Simulation is basically an illusion. We often think about something that`s about to happen, it`s something we do every day. However the next level in thinking about something that`s about to happen is to create a material support/prove/aid for the thing that is about to happen that is not the thing itself.

I think you're so focused on the idea of mental models, so that you forget about actual reality.

For a game that means: You need simulation first and foremost to simulate the actual reality of the game. E.g. Characters walking over terrain and not falling through, inability to walk or shoot through doors, maybe even rigid bodies, ragdolls, vehicles, etc.

In this context, the simulation is no illusion but models reality. Ignoring the point of it being a fictional reality.

And because reality is more essential than intelligence to get a game, you might want to work on the former first.
After that, you see how much of it can be reused to run mental simulations, or if a simplified system is needed.

The problem is: you have to work on the problems you actually have right now, not the problems which are interesting.
If you don't like that, do what i did: Quit working on a game, and focus entirely in research of crazy ideas.
But be warned: I really miss to produce something… playable ; )

I don`t care about vehicles. That`s the last of my worries praise God Yahveh

mental models

Characters walking over terrain and not falling through

valid point. I think the important thing is to see were it drifts, like what the overall message is, is it something technologicaly plausible that could have military application for example.