NATS1595 13 - Mixed Strategies
NATS1595
Course Title: Iterated Games and Mixed Strategies
Introduction to the Iterated Prisoner’s Dilemma
Background: Focus on understanding cooperation through experiments.
Key Question: What sorts of experiments have people conducted to explore cooperation?
Axelrod’s Tournament
Competition Overview: In 1980, Professor Robert Axelrod held a competition to analyze strategies in the Iterated Prisoner’s Dilemma.
Game Mechanics:
Iterative play: Players repeatedly engage in the game.
Simultaneous moves: Both players decide whether to Cooperate (C) or Defect (D) at the same time.
Strategy influenced by previous moves.
Example of Rounds in the Iterated Game
Round Description:
Round 1: C vs C -> (3,3)
Round 2: D vs C -> (8,1)
Round 3: C vs D -> (1,8)
Round 4: C vs C -> (6,6)
Final Score Observations: Reveals the outcomes based on chosen strategies and consequences based on player actions.
Winning Strategy Identified
Best Strategy: Tit-for-Tat (TFT)
Mechanism:
Start with Cooperation on the first move.
In subsequent moves, mimic the previous action of the opponent (Cooperate or Defect).
Character of the Strategy: Considered a "nice" strategy.
Cooperation Findings
Emerging Conclusion: Cooperation may lead to better outcomes in certain situations.
Mixed Strategy Nash Equilibrium
Concept Introduction: What if there is no dominant strategy?
Implication of playing the game iteratively suggests mixing strategies prevents exploitation.
Strategy Consideration: How often and when to alternate between different strategies?
Understanding Expected Value
Definition: The expected value is the average outcome based on probabilities of different possible results.
Formula: E(X) = x1 * p1 + x2 * p2 + ...
Example Using a Die:
If repeatedly rolling a 6-sided die, consider what the average value will be over many rolls.
Practical Example of Expected Value with a Spinner
Spinner Average Moves: 1.75 moves per turn based on given outcomes.
Probabilities in Simple Experiments
Shirt Color Example:
Probability of wearing blue shirts: 0.4
Probability of wearing red shirts: 1 - 0.4 = 0.6
Detailed Mixed Strategy Nash Equilibrium
Condition: A 2x2 payoff matrix lacking a pure strategy leads to a mixed strategy Nash equilibrium.
Strategy Probabilities:
Player 1 selects strategy A with probability q, and B with 1-q.
Player 2 selects strategy A with p, and B with 1-p.
Example Payoff Calculations
Dynamic Payoff Insights:
If Player 2 plays A with probability p and B with 1-p, predict expected payoffs based on Player 1's strategy.
Equalize payoffs to determine mixed strategies for optimal play.
Hawk and Dove Concept
Behavioral Strategy:
Hawks: Aggressive, escalate conflicts resulting in possible injuries.
Doves: Avoid conflict, seldom lead to injury.
Strategy Outcomes in Hawk and Dove Scenarios
Common Resource Competition: Outcomes depend on behavior (Hawk versus Dove) and expected payoffs (Benefit B vs Cost C).
Dominant Strategy Dynamics: If B > C, both players behaving as Hawks leads to dominant strategy identified through Nash equilibrium.
Population Dynamics of Hawks and Doves
Probability of Encounters: Awareness of how prevalent each behavior is in a population.
Varies between 100% probability to lower percentages in mixed populations.
Implications of Increasing Hawk Population**
Assess the impact on strategy as the prevalence of Hawks in the population rises.