Use when adding new error messages to React, or seeing "unknown error code" warnings.
npx skills add defi-naly/skillbank --skill "skin-in-the-game"
Install specific skill from multi-skill repository
# Description
Apply Nassim Taleb's Skin in the Game principles for evaluating trust, designing incentives, and making ethical decisions. Use when assessing advisors, structuring partnerships, evaluating information sources, designing accountability systems, or deciding who to trust. Also use when detecting asymmetric risk-reward situations.
# SKILL.md
name: skin-in-the-game
description: Apply Nassim Taleb's Skin in the Game principles for evaluating trust, designing incentives, and making ethical decisions. Use when assessing advisors, structuring partnerships, evaluating information sources, designing accountability systems, or deciding who to trust. Also use when detecting asymmetric risk-reward situations.
tags: [decide, persuade]
Skin in the Game
Symmetry between risk and reward. Those who get the upside must bear the downside.
The Core Principle
No skin in the game = advice/decisions without consequences = unreliable
Skin in the game = bearing the downside of your decisions = trustworthy signal
Golden Rule (Silver version): Do not do unto others what you would not have them do unto you.
Why it matters: Talk is cheap. Exposure to consequences is the only reliable filter for truth and competence.
The Agency Problem
When someone makes decisions for you without sharing consequences:
Agent (decides) β Outcome β Principal (bears consequence)
Examples:
- Fund managers (fees regardless of performance)
- Politicians (don't live under their policies)
- Consultants (paid for advice, not outcomes)
- Journalists (no cost for being wrong)
Solution: Align incentives. Make agents bear downside:
- Performance fees only
- Clawbacks for bad outcomes
- Reputation systems with memory
- Eat your own cooking
The Bob Rubin Trade
Getting paid for hidden riskβlooking successful until catastrophe.
Pattern:
1. Take hidden tail risk
2. Collect steady profits (looks like skill)
3. When Black Swan hits, others pay
4. Blame "unpredictable circumstances"
Examples:
- Bankers before 2008 (bonuses kept, losses socialized)
- Executives with golden parachutes
- Anyone with asymmetric upside/downside
Detection: Ask "What happens to this person if things go badly?"
Don't Tell Me What You Think
"Don't tell me what you think, tell me what's in your portfolio."
Talk is infinitely cheap. Revealed preference (what people do) beats stated preference (what they say).
Applications:
- Advisor recommends investment? Are they invested?
- Doctor recommends treatment? Would they take it?
- Founder says company is great? Are they buying or selling stock?
- Pundit predicts X? Are they betting on it?
Heuristic: Weight advice by the advisor's exposure to being wrong.
The Lindy Effect
For non-perishable things, expected remaining lifespan = current age.
Why it relates to skin in the game: Time is a filter. Things that survive have passed tests. Experts who recommend the novel have no skin in the "it works long-term" game.
Applications:
- Technologies: Old and working > new and promising
- Books: Classics > bestsellers
- Foods: Traditional diets > nutrition fads
- Practices: Time-tested > expert-recommended
Formula: If it's been around for X years and isn't perishable, expect another X years.
The Minority Rule
A small intransigent minority can dictate choices for the majority.
Mechanism: If 3% absolutely won't eat X, and 97% don't care, everything becomes X-free (cheaper to standardize).
Examples:
- Kosher/Halal food in general markets
- Peanut-free policies (few allergic, all affected)
- Accessibility requirements driving design
Implication:
- Small committed groups have outsized influence
- Flexibility is weakness in negotiation
- The stubborn win
Operators vs Talkers
Two classes of people:
| Operators | Talkers |
|---|---|
| Do things | Describe things |
| Bear consequences | No consequences |
| Learn from mistakes | No feedback loop |
| Respect from peers | Respect from outsiders |
| Skin in the game | No skin in the game |
Talker professions: Journalists, academics, consultants, pundits, most managers
Operator professions: Traders, entrepreneurs, surgeons, pilots, craftsmen
Heuristic: Take advice from operators, not talkers. Ask "Have they done this?"
Ethical Asymmetry
Via negativa ethics: Avoiding harm is more robust than doing good.
- "Do no harm" > "Do good"
- Not lying > actively telling truth
- Not taking > giving
Why:
- Harm is measurable; good is subjective
- Unintended consequences of "doing good" can cause harm
- Those who "do good" often have no skin in the game
Rule: First, do no harm. Then, if you understand consequences, do good.
Symmetry Tests
Before trusting or acting, apply symmetry tests:
1. Hammurabi's Law: Would the builder live in the building?
- "If a builder builds a house and it collapses and kills the owner, the builder shall be put to death"
2. The Wrath of the Customer: Would you be okay receiving what you're giving?
3. The Grandparent Test: Would you do this to your grandmother?
4. The Newspaper Test: Would you be comfortable if this appeared on the front page?
5. The Inversion: Would I take this advice if our positions were reversed?
Detecting Skin in the Game
| Signal | Skin in Game | No Skin |
|---|---|---|
| Compensation | Tied to outcome | Fixed regardless |
| Personal assets | At risk | Protected |
| Reputation | Long-term, memorable | Forgettable |
| Reversibility | Must live with decision | Can walk away |
| Information | Shares downside | Asymmetric knowledge |
Questions:
- What does this person lose if they're wrong?
- Are they invested in what they recommend?
- Do they face the consequences of their advice?
- Can they exit before the results are known?
Application Checklist
When evaluating advice, partnerships, or systems:
- [ ] Does the advisor have skin in the game?
- [ ] What do they lose if they're wrong?
- [ ] Are they an operator or a talker?
- [ ] Would they eat their own cooking?
- [ ] Is there symmetry between upside and downside?
- [ ] Is this Lindy (time-tested)?
- [ ] Am I applying via negativa? (avoid harm first)
- [ ] Does this pass the symmetry tests?
Anti-Patterns
- "Trust the experts" β Which experts? Do they have skin in the game?
- "Studies show..." β Researchers don't bear cost of being wrong
- "Best practices recommend..." β Who pays if the practice fails?
- "We're doing this for your benefit" β Via positiva without consequence
- Upside without downside β Someone else is bearing the risk
- "We'll monitor and adjust" β No accountability for the adjustment
- Hidden asymmetries β If you can't see the downside, you're probably holding it
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.