One thing that hasn't been mentioned in the discussion of the details of the latch pawls is that, in the case of most of the letch assemblies, some portion of those forces are being transmitted through a relatively small pin at some point. In the one BenK has been detailing with the dotted lines, you'll see that any rearward force on the couplet latch is being transmitted through the 'foot' to the bottom part of the coupler assembly as well as the pivot pin on the latch, and the latch locking lever that dogs into the top of the coupler. How robust they are, and what kind of long-term fatigue forces they can withstand remains to be seen.
Also, there was some talk earlier about the rotational moments that the chains or spring bars apply on the trailer. I think that point merits some further discussion as well. With a conventional WDH using spring bars, the springs use the cast hitch head to oppose the force applied to the frame rails. The result is an increase in the vertical force on the ball as the spring bars use the ball as the fulcrum to redistribute the forces. This is perfectly OK to do, in my non-professional opinion, as the hitch ball is being loaded along as though it's a column, and columns can take a LOT of load before failing.
The key difference in the Andersen system is that the chains cause the ball itself to act as the fulcrum. The way it's currently designed will, of course, apply those loads through the coupler's latch. In my opinion, a better design for this setup would use a pair of arms extending from the ball below the coupler parallel to the trailer tongue that the chains would connect to. Tightening these chains would try to pull those arms down and into compression against the ball-end, the same as the current setup pulls the trailer tongue. Then, since the resultant force will be down on those arms, simply hool the ends of those arms over top of the frame rails. That would allow the resultant force applied to the trailer tongue to be only downward instead of forward and down, and since the whole assembly would still turn with the trailer, the sway-control inherint in the Andersen design would still be retained with little or no movement of the ball relative to the coupler.
(I'd love to draw a sketch, but my MS Paint skills suck...)