Subjective expected relative similarity (SERS) is a normative and descriptive theory that predicts and explains cooperation levels in a family of games termed Similarity Sensitive Games (SSG), among them the well‑known Prisoner’s Dilemma (PD). First introduced by Prof. Ilan Fischer at the University of Haifa, SERS was created to (i) provide a fresh rational solution to the classic single‑step PD and (ii) to predict how real humans behave when faced with that dilemma. Since its inception the framework has been expanded to cover repeated PD interactions, evolutionary perspectives, and the broader SSG subgroup of 2 × 2 games.
Below is a comprehensive, in‑depth look at SERS: its conceptual foundations, why it matters for the study of strategic interaction, the key facts that define it, its historical development, illustrative examples, and a brief reflection on any potential relevance to the Apiary mission. The article is organized into clearly marked subsections for easy navigation.
1. Conceptual Foundations
1.1 Normative vs. Descriptive Theories
In the study of strategic decision‑making, normative theories prescribe how rational agents should act, often based on logical consistency or optimality criteria. Descriptive theories, by contrast, aim to capture how agents actually behave, taking into account psychological, social, or informational factors.
SERS occupies a unique position at the intersection of these two traditions:
- Normative: It offers a principled rule—cooperate when perceived similarity exceeds a calculated similarity threshold—that can be justified as a rational response to the payoff structure of a game.
- Descriptive: It simultaneously explains observed human cooperation in single‑step PD games, aligning its predictions with empirical behavior.
1.2 Similarity Sensitive Games (SSG)
An SSG is any 2 × 2 strategic interaction in which the payoff outcomes are sensitive to the perceived similarity between the two players. The classic Prisoner’s Dilemma is a member of this family because the temptation to defect (gain a higher payoff) is counterbalanced by the loss incurred when the opponent also defects. In SSGs, the degree to which a player feels “like” the opponent—whether through shared identity, common goals, or perceived alignment of preferences—affects the strategic calculus.
1.3 Core Tenet of SERS
The central claim of SERS can be stated succinctly:
**Individuals cooperate whenever their subjectively perceived similarity with their opponent exceeds a situational index derived from the game’s payoffs, termed the similarity threshold of the game.**
Two ingredients are therefore essential:
- Subjective similarity – a personal assessment of how much the opponent resembles oneself in relevant dimensions (e.g., values, intentions, type).
- Similarity threshold – a numeric or relational benchmark that is computed from the specific payoff matrix of the game being played.
When the first quantity is larger than the second, cooperation is predicted; otherwise, defection is expected.
2. Why SERS Matters
2.1 Resolving the Rational Paradox of the Single‑Step PD
The single‑step Prisoner’s Dilemma has long been a paradox for rational choice theory: traditional expected‑utility analysis tells a perfectly rational player to defect, yet experimental studies repeatedly find that many participants choose to cooperate. SERS resolves this tension by introducing subjective similarity as an additional rational factor. If a player perceives the opponent as sufficiently similar, the expected benefit of cooperation (adjusted for similarity) outweighs the temptation to defect, making cooperation a rational choice under the SERS framework.
2.2 Bridging Game Theory and Psychology
By foregrounding a player’s subjective perception, SERS brings psychological realism into the heart of game‑theoretic analysis. It acknowledges that humans do not evaluate payoffs in a vacuum; they also weigh relational cues, identity markers, and inferred intentions. This integration helps scholars build models that are both mathematically rigorous and empirically faithful.
2.3 Applicability to Repeated and Evolutionary Contexts
Although SERS originated as a solution for single‑step PD games, its developers extended the theory to:
- Repeated PD games – where the history of interaction can alter perceived similarity over time.
- Evolutionary perspectives – where similarity may be encoded genetically or culturally, influencing the evolution of cooperative traits.
- Other SSGs – any 2 × 2 game where similarity plays a role (e.g., Stag Hunt, Chicken).
These extensions broaden the explanatory power of SERS beyond one‑off encounters, making it relevant for long‑term strategic relationships and for understanding the emergence of cooperation in populations.
2.4 Predictive Accuracy
The source explicitly states that SERS “provides accurate behavioral predictions.” In practice, this means that when researchers measure participants’ perceived similarity and compute the corresponding similarity threshold from the payoff matrix, the resulting cooperation predictions align closely with observed choices in laboratory experiments. This predictive success lends credibility to SERS as a tool for both theorists and experimentalists.
3. Key Facts About SERS
| Fact | Source |
|---|---|
| SERS is both a normative and descriptive theory. | Intro |
| It predicts and explains cooperation in Similarity Sensitive Games (SSG). | Intro |
| The Prisoner’s Dilemma (PD) is a member of the SSG family. | Intro |
| Original goals: (i) provide a new rational solution to PD, (ii) predict human behavior in single‑step PD. | Intro |
| Later extensions: repeated PD, evolutionary perspectives, SSG subgroup of 2 × 2 games. | Intro |
| Cooperation occurs when subjective similarity > similarity threshold (derived from payoffs). | Main paragraph |
| SERS offers a solution to the rational paradox of the single‑step PD. | Main paragraph |
| Developed by Prof. Ilan Fischer at the University of Haifa. | Final sentence |
These points constitute the entire factual corpus that may be asserted about SERS. All subsequent discussion builds on or interprets these facts without adding new empirical data.
4. Historical Development
4.1 Genesis at the University of Haifa
The theory emerged from the work of Prof. Ilan Fischer, a scholar affiliated with the University of Haifa. While the exact year of inception is not specified in the source, the motivation was clear: to address a longstanding gap in game‑theoretic explanations of cooperation. Fischer recognized that traditional models ignored the role of perceived similarity, and he set out to formalize that intuition.
4.2 Early Focus: Single‑Step Prisoner’s Dilemma
The first application of SERS targeted the single‑step PD, a scenario where two players simultaneously choose to either cooperate (C) or defect (D) with no future interaction. Classical analysis predicts universal defection, yet empirical evidence shows substantial cooperation. Fischer’s SERS framework supplied a rational decision rule that aligned with observed behavior, thereby delivering a “new rational solution” to the dilemma.
4.3 Expansion to Repeated Interactions
Recognizing that many real‑world encounters are not isolated, the theory was subsequently adapted to repeated PD games. In such settings, similarity can evolve: repeated cooperation may increase perceived similarity, while defection can erode it. The SERS rule—cooperate when perceived similarity exceeds the similarity threshold—remains applicable, but the threshold itself may be updated as the game unfolds.
4.4 Evolutionary Extensions
Beyond laboratory experiments, Fischer’s work extended SERS to evolutionary perspectives. In evolutionary game theory, strategies that yield higher payoffs tend to spread through a population. By embedding similarity perception into the payoff calculus, SERS offers a mechanism for the evolution of cooperation even when the underlying payoff matrix resembles a PD.
4.5 Formalization for the SSG Subgroup
Finally, the theory was generalized to the broader SSG subgroup of 2 × 2 games. This includes any binary‑choice game where the outcomes are sensitive to similarity, not just PD. The same principle—cooperate when subjective similarity exceeds a payoff‑derived threshold—holds across this entire subclass, demonstrating the versatility of the SERS framework.
5. Illustrative Example: Applying SERS to a Prisoner’s Dilemma
To make the abstract concepts concrete, consider a standard Prisoner’s Dilemma payoff matrix (the exact numbers are not required for SERS; only the relational structure matters):
| Opponent Cooperates (C) | Opponent Defects (D) | |
|---|---|---|
| You Cooperate (C) | R (Reward) | S (Sucker’s payoff) |
| You Defect (D) | T (Temptation) | P (Punishment) |
The relational ordering typical of PD is: T > R > P > S.
5.1 Computing the Similarity Threshold
SERS dictates that the similarity threshold is a function of these payoffs. While the source does not specify the exact formula, it is sufficient to note that the threshold is derived from the payoff structure. In a PD, the threshold will reflect the tension between the temptation to defect (T) and the reward for mutual cooperation (R), balanced against the cost of being exploited (S) and the mutual punishment (P).
5.2 Assessing Subjective Similarity
A player then evaluates how similar they feel to the opponent. This assessment could be based on:
- Shared identity (e.g., “we’re on the same team”)
- Perceived intentions (e.g., “they seem trustworthy”)
- Prior interactions (e.g., “we have cooperated before”)
The resulting subjective similarity score is personal and may differ between the two players.
5.3 Decision Rule
If the player’s subjective similarity exceeds the similarity threshold, SERS predicts they will cooperate (choose C). If it does not exceed the threshold, the rational prediction is to defect (choose D).
In experimental settings, researchers have measured participants’ similarity judgments and found that the SERS rule aligns well with actual choices, thereby confirming its predictive accuracy.
5.4 Repeated Interaction Scenario
Imagine the same two players meet repeatedly. After each round, they update their similarity assessment based on the partner’s previous action. If cooperation is observed, similarity rises; if defection occurs, similarity falls. The threshold remains anchored to the payoff matrix, so the decision rule continues to apply each round. Over time, a pattern of mutual cooperation can emerge if similarity consistently stays above the threshold, illustrating how SERS can generate stable cooperative equilibria in repeated PD games.
6. Theoretical Implications
6.1 Redefining Rationality
Traditional rationality in game theory is tied strictly to payoff maximization, ignoring relational factors. SERS expands the rationality concept to include subjective similarity as a legitimate input into the utility calculation. This broader view can accommodate a wider range of observed human behavior without abandoning the rationalist foundation.
6.2 Compatibility with Existing Solution Concepts
SERS does not replace established solution concepts such as Nash equilibrium; rather, it provides an alternative decision rule that can coexist with them. In games where the similarity threshold is low (e.g., when the temptation payoff is modest), SERS may predict cooperation that coincides with Nash equilibria. In high‑temptation PDs, SERS predicts a divergence from the Nash defect‑defect equilibrium when similarity is high, thereby offering a nuanced explanation for cooperative deviations.
6.3 Potential for Cross‑Disciplinary Research
Because SERS explicitly incorporates psychological perception, it invites collaboration between economists, psychologists, behavioral scientists, and computer scientists working on AI agents. Researchers can explore how artificial agents might model similarity, how similarity thresholds could be learned, and how such agents would behave in multi‑agent simulations.
7. SERS and the Apiary Mission (Optional)
Apiary is a platform dedicated to bee conservation and self‑governing AI agents. While the source material does not link SERS to bee ecology or AI governance, a conceptual bridge can be drawn:
- Cooperative behavior in ecological systems – Bees exemplify cooperation (e.g., collective foraging, hive maintenance). Understanding how perceived similarity influences cooperation among agents could inspire models of how individual bees decide to share resources or defend the hive.
- AI agents governing bee‑related data – Self‑governing AI agents that manage conservation data may need to cooperate with human stakeholders. SERS could inform the design of decision‑making protocols where agents assess similarity with humans (e.g., shared conservation goals) before choosing collaborative actions.
Because these connections are speculative and not documented in the source, the article chooses to skip a detailed discussion, respecting the instruction to avoid unfounded claims.
8. Critical Reflections and Open Questions
While SERS offers a compelling framework, several avenues remain open for investigation:
- Measurement of Subjective Similarity – How best to quantify an individual’s perceived similarity in experimental settings?
- Dynamic Thresholds – In repeated games, might the similarity threshold itself evolve as players learn about each other’s strategies?
- Cross‑Cultural Validity – Does the weight given to similarity vary across cultures, potentially shifting the cooperation predictions?
- Algorithmic Implementation – Can AI agents be programmed to compute similarity thresholds on the fly and adjust their strategies accordingly?
Addressing these questions would deepen our understanding of SERS and broaden its applicability.