Skip to content

Mental Simulation

Why It Matters

Formal models are expensive to build and easy to overfit. Mental simulation is cheap and often sufficient. Many structural errors are obvious when you "run" the model in your head - this stock only ever increases, that loop has no limiting mechanism, this delay is implausibly long. Simulation before formalization saves effort and improves intuition.

The Key Move

Talk through the system out loud, moving through time. "Suppose this stock increases. That speeds up this outflow, which fills this other stock, which activates this feedback... wait, that can't be right because there's no limit..." The places where narration breaks down reveal structural problems or knowledge gaps.

Classic Application

Business case analysis. Before building a spreadsheet model, mentally simulate: if we enter this market, competitors respond how? Our costs change how? Demand evolves how? Walk through several years. Where does the story break down? Those breakdowns are either model errors or genuine uncertainties to investigate.

Surprising Application

Difficult conversations. Before a tough conversation, mentally simulate: I say X, they likely respond Y, I then feel Z, which leads me to say W... Playing it forward reveals likely failure modes. "If I start by criticizing, they'll get defensive, I'll escalate, they'll withdraw" - the simulation suggests restructuring the approach.

Failure Modes

  • Simulation as wishful thinking: it's easy to mentally simulate favorable scenarios. Discipline requires simulating adversarial and unlucky paths.

  • Overconfidence in simulation: mental models have systematic blind spots. Simulation surfaces some errors but not others. It's a check, not validation.

  • Avoiding simulation because it's "not rigorous": sometimes rough intuition is better than precise nonsense. Formal models can obscure rather than reveal.

Go Deeper

  • Sterman, "Learning in and about complex systems" - simulation as learning tool
  • Gary & Wood, "Mental models, decision rules, and performance heterogeneity" - empirical evidence on simulation effectiveness