Simulating Police Training - What's the pedagogy here?

My friend RC alerted me to this piece on a police training simulator for learning how to reduce misuse of force when encountering autistic people or people with various mental disabilities. It touts the virtues of the virtual playback versus role-playing-based training.

Through the simulator, deputies are immersed in true-to-life scenarios — exactly the kind of situations they often find themselves in: A father calls the police because his bipolar daughter is off her meds and is destroying the kitchen; a man is wandering around a parking lot with a knife talking to himself; an angry young veteran, suffering from post-traumatic stress disorder, is walking down the middle of the street.

The simulator’s operator changes the scenario’s outcome based on how the deputy responds – and the system recognizes and records each type of response, whether it’s with a taser, pepper spray, a gun or verbal interaction. Just like in real like, no situation is ever the same

“The key here is decision-making. We can test and practice-decision making among deputies in a way that wasn’t possible before,” Lt. John Gannon said.
And Gannon can watch it all unfolding, as if he was on the call himself. He can see if a young deputy misses a knife sitting within arm’s reach of a suspect, or unnecessarily escalates a situation involving someone with mental illness. Then the supervisor and deputies can debrief, going frame by frame through the situation to discuss what was done right or wrong.

“How often do you have a chance to go back and see what someone did, see what someone missed?” Gannon said. “Rather than role-playing, rather than having them learn on the job, we can have them learn and make their mistakes here.”

A few thoughts.

1. The pedagogy of police training continues to fascinate me. The police trainers I've talked to are often critical of the method but not empowered to try active learning models. It's mostly talk at officers, dumping information in lecture format. 
2. This simulator, like the more standard role-playing, is obviously active learning.
3. I don't see the evidence in this piece, although it's interesting, that a simulator is better pedagogically than role-playing. In standard role-play scenarios officers work with actors who then go back after and debrief on what the officers did well or poorly in their decision making.
4. Decision making in tense situations is hard.
5. Finally ..

The piece concludes with this insight:

Lessons Learned:

Some of the same techniques can also help deputies when they are responding to someone with PTSD, schizophrenia or bipolar disorder.

“You don’t always have to shoot, taze, pepper spray — sometimes talking works,” Deputy Shawn Walters said.

That's true. The key is to take that insight out of disability-related training and apply it to every encounter, because people don't broadcast their disabilities to the world. It's why I'm skeptical of diagnosis-based training in general. It says: Apply these specific resources to specific cases.

I like police training models which use insights from the disability world to help officers fundamentally change the way they think.