Nick, who do you figure did the best homework as far as optimising for yaw condition, and why? Also, it was mentioned earlier that most of the details for WRC technical regulation is under 400/01 which is not publicly available. I never realised this, and find it rather annoying, because I enjoy reading other regs (e.g. LM P1, LM GT) when I'm not neck-deep in 400/01, which give full description of legality zones (among others), and are openly available on the FIA website. To comment on your earlier post about underbody treatment, check section 902 of article 255a (the public release). Basically, it doesn't say much, and certainly some take a more liberal approach than others with regards to treating it as an aero device.
Re: diffusers. Without disclosing any numbers or commenting on Nick's estimates ;), I tend to agree that the physical effect of losing the diffuser would be only a small shift in aero balance, but enough to disturb the driver's confidence, which is far more important than a few kilos of downforce. So, as Nick said, some real effects, but mostly what's in the head.
Re: doors. Don't forget that these doors are first gutted out (OEM side impact beam and all) then filled back with energy-absorbing foam. Of course, the OEM beam is replaced by the roll cage, but this is not realistic for a road car, and neither is filling the entire door + outer extension with foam. Among other things, the window won't go down! Incidentally, in my previous job I was involved in a lightweight electric car project with composite monocoque and bodywork. Inside the door was a CFRP side impact beam which housed an outside-firing airbag. I believe it was for pedestrian protection more than occupant protection, but interesting nonetheless. So maybe something like this could be applied to road cars, before we see extra thick rally-style doors.
My earlier comment about the looks of the cars with the side impact is perhaps best appreciated in the context that the regulation of the side impact was finalised rather late for the '17 cars (Article 255a says 10/03/2017 for the official publication date for the side impact :rolleyes:). That's why most "stage 1" test cars didn't have them, including the Polo until its last few runnings. I do like how Toyota took full advantage of the "requirement" and essentially created a continuous aero surface along the top of the extension which runs flat until the rear wing elements. That car is like a giant wedge in that sense, and seems to be making good downforce as well.
What I find also interesting is how the aero concept of the cars seems to have converged so well, first time around. Yes, the individual execution differs, but overall the flow management is much the same. I'm sure we were all thinking that we were being smart and outplaying others within the regulations without going full DTM-mode, but I suppose there is only so much that can be done on a road car base, aerodynamically. Do note that there seems to be two distinct approaches for the rear wing. Had VW been around, it would have been 3 for "fully separate upper and lower element" vs. 2 for "double decker with end plate + side element".