Tire modelling is a very complex area. First of all, we have to keep in mind that rF2 tire model isn't even finished...
What you describe Mo, and that Ben Collins, is a well known problem (grip level after the peak) in racing simulations, basically because there is no way to measure such things. On a tire testing machine you can measure the lateral forces at several slip angles, cambers, tire pressures, speeds, vertical loads, etc, but it is usually done in a small range of slip angles (usually less than 12deg). The raw data is also very noisy, I think Niels wrote an interesting document about this some time ago. So, it's not really possible to know certainly how the tire will behave at really big slip angles.
That's mainly the problem of the games who are based on a Pacejka tire model (ISI's games, the old iRacing tire model also had a similar system, etc), which means that you have to design a basic slip curve, which is affected by different parameters on the sim (the same parameters I mentioned for the tire testing machine). With such method you can make a set of tires that can match pretty good the real ones (if you have such data) in the lower slip angles, but after that it's just a guessing game. Some sim developer can think that the tire must loose a crazy amount of grip, and another can think that it shouldn't loose grip at all. At the end they are just programmers... It has also some other issues, for what I've been told, for example when combining lateral and longitudinal forces.
That's why the current generation of sims are trying a different approach (at least rF2 and iRacing, no idea about AC), based more on the tire construction rather than on a empirical approach. Both sims are still struggling in some areas, but I believe they will behave much better in some time, with more experience and knowledge. It's something incredibly complex to develope, and needs time, time and time. But at the end, if it's done properly, they won't need to worry about the lack of tire testing data and the lack of data mentioned at higher slip angles.
Edit: Here it is explained by someone with better knowledge than me (and better english I guess
:
http://www.h-engineering.nl/uploads/Tires in race simulations 20120209.pdf