question

That's the tip of the iceberg. The description "friction/torque" is a COMPLETE misnomer - the effect is purely a TORQUE and the consequence is severe for anyone that tries to actually stop the engine in any situation other than parked on the runway. I.e. it is NOT possible to stop an engine by applying a combination of torques unless all those torques somehow add up to exactly zero and if there's any airspeed at all this is impossible. You need friction for that and instead MSFS relies on some custom code to get the propeller to stop on the runway which is the only situation power pilots experience.

Trust me on this - I have spent many many hours programming engines/propellers in gliders that actually need to stop (& start) while airborn. The Asobo DG1001 glider doesn't manage to reliably stop its propeller either, it's easy to put the plane into situations where the prop windmills out of control and becomes unstoppable at any flyable speed, so in all gliders we fake the animation for the prop to look stopped if the RPM is below some threshold (100 rpm for the AS33, 700 rpm for the Discus). This is the reason no airborn stopped fixed-pitch propeller is ever simulated on a GA aircraft in MSFS (they all have very high windmilling speeds). Yes it's been reported as a 'bug' but power pilots never stumble across it so understandably it's very low priority.

So IMHO the SDK docs should be corrected to just "torque" to reflect the reality of the simulation.