Learn & Review: How Decision Making is Actually Science: Game Theory Explained

Game Theory: Understanding Social Interactions

This summary explores the field of game theory, a branch of mathematics and science that analyzes social interactions and decision-making. It covers the core concepts, its applications, and its two main branches: non-cooperative (competitive) and cooperative game theory.

Main Idea

Game theory provides a framework for understanding situations where individuals' outcomes are influenced by the decisions of others. It offers tools to analyze both competitive scenarios, aiming for individual advantage, and cooperative scenarios, aiming for fair distribution of benefits.

Key Concepts and Branches

• Definition of a Game: In game theory, a "game" is any interaction between multiple people where each person's payoff is affected by the decisions of others. This extends beyond traditional games to encompass most social interactions.
• Pioneering Figure: Game theory was significantly developed by mathematician John Nash in the 1950s.
• Wide-Ranging Applications: It is used by economists, political scientists, biologists, military tacticians, psychologists, and many others.
• Two Main Branches:
  • Non-Cooperative (Competitive) Game Theory: Focuses on competitive interactions with clear winners and losers.
  • Cooperative Game Theory: Deals with situations where players agree to work together towards a common goal.

Non-Cooperative Game Theory: The Prisoner's Dilemma

The Prisoner's Dilemma is a famous thought experiment illustrating competitive game theory.

• Scenario: Two prisoners, Wanda and Fred, are arrested.
  • If one confesses and the other stays silent, the confessor goes free, and the silent one gets 10 years.
  • If both confess, they both get 5 years.
  • If neither confesses, they both get 2 years.
• The Dilemma: The prisoners are separated and must decide independently. They have no reason to trust each other.
• Rational Decision: From an individual perspective, confessing is the best strategy regardless of what the other person does.
  • If the other confesses, confessing leads to 5 years instead of 10.
  • If the other stays silent, confessing leads to freedom instead of 2 years.
• Nash Equilibrium: This is a key concept where each player makes the choice that benefits them most, no matter what their opponent decides. In the Prisoner's Dilemma, both confessing is the Nash Equilibrium.
  • This outcome is stable because neither player can improve their situation by unilaterally changing their decision.
• Outcome: The logical outcome is that both Wanda and Fred confess, resulting in a suboptimal outcome for them collectively (5 years each) compared to if they had both stayed silent (2 years each). This highlights how individual rationality can lead to a worse collective result in competitive situations.

Cooperative Game Theory: Fairness and Distribution

Cooperative game theory addresses how to fairly divide gains or costs among players working together.

• Coalition: A group of players in a cooperative game.

• Key Question: How much should each player contribute and how much should they benefit?

• Shapley Value: This is the central concept in cooperative game theory, providing a method for dividing outcomes based on individual contributions. It is guided by several axioms:

  1. Marginal Contribution: A player's value is determined by what is gained or lost by their removal from the game.
     • Example: If a group bakes 50 fewer cookies when one member is absent, that member's marginal contribution is 50 cookies.
  2. Interchangeable Players: Players with the same skills or contributions should have equal value and receive equal rewards/pay the same costs.
     • Example: Two people ordering the same meal at a restaurant should pay the same amount.
  3. Dummy Players: A player who contributes nothing has zero value and receives nothing.
     • Note: This axiom can be controversial, as situations like maternity leave or disability might warrant consideration beyond direct contribution.
  4. Additivity (Decomposition): Costs or payments should be broken down across different parts or time periods of the game. Rewards should reflect contributions made at different times or for different tasks.
     • Example: If a player works hard on Monday but slacks off on Tuesday, their rewards should differ accordingly.

• Shapley Value Calculation (Conceptual): It involves calculating a player's marginal contribution to every possible subgroup of other players and then averaging these contributions.

  • Example: Two bakers, one making 10 cookies/hour alone and the other 20 cookies/hour alone, produce 40 cookies together.
    • If the first baker (10 cookies/hr) is removed, the second baker's contribution to the total is 30 cookies (40 total - 10 from the first).
    • If the second baker (20 cookies/hr) is removed, the first baker's contribution to the total is 20 cookies (40 total - 20 from the second).
    • The Shapley value suggests averaging these marginal contributions: (30 + 20) / 2 = 25. So, the second baker (who is more productive alone) gets $25, and the first baker gets $15 (assuming cookies sell for $1 each).

Conclusion

Game theory offers valuable insights into decision-making:

• In competitive situations, it helps individuals strategize to maximize their own benefit.
• In cooperative situations, it provides methods for ensuring fairness and equitable distribution of outcomes.