Keep in mind that as your RV is slowing descending a steep grade (using mostly transmission braking) the inertial brake is applying ever increasing brake pressure in response to the grade and deceleration forces.
Should not be "ever increasing". The braking force should always be proportional to the rate of acceleration. That's the advantage of inertial vs time-based controllers.
The potential problem with these systems is twofold:
1. On long grades the brake is on for long periods and can overheat the car's brakes. The car brakes typically cannot shed heat as well as the larger motorhome brakes.
2. The Proportion of braking may be too high and the toad essentially tries to drag the motorhome to a stop. This can be difficult to detect when the toad is such a small fraction of the coach weight, yet the amount of braking causes overheating and is fatal to the car brakes.
Both of these are handled by proper adjustment of the brake sensitivity controls, but there is no simple yet foolproof method of figuring that out. It's pretty much a "seat of the pants" thing, trying to sense any subtle pull-back as the toad brakes come on. Not very scientific and highly dependent on each individual's sensitivity to forces. Ideally there would be a meter on the tow bar showing whether the forces on the tow bar are neutral, positive (forward) or negative (backward). The goal is that the force on the tow bar is as close to neutral as possible, i.e. the car is neither pulling nor pushing the coach.