A while ago I was listening to this on line lesson by a university professor supposedly expert on learning with games and the reference game she was basing her lesson on was a physics enabled environment where you could handle simple object and ropes and see the physical consequences of the forces you applied.
As the “lesson” went on I felt a growing sense of discomfort: that wasn't a typical case of learning with games. That was a simulation, where you could interact and test. That is not a game. I progressively realized that the “expert” just knew little of games.
A classification of interactive systems inspired by Keith Burgun writings.
The learning process in games is rarely by their being explicitly set as a simulation. There is often indeed an underlying combinatorial model, but the game play experience is a learning experience in a loop of attempts, not a laboratory session.
When you propose or conceive a game for learning a topic, field expert will often come with a simulation model as their first idea. But games are quite special as learning environments not because they allow for A-B test, like in a lab, but more because of the built in defining mechanics of play where to fail and retry is just fine. Try, fail, retry, fail, retry, get better, next level. And learning is in the process, and can be built in the evolving narrative of your game play.
A last note: shouldn’t be schools be built around the same principle? Instead of being built on “failing is a disaster”?