Trigger Strategies in Repeated Games

The Ultimate Sanction: The Grim Trigger Strategy

The simplest and most severe way to enforce an agreement is with a "nuclear option." In game theory, this is known as the Grim Trigger strategy. It is a conditional strategy that is easy to state:

1. Start by cooperating.
2. Continue to cooperate as long as your partner cooperates.
3. If your partner ever defects (breaks a rule), defect forever.

The punishment is total and permanent. This strategy's power lies in its terrifying finality. To understand its mechanics and the conditions under which it works, the following video provides a clear, foundational explanation.

As the video demonstrates, the viability of Grim Trigger depends entirely on the discount factor, $\delta$. This variable represents how much a player values future payoffs relative to present ones. A high $\delta$ means a player is "patient" and cares a great deal about the long-term health of the relationship. A low $\delta$ means they are "impatient" and prioritize immediate gratification.

Scenario: The Unbreakable Rule

Let's ground this in a concrete hotwife scenario. Chloe (the hotwife) and her husband, David, have a relationship agreement. One of their most critical rules is an absolute prohibition on unprotected sex with any outside partner (a "bull"). This is a hard boundary.

They implicitly or explicitly agree on a Grim Trigger enforcement:

• Cooperation: As long as Chloe uses protection, the lifestyle continues, bringing shared enjoyment (let's say a payoff of 3 for each per encounter).
• Defection: If Chloe has unprotected sex even once, she gains a higher immediate payoff from the heightened intimacy (payoff of 4), but David is emotionally devastated and his trust is shattered (payoff of 1).
• Punishment: The trigger is pulled. David permanently withdraws his emotional support and active participation in the lifestyle. The shared fun ends, and they revert to a baseline, less-fulfilling state (a perpetual payoff of 2 for each).

Chloe's decision at any given moment is a tradeoff. She can defect now and get a one-time payoff of 4, but this is followed by an infinite stream of 2s. Or, she can stick to the agreement and receive a perpetual stream of 3s. As the video's math shows, she will stick to the agreement only if her discount factor $\delta$ is greater than or equal to 1/2. If she values the future of her relationship with David enough, the threat of the Grim Trigger is sufficient to ensure the rule is followed.

The "Grudger" strategy, visualized in the image below, is another name for Grim Trigger. Notice how after an initial period of cooperation (green circles), a defection occurs. From that point on, the payoffs for both players remain negative or low (red circles), leading to a much lower total score than if cooperation had been maintained.

Strategic Errors When Facing the Grim Trigger

Playing against a Grim Trigger strategy is simple in theory but can be challenging in practice. An experiment where human subjects played a Prisoner's Dilemma against a computer programmed to use Grim Trigger reveals common mistakes people make.

Let's apply these errors to our scenario:

• Cooperating after Defection (CaD): Chloe breaks the rule. David triggers the punishment and withdraws. The next week, Chloe says, "I'm sorry, I'll follow the rule from now on." If David is truly using a Grim Trigger strategy, this is pointless. The punishment phase is permanent. Attempting to "re-cooperate" is a strategic error based on a misunderstanding of the agreement's severity.
• "Sniping" (DaC): Chloe plans one last date with a bull she never intends to see again. She thinks, "I'll break the rule this one time. The damage to my relationship with David won't start until after this date is over, and I'll have already gotten the thrill." This is an attempt to capture the temptation payoff while trying to outrun the consequences, a risky gamble that underestimates the long-term cost of the punishment phase.

Forgiveness and Reciprocity: Tit-for-Tat and Its Variants

The Grim Trigger is a powerful tool, but its inflexibility makes it brittle. What if a rule is broken by accident? What if there's a misunderstanding? A single misstep, even an unintentional one, leads to total collapse. In the real world, a more forgiving strategy is often more robust.

The most famous of these is Tit-for-Tat (TFT).

1. Start by cooperating.
2. In each subsequent round, do whatever your partner did in the previous round.

If they cooperate, you cooperate. If they defect, you punish them by defecting in the next round, but—and this is the crucial difference—you are willing to return to cooperation as soon as they do. It punishes, but it also forgives.

Scenario: The Flexible Rule

Let's revise Chloe and David's agreement. The rule is now: "Be home by 1 AM after a date." This is a softer boundary than the previous one. The enforcement mechanism is Tit-for-Tat.

• Cooperation: Chloe comes home on time. David is happy and supportive (Payoff: 3, 3).
• Defection: Chloe stays out until 3 AM. She has extra fun (Payoff: 4), but David is anxious and upset (Payoff: 1).
• Punishment (TFT): For the next scheduled date, David punishes by being emotionally distant and unenthusiastic (e.g., he doesn't help her pick an outfit, doesn't want to hear about it afterwards). This reduces Chloe's enjoyment of the next date. If she then respects the curfew on that date, he returns to being fully supportive for the one after.

The punishment is immediate and reciprocal, but not permanent. It allows the relationship to recover from minor infractions. Interestingly, for the payoffs used in the video, TFT is also an equilibrium as long as $\delta \ge 1/2$.

Leniency: Thriving in a "Noisy" World

Real relationships are full of ambiguity, or what game theorists call "noise." Was the rule broken intentionally, or was it a mistake? "I stayed out late because my car broke down." "I didn't answer your text because my phone died." A simple TFT strategy might punish these accidental "defections," leading to a needless and destructive cycle of retaliation.

Successful real-world strategies are often more sophisticated. They are lenient (slow to anger) and forgiving (fast to forgive). Given your background in stochastic processes, you'll appreciate that these strategies are designed to be robust in an environment where signals are imperfect.

The academic paper "Slow to Anger and Fast to Forgive" by Fudenberg, Rand, and Dreber explores a variety of such strategies.

The choice of a trigger strategy is a crucial part of designing a relationship agreement.

• Grim Trigger is best for bright-line, inviolable rules (e.g., health and safety boundaries), where a single violation destroys the foundation of trust.
• Tit-for-Tat is effective for clear, simple rules where reciprocity is desired and mistakes are unlikely.
• Lenient strategies (like TF2T or Grim2) are arguably the most practical for the majority of relationship rules. They build trust by giving the benefit of the doubt while still retaining the power to punish persistent or deliberate violations. They account for the "noise" of real life and prevent the relationship from spiraling into conflict over misunderstandings.

Conclusion

In this lesson, we have moved from abstract cooperation to the concrete mechanics of enforcement. You've seen how trigger strategies provide the teeth that make relationship agreements credible and self-enforcing.

Key Takeaways:

• Trigger strategies are pre-committed plans for responding to rule violations (defections) in a long-term relationship (repeated game).
• Grim Trigger is the most severe punishment strategy, enacting permanent defection after a single violation. It is powerful but brittle.
• Tit-for-Tat is a reciprocal strategy that punishes defection immediately but is also quick to forgive, allowing for recovery from transgressions.
• Leniency and Forgiveness are key features of successful strategies in real-world "noisy" environments, preventing destructive conflict spirals caused by misunderstandings.
• The effectiveness of any trigger strategy hinges on the players' patience ($\delta$). The threat of future punishment must outweigh the temptation of a present gain from defection.

Preview of the Next Lesson:

We have now spent considerable time on the dynamics of interaction within an established partnership. But how do these partnerships form in the first place? In our next lesson, we will shift gears to begin the module on Cooperative Games and Stability. We will start by modeling the partner selection process itself, using a classic tool from cooperative game theory: the Gale-Shapley algorithm for stable matching. This will provide a formal framework for analyzing how stable pairings emerge in a market of potential partners.