In my opinion this discussion is flawed. It neglects the depth of field in the subject space at the actual taking aperture in use, which is greater for wide angle lenses at a given f-number. The depth of focus at the film plane, to first order, depends on the f-number in use, and not on the focal length. Because the f-number defines the taper of the cone of light toward/away from the point of best focus at the ideal film plane. If you have a 0.2mm focus offset, but an acceptable circle of confusion of 0.03mm, then the f-number that allows depth of focus to cover the error is 0.2/0.03 = f/6.7. This is true whether you are using a 200mm lens (where the focus offset at the subject is small, but the depth-of-field is small), or a 20mm lens (where the focus offset is past infinity, but the depth-of-field is large in the subject space).
Pictorially, the effects of a film plane offset might still be worse for wide angle lenses because one is trying to get the entire picture sharp from near to far, as opposed to tele lenses where one accepts or wants some blur of the subject away from the focus point. But this text gives a misleading impression.
I agree with Dan that the proof is empirical, at the small-offset level one has to test the system on film before knowing whether to get agitated or modify the camera over a nominal 0.2mm offset.