Prisoner's Dilemma
A classic problem showing how individual rationality can lead to collective irrationality.
Payoff Matrix
Choose Your Strategy
Nash Equilibrium Analysis
In Prisoner's Dilemma, Defect is always your best response regardless of what your opponent chooses. Therefore (Defect, Defect) is the unique Nash equilibrium. However, if both cooperated, each would get 3 instead of 1. This is an inadequate equilibrium - the system is stuck in a suboptimal state.
Stag Hunt - Coordination Game
A game showing how coordination problems can lead to suboptimal equilibria.
Payoff Matrix
Choose Your Strategy
Multiple Nash Equilibria
Stag Hunt has two Nash equilibria: (Stag, Stag) and (Rabbit, Rabbit). The former gives higher payoff (payoff-dominant), but the latter is less risky (risk-dominant). If you're uncertain your opponent will hunt stag, hunting rabbit is safer. This explains why systems can get stuck in suboptimal equilibria.
Population Dynamics
Watch how strategies evolve over generations in a population playing repeated games.
Strategy Distribution
Evolution Over Time
Understanding Inadequate Equilibria
What are Inadequate Equilibria?
An inadequate equilibrium is a situation where a system gets stuck in a suboptimal state because no individual has the incentive to change their behavior. Even though everyone would be better off if they could coordinate to change, the individual incentives maintain the status quo.
1. Individual Rationality ≠ Collective Optimality
What's rational for each individual can lead to poor outcomes for the group. The Prisoner's Dilemma demonstrates this perfectly.
2. Coordination Problems
Even when everyone wants to reach a better equilibrium, they may not be able to coordinate. The Stag Hunt shows how fear of being the 'sucker' can trap people in suboptimal choices.
3. Path Dependence
Historical accidents can lock systems into inadequate equilibria. Once a system is in place, switching costs and network effects maintain it.
Real-World Examples from the Book
Hospital Mortality Statistics
Hospitals don't publish mortality data because doing so would make them look bad compared to others. If everyone published, patients could make informed decisions and overall healthcare would improve. But no individual hospital has incentive to be the first to publish potentially unfavorable data.
Academic P-Values
The scientific community is locked into using p < 0.05 as the threshold for statistical significance, despite known flaws. Better alternatives exist, but journals, reviewers, and researchers are all incentivized to maintain the status quo. Anyone deviating risks publication and career advancement.
'Red-Haired Entrepreneur' Superstition
If a successful entrepreneur happens to have red hair, people might incorrectly associate red hair with entrepreneurial success. This superstition persists because: (1) it's costly to test, (2) counter-examples are dismissed as exceptions, and (3) conforming signals you understand the 'culture.'
QWERTY Keyboard Layout
The QWERTY keyboard layout was designed to prevent typewriter jams, not for optimal typing speed. Better layouts exist, but switching costs and network effects keep QWERTY dominant. Manufacturers, teachers, and typists all have incentives to maintain the status quo.
When to Defer to Expertise vs. When to Question Systems
Defer to Expertise When:
- The domain has strong feedback mechanisms and error correction
- Experts have skin in the game (personal consequences for being wrong)
- There's active competition and innovation
- You're significantly less knowledgeable than experts
- Example: Trusting engineers about bridge design
Question Systems When:
- There are weak feedback mechanisms or long delays
- Experts face different incentives than outcomes
- The field is stagnant with little innovation
- You notice systematic failures that persist
- Example: Questioning academic publication practices
How to Escape Inadequate Equilibria
1. Coordinate Simultaneous Action
Get everyone to agree to switch at the same time. This requires trust and communication channels.
2. Change Incentives Structure
Modify payoffs so the desired equilibrium becomes individually rational. Regulations or rewards can help.
3. Reduce Switching Costs
Make it easier for individuals to try alternatives. Lower barriers enable exploration.
4. Create Critical Mass
Build enough adoption of the alternative that network effects start working in your favor.
Further Reading
- "Inadequate Equilibria" by Eliezer Yudkowsky The primary source for these concepts
- "The Evolution of Cooperation" by Robert Axelrod Classic work on Prisoner's Dilemma and Tit for Tat
- "Thinking Strategically" by Avinash Dixit and Barry Nalebuff Excellent introduction to game theory