Mental Models

Mental models are frameworks for understanding how things work. They simplify complex reality into patterns we can reason about. The quality of your thinking depends on the quality of your models.

What is a Mental Model?

A mental model is a simplified representation of reality that helps you:

Predict outcomes
Make decisions
Understand systems
Communicate with others

Every model is incomplete. Every model is wrong in some ways. The goal isn't perfect models. It's useful models.

Why Models Matter

You Already Use Models

Whether you know it or not, you're already thinking in models:

"If I drop this, it falls" (gravity model)
"If I raise prices, demand decreases" (supply/demand model)
"If I work harder, I'll succeed" (effort/reward model)

The question isn't whether to use models. It's whether your models are any good.

Bad Models Cause Bad Decisions

If your model of reality doesn't match reality, your predictions will be wrong and your decisions will fail.

Example: A manager believes "employees are motivated primarily by money." This model leads to compensation-focused retention strategies. When employees leave for jobs with lower pay but better culture, the manager is confused. The model was wrong.

Multiple Models Beat Single Models

No single model captures all of reality. The more models you have, the more perspectives you can take. Charlie Munger calls this "mental model diversity."

Core Models for Problem-Solving

First Principles

Break problems down to their fundamental truths and build up from there. Don't reason by analogy unless you understand why the analogy holds.

Inversion

Instead of asking "how do I succeed?", ask "what would guarantee failure?" Then avoid those things.

Second-Order Effects

Ask not just "what happens next?" but "and then what?" Most mistakes come from ignoring downstream consequences. See Second-Order Effects.

Map vs Territory

The model is not the reality. The org chart is not the organization. The metric is not the outcome. The description is not the thing. See Field Note: The Map Is Not The Territory.

Occam's Razor

Among competing explanations, prefer the simplest one that fits the evidence. Don't multiply entities unnecessarily.

Hanlon's Razor

Never attribute to malice what can be adequately explained by ignorance, confusion, or miscommunication.

Margin of Safety

Design for more stress than you expect. Build buffers. Assume your estimates are optimistic.

Feedback Loops

Understand how outputs become inputs. See Feedback Loop Analysis.

Developing Better Models

Collect Models Actively

Read widely across disciplines. Physics, biology, economics, psychology, history, engineering: each field has developed models that apply far beyond their original domain.

Test Models Against Reality

When a model predicts something, check whether it happens. When predictions fail, update the model.

Notice When Models Conflict

If two of your models give different answers, at least one is wrong (or you're applying them incorrectly). Investigate.

Hold Models Loosely

Strong opinions, loosely held. Be willing to abandon a model when evidence contradicts it.

Understand Model Limits

Every model has boundary conditions where it stops working. Know what those are.

Common Model Failures

Using the Wrong Model

Applying a model outside its domain. Not every problem is a nail. Not every solution is a hammer.

Confusing Model with Reality

Forgetting that the model is a simplification. The stock price is not the company. The grade is not the learning.

Ignoring Unfamiliar Models

Dismissing models from other fields because they're unfamiliar. Physics has much to teach business. Biology has much to teach software.

Model Lock-In

Using the same model for every problem because it worked before. Past success with a model doesn't guarantee future applicability.

Building Your Model Toolkit

Start with fundamentals:

Understand the basic models in this wiki
Practice applying them to real situations
Notice when they work and when they don't
Add new models from diverse sources
Develop intuition for which model fits which situation

The map is not the territory, but a good map still beats wandering blind.