It looks to me like its really close to a posi with a "non-smooth" breakaway torque. At this point I think it's best for drifting, but like you Ken, I need to digest the photos more.
OK, upon further "mental-modeling" I seem to understand how it works.
The balls and plates, are free to move left and right inside the case. Imagine holding the case in a vice. Apply torque to ONE of the output shafts. The "zig-zag-ball-track" from the rotating output shaft will cause the ball-plate assemblies to slide laterally and "cam" the other output shaft (which is out of phase with the opposite track) into turning the opposite direction. Think "barrel-cam".
Now that we see that the left and right shaft are mechanically linked via the ball-plates and cam tracks, resistance to opposite rotation (breakaway-torque) is purely a function of the spring loaded clutch pack in between the two output barrels. Ignoring ball and track friction.
Understand?
In theory, this should be a smooth acting breakaway, but with torque being applied from the input shaft, the friction in the ball-tracks could cause some stiction and inhibit smooth breakaway. In fact, under large input torque, depending on the phase of the balls with the cam track, it could be a rising breakaway torque with rising input torque. (I really can't fully model this in my head without seeing it first).