Interactive model · Expectation States Theory

Pecking Order Lab

Six agents with hidden, unequal abilities repeatedly pair off and play a coordination game: each round, either one backs down and lets the other lead, both back down at once, or neither does and a costly conflict breaks out. Nobody is optimizing — each agent follows a fixed, psychologically "hardwired" rule for forming expectations about who outranks whom, based on what they've personally experienced and what they've watched happen to others. The paper's central finding: these rules reliably stop conflict, but they only sometimes build a hierarchy that actually tracks ability — there's a tradeoff between sorting quickly and sorting correctly. Adjust the rules on the left and run the group to see where a given rule set lands on that tradeoff.

Sirianni, A. D. (2018). "Expectations and Coordination in Small Groups." Advances in Group Processes, Vol. 35, pp. 181–207. · DOI · source notebook

Three kinds of dial drive every run. Game sets the stakes of a single interaction — how many agents, how many rounds, what's at risk when nobody backs down. Rule is the hardwired automaton every agent uses to decide how to act and update its expectations. Ability controls how talent is spread across the group and how strongly it converts into reward, independent of anyone's behavior.

Real-world scenarios

Each one sets the sliders below to match a familiar situation.

Conflict dynamics

Reward dynamics

Simulation setup

How long the run is — doesn't touch what happens in any one interaction.

Single-game mechanics

The three dials that define one pairwise interaction, independent of who the two agents are.

Payoff over time

Moving average of collective payoff per interaction, against the paper's own bounds: 1/2 for ability-blind coordination, 2/3 for a group that always lets the abler agent lead.

How interactions resolve

Share of each outcome type across the run, in 30 sequential windows.

Resources vs. ability

Final payoff earned by each agent (bars) against their underlying ability (markers). Bars tracking the markers means the hierarchy rewards the more able agent.

Expectation structure

Every pair's final standing: has it settled into a one-sided hierarchy, is it a stand-off, or still rivalrous?

Pecking Order Lab

Payoff over time

How interactions resolve

Resources vs. ability

Expectation structure

The coordination game, for one pair

How conflict resolves

The first 45 interactions, one at a time