The framework
It frames learning through constraints — the individual, the task, and the environment — interacting over time, with expertise emerging from tight perception-action coupling rather than a linear 'input, decide, execute' chain. Practically, that means designing representative games where the feedback comes from the game itself, minimizing external instruction, and being careful not to over-constrain to the point that the practice stops carrying real game information.
What it means for me on the grass
It reframes my whole job. I'm not a dispenser of solutions; I'm a designer of environments and an observer of what emerges. That fits my belief that players are more capable than we assume — given a rich, representative game, they'll find solutions I never scripted, and those solutions stick because they discovered them. It's also a caution: the article warns that a rule like 'everyone must touch it before you score' can actually destroy the realistic information in the game. So I design with a light touch and let the game teach, then adjust the constraints when the behavior I want isn't showing up.