Evolutionary dynamics in repeated optional games

Fang Chen, Lei Zhou, Long WangarXiv 2022 · DOI · arXiv

Abstract

Direct reciprocity facilitates the evolution of cooperation when individuals interact repeatedly. Most previous studies on direct reciprocity implicitly assume compulsory interactions. Yet, interactions are often voluntary in human societies. Here, we consider repeated optional games, where individuals can freely opt out of each interaction and rejoin later. We find that voluntary participation greatly promotes cooperation in repeated interactions, even in harsh situations where repeated compulsory games and one-shot optional games yield low cooperation rates. Moreover, we theoretically characterize all Nash equilibria that support cooperation among reactive strategies, and identify three novel classes of strategies that are error-robust, readily become equilibria, and dominate in the evolutionary dynamics. The success of these strategies hinges on the effect of opt-out: it not only avoids trapping in mutual defection but also poses additional threats to intentional defectors. Our work highlights that voluntary participation is a simple and effective mechanism to enhance cooperation in repeated interactions.

Introduction

In evolutionary game theory, social dilemmas such as the Public Goods Game (PGG) illustrate the tension between individual gain and collective benefit. While direct reciprocity—the idea that individuals can sustain cooperation through repeated interactions (e.g., Tit-for-Tat)—is a well-known solution, most models assume that participation is mandatory. However, real-world social and economic interactions are frequently voluntary. If a partnership becomes exploitative, individuals often have the option to walk away. This study investigates whether the freedom to opt out in a repeated setting provides a stable path to ubiquitous cooperation, moving beyond the unstable "rock-paper-scissors" cycles often seen in one-shot optional games.

Methods: The Repeated Optional PGG Framework

The researchers developed a mathematical framework for repeated optional Public Goods Games. In each round, $n$ individuals choose between three actions: Cooperation (C), Defection (D), or Opting-out (O). Cooperators pay a cost to provide a benefit shared by all participants, while defectors exploit these contributions. Loners (those who opt out) receive a fixed "safe" payoff ($\sigma$), which is designed to be better than the outcome of mutual defection but less than the rewards of mutual cooperation.

Comparison of compulsory, one-shot optional, and repeated optional Public Goods Games (PGGs).

The study focuses on reactive strategies, where an individual's action in the current round depends on the actions of others in the previous round. Using Markov chain modeling and evolutionary simulations (pairwise comparison in a well-mixed population), the authors analyzed the long-term stability and success of these strategies compared to standard compulsory and one-shot models.

Key Results and Strategy Evolution

The introduction of the "opt-out" option in a repeated setting leads to a dramatic shift in evolutionary outcomes. While compulsory games often struggle to maintain cooperation rates above 20% in harsh conditions, the repeated optional version achieves nearly 100% cooperation.

Evolutionary dynamics and strategy analysis showing the dominance of cooperation in repeated optional games.

As shown in Figure 4a, the cooperation rate for repeated optional PGGs quickly stabilizes at a high level, far outperforming both one-shot optional and repeated compulsory models. The success of this mechanism relies on the opt-out option acting as a "reset button." In compulsory games, groups can become trapped in mutual defection. In optional games, individuals can exit to receive the safe payoff, breaking the deadlock and allowing cooperation to re-emerge.

The researchers identified three dominant strategy classes: OD-TFT, DO-OFT, and OD-TFT$_C$. These strategies are "error-robust," meaning they can recover cooperation even after accidental defections. By strategically switching between opting out and defecting (as seen in the strategy probabilities in Figure 4c), these strategies make exploitation unprofitable for others, effectively using the threat of exit as a more potent punishment than simple defection.

Conclusion

This work demonstrates that voluntary participation is a powerful catalyst for cooperation. By allowing individuals to "vote with their feet," the repeated optional game framework avoids the pitfalls of forced interaction. The identified reactive strategies dominate the evolutionary landscape, occupying the population over 96% of the time. Ultimately, the "right to exit" appears to be a fundamental mechanism for fostering stable, cooperative societies, even in environments where cooperation would otherwise fail.

Summary updated

Abstract

Direct reciprocity facilitates the evolution of cooperation when individuals interact repeatedly. Most previous studies on direct reciprocity implicitly assume compulsory interactions. Yet, interactions are often voluntary in human societies. Here, we consider repeated optional games, where individuals can freely opt out of each interaction and rejoin later. We find that voluntary participation greatly promotes cooperation in repeated interactions, even in harsh situations where repeated compulsory games and one-shot optional games yield low cooperation rates. Moreover, we theoretically characterize all Nash equilibria that support cooperation among reactive strategies, and identify three novel classes of strategies that are error-robust, readily become equilibria, and dominate in the evolutionary dynamics. The success of these strategies hinges on the effect of opt-out: it not only avoids trapping in mutual defection but also poses additional threats to intentional defectors. Our work highlights that voluntary participation is a simple and effective mechanism to enhance cooperation in repeated interactions.
BibTeX
@misc{chen2022evolutionary,
  author = {Fang Chen and Lei Zhou and Long Wang},
  title = {{Evolutionary dynamics in repeated optional games}},
  year = {2022},
  eprint = {2210.08969},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2210.08969},
}
Figure