Related Experiment Videos
Individual variation evades the prisoner's dilemma
Dominic D P Johnson1, Pavel Stopka, Josh Bell
1Faculty of Arts and Sciences, Harvard University, Cambridge MA 02138, USA. dominic@post.harvard.edu
BMC Evolutionary Biology
|September 12, 2002
Summary
The Prisoner's Dilemma (PD) model, used to study cooperation, often fails in real-world scenarios because payoffs are not fixed. This research shows that variable payoffs make valid PD games rare in nature.
Area of Science:
- Evolutionary biology
- Game theory
- Social sciences
Background:
- The Prisoner's Dilemma (PD) is a key model for cooperation studies across various disciplines.
- Traditional PD assumes fixed payoffs, which is unrealistic as payoffs vary due to individual factors.
- Factors like age, status, and preferences influence an individual's utility and payoffs.
Purpose of the Study:
- To investigate the impact of variable payoffs on the validity of the Prisoner's Dilemma.
- To determine the frequency of valid PD games when payoff structures are not fixed.
Main Methods:
- Relaxing the assumption of fixed payoffs in the Prisoner's Dilemma model.
- Analyzing the resulting payoff matrices for validity under variable conditions.
Main Results:
- Relaxing fixed payoffs leads to frequent violations of the required PD structure.
- With payoff variance twice the interval, only 16% of matrices remain valid PD.
- A single player with a non-valid matrix invalidates the PD game, reducing valid games to 3%.
Conclusions:
- The Prisoner's Dilemma is rarely observed in nature due to the prevalence of variable payoffs.
- The study explains the scarcity of empirical evidence for PD despite its theoretical importance.
- Rethinking the fixed-payoff assumption is crucial for understanding cooperation evolution.