🔬 physics

Stereotyping by strategy standing diversifies cooperation patterns in indirect reciprocity

This paper proposes a framework of "stereotyping by strategy standing," where individuals base their actions on the overall reputation of a strategy rather than specific recipients, demonstrating that this mechanism diversifies and stabilizes cooperative evolutionarily stable norms in indirect reciprocity that would otherwise be unstable.

Original authors: Ming Wei, Xin Wang, Wenqiang Zhu, Longzhao Liu, Hongwei Zheng, Feng Fu, Shaoting Tang

Published 2026-06-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ming Wei, Xin Wang, Wenqiang Zhu, Longzhao Liu, Hongwei Zheng, Feng Fu, Shaoting Tang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a large community where people decide whether to help each other based on their "reputation." If you've been nice before, people help you. If you've been mean, they might ignore you. This is called indirect reciprocity.

Usually, to make these decisions, you need to know the specific history of the person standing in front of you. Did this person help that person last week? Tracking every single person's history is hard work for your brain. It's like trying to remember the credit score of every single stranger you meet in a city.

The Problem: Too Much Brain Power

The paper asks: What happens if we stop tracking individuals and start using stereotypes? Instead of asking, "What is your reputation?", we ask, "What is the general reputation of the group you belong to?"

In many computer models, researchers force groups to exist by giving people pre-assigned labels (like "Team Red" vs. "Team Blue"). But in real life, groups often form naturally because people act in similar ways. If everyone in a neighborhood is known for being helpful, you might assume a new person from that neighborhood is helpful too, even if you don't know them personally.

The New Idea: "Strategy Standing"

The authors propose a new way to model this. They call it stereotyping by strategy standing.

Think of it like this:

The Resident Group: Imagine a town where everyone follows a specific "rulebook" for being nice or mean. Over time, this town develops a general vibe or "standing." Maybe 90% of the people there are considered "Good."
The Newcomer (Mutant): A stranger arrives. They have their own rulebook.
The Decision: When the stranger needs to decide whether to help a local, they have two choices:
1. Check the individual: "Is this specific local person Good or Bad?" (The hard way).
2. Check the group vibe (Stereotyping): "This town has a 'Good' standing. I'll assume this person is Good too." (The easy way).

The paper introduces a "stereotyping knob" (called $\nu$ ). If the knob is at 0, the stranger checks individual reputations. If the knob is turned up, the stranger relies more on the group's general standing.

The Big Surprise: Stereotypes Can Help Cooperation

Usually, we think stereotypes are bad because they are inaccurate. They might make us trust a bad person just because they are from a "good" group.

However, this paper found a surprising twist: A little bit of stereotyping actually creates more stable ways for people to cooperate.

When the authors turned up the "stereotyping knob," they discovered that the system didn't just stay the same; it unlocked eight new, highly cooperative strategies that were previously unstable.

The "Counterparts of the Leading Eight"

In the world of game theory, there is a famous set of eight rules called the "Leading Eight." These are the gold standard for keeping cooperation alive. They are very strict and effective.

The paper found that when you add stereotyping, you get eight new rules that are almost identical to the original "Leading Eight." The authors call these the "Counterparts of the Leading Eight."

Here is the difference:

The original "Leading Eight" rules are very careful about how they treat "Bad" people meeting other "Bad" people.
The new "Counterparts" are slightly different in how they handle those specific "Bad vs. Bad" situations.
Without stereotyping: These new "Counterparts" fail. They get invaded by the original "Leading Eight" and die out.
With stereotyping: The new "Counterparts" become rock-solid. They can resist the original rules and survive.

Why Does This Happen?

Think of it like a crowded party.

Without stereotypes: Everyone is trying to remember exactly who everyone else is. If you make a tiny mistake in your memory (a "noise" or error), the whole system can get confused, and some good strategies collapse.
With stereotypes: You use the "vibe" of the room. If the room has a "Good" vibe, you assume people are good. This coarse, simple view actually protects the new "Counterpart" strategies from being knocked over by tiny errors. It acts like a safety net.

The Bottom Line

The paper claims that if we allow people to use group-level impressions (based on how a group generally behaves) instead of just tracking every single individual, we actually expand the number of ways cooperation can survive.

It's not that stereotypes are always good; the paper specifically warns that if stereotypes are based on fake labels or bias, they can be harmful. But, if the "stereotype" is just a rough summary of how a group of people actually behaves (their "standing"), it can be a powerful tool that helps society stay cooperative, even when things get messy or noisy.

In short: Sometimes, trusting the "vibe" of a group is a smart shortcut that helps good people stick together, even if it means you don't know every single person perfectly.

Technical Summary: Stereotyping by Strategy Standing Diversifies Cooperation Patterns in Indirect Reciprocity

Problem Statement
Indirect reciprocity explains the evolution of cooperation through social reputations, where individuals condition their actions on the observed reputations of others. While third-order norms (which consider the donor's reputation, the recipient's reputation, and the action taken) can sustain high levels of cooperation (e.g., the "leading eight" strategies), they are cognitively demanding and sensitive to noise. Stereotyping offers a cognitive shortcut by replacing individual-level reputation with group-level information. However, existing theoretical models typically treat stereotyping as an exogenous process based on predefined group labels or structured populations. This leaves an open question: how does cooperation evolve when stereotyping is linked not to external labels, but to group-level impressions that are endogenously generated by the behavioral strategies themselves? Specifically, can relying on the "standing" of a strategy (the stationary reputation level of a population using that strategy) rather than individual reputation stabilize new cooperative patterns?

Methodology
The authors introduce a framework of stereotyping by strategy standing within the context of third-order indirect reciprocity.

Model Setup: The study utilizes an infinitely large, well-mixed population engaging in one-shot donation games. Individuals possess binary reputations (Good/Bad) and adopt behavioral strategies defined by a $2 \times 2$ table of actions based on donor/recipient reputations. Social norms assign reputations based on these interactions.
Strategy Standing: The "standing" of a resident strategy $p$ , denoted $O_d(p)$ , is defined as the stationary reputation level (fraction of good individuals) reached by a population dominated by $p$ under a specific social norm $d$ . This value is independent of initial conditions.
Stereotyping Mechanism: The model introduces a mutation process where rare mutants ( $q$ ) interact with a resident population ( $p$ ). When a mutant acts as a donor, it conditions its action on the recipient's individual reputation with probability $1-\nu$ , but with probability $\nu$ (the stereotyping strength), it conditions its action on the standing of the resident strategy ( $O_d(p)$ ) instead of the recipient's specific reputation.
Analysis: The authors perform an exhaustive search over the full space of social norms and behavioral strategies to identify Evolutionarily Stable Strategy (ESS) pairs $(d, p)$ . They analyze the stability of these pairs against invasion by mutants, varying the stereotyping strength $\nu$ and error rates (execution error $\mu_e$ and assignment error $\mu_a$ ).

Key Contributions and Results

Expansion of Cooperative ESS Pairs: The primary finding is that stereotyping by strategy standing diversifies and expands the set of stable cooperative strategies. As $\nu$ increases from 0 to 1, the number of cooperative ESS pairs increases non-uniformly, rising from 25 (at $\nu=0$ ) to 60 (at $\nu=1$ ). Crucially, a significant number of new stable pairs emerge even at very low levels of stereotyping.
The "Counterparts of the Leading Eight": The study identifies eight new, highly cooperative ESS pairs that appear in the highest performance range (relative payoff $>0.9$ $> 0.9$ ) as soon as $\nu$ $ν$ exceeds a small threshold. These pairs are termed the counterparts of the leading eight.
- They share the exact same social norms as the classical "leading eight."
- They differ from the leading eight only in a single entry of their behavioral strategy: the action taken when a bad donor meets a bad recipient ( $p_{B||B}$ ).
- Without stereotyping ( $\nu=0$ ), these counterparts are unstable because they can be invaded by their corresponding leading strategies. However, once $\nu$ exceeds a critical threshold ( $\nu_c$ ), the counterparts become stable and resist invasion.
Mechanism of Stabilization: The instability of the counterparts at $\nu=0$ is not due to invasion by unconditional defectors (ALLD), which they can already resist. Instead, it is due to competition with their corresponding leading strategies. Stereotyping alters the payoff dynamics such that the specific behavioral nuance of the counterpart (cooperating vs. defecting in the $B||B$ context) becomes advantageous under the influence of group-level cues.
Critical Thresholds: The authors derive asymptotic expansions for the critical stereotyping strength $\nu_c$ required to stabilize each counterpart. They find that $\nu_c$ depends on error rates and payoff parameters ( $b, c$ ). Notably, for some counterparts, increasing the benefit $b$ or assignment error $\mu_a$ can either increase or decrease the required $\nu_c$ , depending on whether the strategy involves cooperation or defection in the $B||B$ context.

Significance and Claims
The paper claims that stereotyping, often viewed as a source of bias or reduced accuracy, can play a constructive role in the evolution of cooperation when it is based on strategy standing.

Behavioral Grounding: Unlike models using exogenous group labels, this framework links stereotypes to endogenous behavioral regularities. It suggests that coarse-grained information derived from population-level strategy performance can serve as a viable informational route to stable cooperation.
Diversification: The results demonstrate that stereotyping does not merely preserve existing cooperation but actively diversifies the landscape of stable norms. It allows for the stabilization of strategies that are otherwise unstable, effectively expanding the "space" of possible cooperative societies.
Cognitive Efficiency: The findings support the idea that when individual-level information is costly or noisy, relying on group-level impressions (strategy standing) can be a cognitively efficient heuristic that sustains high levels of cooperation, provided the group impression is tied to actual behavioral performance.

The authors conclude that while stereotyping can be detrimental when detached from behavior or based on biased beliefs, in the specific context of strategy-standing-based indirect reciprocity, it provides a robust mechanism for stabilizing a broader and more diverse set of cooperative strategies.