I'll address this question later because it's largely irreverent to RWS discussion. Just consider the radial distance difference between the FWS and RWS CoM during a turn. Note that the RWS vehicle's CoM is further from the turn's virtual axis, therefore, its angular velocity is higher (at a given speed). This higher angular velocity elevates centripetal forces and likewise increased lateral contact patch loading. But that's just tip of the iceberg...

RWS steering complicates the design because you CANNOT employ conventional FWS thinking. For example;.. Adding trail to the RWS layout does, indeed, improve self-centering. However, steering 'feel' becomes very light as you begin turning (at speed), but upon recovery, becomes quite heavy, making it virtually impossible to subdue the over-steer. In other words, steering does not have a uniform and predictable 'feel'. If you design without trail, the steering 'feels' more uniform and predictable during maneuvers, but you lose self-centering on the straights.

I'm not one to say (or even think) that RWS cannot eventually be conquered and evolve into a safe and viable production vehicle. But don't expect to see me on the engineering crew trying to perfect it. I'm rather fond of my successful FWS designs to drift into such a complex design.