I think the AF fine adjustment is worth doing for best results. The way I have been going about it has been using center point AF and comparing shots at different settings for best absolute focus in a flat, contrasty area. E.g. comparing +5, 0 and -5, 0 first and then taking it from there. Of course, if it turns out it that the best results are obtained at the setting zero then that is the setting to use
. I suppose it is a good idea to test at a "short" and "medium" distance at least: if it turns out that getting good focus with one precludes it with the other then service/return/exchange is called for.
While it is easy enough to understand that the image and AF sensors might have a bit of misaligment amongst themselves and this requires compensation, the idea that the best setting can vary among lenses with the same body is something for which I have not seen a detailed, convincing explanation so far, yet this is apparently the case, otherwise there would not be a point of having per-lens setting.