Mental Models

Don't stop at the obvious consequence. Ask "and then what?" three levels deep:

Action → First-order effect → Second-order effect → Third-order effect

Apply when: Evaluating a decision, policy, or deal. The first-order effect is what they tell you. The second and third are where the real consequences live.

3. Steel Man

Before criticizing a position, construct the strongest possible version of it:

• What would a reasonable, intelligent person mean by this?
• What context might make this position rational?
• What evidence supports it?

Then critique the steel man, not the straw man. If your critique holds against the strongest version, it's a real critique.

Apply when: You're about to dismiss something. Build the best case for it first.

4. Incentive Analysis

People respond to incentives. Map them:

• What are each actor's incentives? (financial, political, social, personal)
• Where do incentives align with stated goals?
• Where do incentives contradict stated goals?
• What would a rational actor do given these incentives, regardless of what they say?

Apply when: Someone's actions don't match their words. The incentives usually explain the gap.

5. Hanlon's Razor

Never attribute to malice that which is adequately explained by incompetence. But also:

• Once incompetence is ruled out, consider malice
• Pattern of "mistakes" that consistently benefit the same party suggests more than incompetence
• Incompetence and malice are not mutually exclusive

Apply when: Something goes wrong. Start with incompetence, escalate to malice only with evidence.

6. Survivorship Bias

What are you NOT seeing?

• Success stories are visible; failures are silent
• The data you have is filtered by survival — the dead don't report
• Ask: "What would the failures look like? Would I even know about them?"

Apply when: Someone cites examples of success. Ask how many tried and failed invisibly.

7. Inversion

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

• What would make this deal catastrophic?
• What assumptions, if wrong, would destroy the plan?
• What has historically caused similar things to fail?

Apply when: Planning or evaluating a strategy. The failure modes are often more informative than the success path.

8. Base Rate

What's the historical success rate of similar things?

• Before believing "this time is different," check the base rate
• If 90% of similar ventures fail, you need strong evidence for why this one won't
• Extraordinary claims require extraordinary evidence — and extraordinary base rates

Apply when: Someone claims exceptional outcomes. Check what normally happens.

9. Map vs Territory

The description of reality is not reality itself:

• Models are simplifications — useful but incomplete
• When the map contradicts the territory, trust the territory
• People often argue about maps (narratives, models) while ignoring the territory (actual data)

Apply when: There's a gap between official narrative and observable reality.

10. Opportunity Cost

Every choice has a cost — the best alternative you didn't choose:

• "Is this good?" is the wrong question. "Is this better than the alternatives?" is the right one
• Time, money, and attention spent here can't be spent elsewhere

Apply when: Evaluating whether to do something. Always ask "compared to what?"

How to Apply

When analyzing anything substantive, select 2-4 relevant models and apply them explicitly:

## Analysis: [Topic]

### Cui Bono
[Who benefits and how]

### Second-Order Effects
1. [First order]: [obvious consequence]
2. [Second order]: [what follows from that]
3. [Third order]: [what follows from that]

### Steel Man
[Strongest case for the position being analyzed]

### Verdict
[Your assessment after applying the models]

Guidelines

• Not every model applies to every situation — select the relevant ones
• Models are lenses, not answers. They help you see; they don't decide for you
• When models conflict (Hanlon's Razor vs Cui Bono), note the tension explicitly
• Always apply steel-man before critiquing — it sharpens your analysis even when you ultimately disagree
• Pair with historical-parallel to ground the models in real precedent