle super fast primes. Croppers can get very, very wide, and obviously get the reach. I just think using a total of about 2-3 different lenses as the leading reason a FF is better isn't enough fodder for an argument personally.
Also if I wanted a wider view with more DOF using the same lens as on the FF, I would probably try to use the brenizer method. That would widen the shot and give me that DOF if not less. Give me a 20 1.4 and a cropper, and I can give you a FF view using about 6-9 shots. More hassle (shoot in manual, merge the images, then crop), but then I go back to the first paragraph on what the scenarios are.
I still think the same generation crop compared to the same generation FF will show that the FF still holds at least a 1 stop advantage for ISO, a bit more sharpness, better colors, more creativity with DOF control, etc. This is where I choose to use FF. An 80D isn't going to be "almost indistinguishable" to a 5D4 or 6D2 should one come out this year. Low ISOs, sure, but at higher ISOs, I think you will see this 1 stop difference, especially if you have to crop in any on both and then compare.