Games to world models

Ever played Chinese Whispers, or Concentration? I recall playing these group games at primary school with the entire class sitting around in a big circle. I still remember the rhythm and surprising difficulty to Concentration, and the absurd results of Chinese Whispers. Thinking back, it is odd that such games were part of the school curriculum (presumably).
The results of the games are still quite memorable, why?


Chinese Whispers has a few quite simple rules:
1. A person starts by coming up with a phrase or sentence and whispering to the next in line.
2. Each person whispers to the next what they (think they heard) from the previous.
3. The last person announces what they heard and compares it with the starter.

At the time, it was quite a surprise that the end result differed so much from the original.
Why?
A. Someone (or several) changed in intentionally.
B. Someone (or several) repeated incorrectly.

The main learning outcome, for me, was that it is in human nature to error. Perhaps in context it could be applied to mean that rumors are not to be taken at face value.

Concentration
Rules:
1. All players clap in harmony to a 2-2 beat.
2. Someone starts by singing their own name twice, and then another players name twice in rhythm.
3. That named player does the same.
4. The loser is the first one to mess up the rhythm or the names.

This appeared to be a simple game, but when played, it was surprising to see who lost each round - and how. To me, the game taught two things:
1. The Role-taking of the singer. There felt a huge pressure as everyone watched the current singer. The game taught me about this pressure (and with time - likely how to deal with it)
2. Strategies on how to deal with that pressure. Ie: thinking of another players name before you were chosen.

Now, looking at the learning outcomes compared with the games themselves - it seems there is quite a difference. How can such simple rules, lead to such valuable learning?
This is an example of using the play in games to create abstract models for use in the real world. Is there a way I can reverse engineer this idea? Ie: Come up with a model for use in the real world, then create gameplay that would result in that abstract model?