Model Library
Evidence So Far tracks how confident the site is in each result. This page is the complementary map: every simulation model on the site, organized by family, with what each one tests and a link to go read it. Use it to find a specific model rather than to judge how conclusive it is.
How to Read This Page
Each table has an Adaptation column that says what kind of process the model runs:
- Evolution — selection on an inherited trait or strategy across generations, with no within-lifetime learning.
- Learning — within-lifetime adaptation only, with no evolutionary component.
- Evolution + Learning — both act together, which is the site's central two-timescale claim.
Nowak Mechanisms
Five thin wrappers over a shared Moran-process engine, one per classical mechanism for the evolution of cooperation identified in Nowak (2006).
| Model | Adaptation | What it demonstrates | Explore |
|---|---|---|---|
| Kin Selection | Evolution | Cooperation spreads when benefit routed to relatives, weighted by relatedness, exceeds its private cost — Hamilton's rule rB > C. | Kin Selection |
| Direct Reciprocity | Evolution | Cooperation sustained by repeated encounters: a cooperator can reward or punish the same partner next round, but only if they meet again. | Direct Reciprocity |
| Indirect Reciprocity | Evolution | Cooperation sustained by reputation: agents help others whose public reputation signals they are good cooperators, without needing repeated encounters. | Indirect Reciprocity |
| Network Reciprocity | Evolution | Cooperation sustained by spatial structure: local-neighborhood interaction lets cooperator clusters form and shield each other from exploitation. | Network Reciprocity |
| Group Selection | Evolution | A second level of selection on top of individual competition: the most successful group is periodically copied into the least successful group. | Group Selection |
Ecological Models
Three case studies that embed a cooperation problem inside a spatial, ecological setting rather than an abstract Moran process. (Spatial Altruism, once a fourth case study here, moved to Evolved Cooperation's Foundational Replications — it is a faithful implementation of one named paper rather than an original scenario built for this site; see that page's own explanation.)
| Model | Adaptation | What it demonstrates | Explore |
|---|---|---|---|
| Cooperative Hunting | Evolution | Whether an inherited, continuously-valued hunt-investment trait can spread when it costs the individual predator but pays off only through successful shared hunts. | Cooperative Hunting |
| Spatial Prisoner's Dilemma | Evolution | What inherited response rules spread when agents repeatedly play a local Prisoner's Dilemma and reproduce into nearby empty cells using accumulated energy. | Spatial Prisoner's Dilemma |
| Retained Benefit | Evolution | The most abstract case study: strips away mechanism-specific structure to ask how much of the value cooperation creates has to be routed back to cooperators for it to spread. | Retained Benefit |
Learning × Selection Interaction Models
The models that test the site's central claim directly: that within-lifetime learning and between-generation selection can work better together than either alone.
| Model | Adaptation | What it demonstrates | Explore |
|---|---|---|---|
| PredPreyGrass | Evolution + Learning | A richer predator-prey-grass ecology used to test whether the genome needs a direct channel into individual behavior for selection-driven behavioral evolution to become visible. | PredPreyGrass |
| Model 1: Trust Learning | Evolution + Learning | Baseline two-timescale model: agents learn a scalar trust value per partner within a lifetime, and evolution selects the inherited parameters that shape that learning. | Model 1 |
| Model 2: Q-learning | Evolution + Learning | Same two-timescale framework as Model 1, but with action-value (Q-learning) reinforcement learning replacing scalar trust. | Model 2 |
| Model 3: Extended | Evolution + Learning | Adds reputation, partner choice, and forgiveness on top of Q-learning, shifting the outcome toward conditional cooperation with active monitoring. | Model 3 |
| Network Diversity Experiment | Evolution + Learning | Not a new model — runs Models 1 through 3 under systematically varied network conditions to test when partner-memory alone stops being enough. | Network Diversity |
Learned Cooperation Models
The Prisoner's Dilemma and Repeated Prisoner's Dilemma pages set up the game-theoretic background; the PPO Study is the model that actually runs it.
| Model | Adaptation | What it demonstrates | Explore |
|---|---|---|---|
| PPO Study | Learning | Two independently trained PPO agents repeatedly play a Prisoner's Dilemma, testing whether learned behavior converges to persistent defection, cooperative conventions, or mixed horizon-dependent patterns. | PPO Study |
Foundational Replications
From-scratch replications of specific named papers, kept separate from this site's own original models. Three established the two-timescale claim this site builds on (Evolution + Learning); the rest are faithful implementations covering pure evolution and pure learning — not every one of them reproduces its paper's headline claim, and that's recorded here too rather than only in each page's own status section. See Evidence So Far for the cross-cutting status summary.
| Model | Adaptation | What it demonstrates | Explore |
|---|---|---|---|
| Ackley & Littman (1991) | Evolution + Learning | A neural-network-controlled creature that senses, moves, eats, fights, and reproduces, testing whether combining an evolving genome with within-lifetime reinforcement learning beats either alone. This site's strongest confirmed result. | Ackley & Littman (1991) |
| Hinton & Nowlan (1987) | Evolution + Learning | The needle-in-a-haystack landscape that gave the Baldwin effect its modern computational footing: whether within-lifetime learning can manufacture a gradient where genetic search alone has none. Includes an interactive demo. | Hinton & Nowlan (1987) |
| Prosser (2022) | Evolution + Learning | Tests what changes when learning acts on genetically correlated groups of traits, with the correlation structure itself evolving, instead of on independent single loci. | Prosser (2022) |
| Ackley & Littman (1994) | Evolution | The same authors' follow-up, with learning dropped entirely: whether costly, honest signaling can evolve and stabilize with no built-in reward, using only population viscosity and communication's naturally short reach to make signalers and listeners relatives. | Ackley & Littman (1994) |
| Spatial Altruism | Evolution | Mitteldorf & Wilson's (2000) population-viscosity model: whether an altruistic inherited type can survive local exploitation once reproduction is local and empty space is a competitor. See its own page for how the same result also reads as a group-selection and a kin-selection argument. | Spatial Altruism |
| Leibo et al. (2017) | Learning | Whether independent reinforcement learners dropped into spatial, temporally extended social dilemmas still produce policies that behave like a classical matrix-game dilemma once cooperation is measured at the policy level, not the action level. | Leibo et al. (2017) |
| Boyd & Richerson (2009) — Culture | Evolution | Whether culture can sustain a far lower migration/selection ratio than genetics, preserving the between-group variation that group selection needs. Reproduces, plus an independently confirmed refinement to the paper's own homogenization threshold. | Boyd & Richerson (2009) — Culture |
| Boyd & Richerson (2009) — Voting With Your Feet | Evolution | Whether payoff-biased migration reliably spreads group-beneficial behavior. Reproduces the paper's own claim, plus a further finding — checking welfare directly, not just behavior frequency — that the paper's own conclusion glosses over. | Voting With Your Feet |
| Hughes et al. (2018) | Learning | Whether an intrinsic aversion to unequal outcomes (inequity aversion) raises cooperation in Cleanup and Harvest. Does not currently reproduce the paper's headline Cleanup claim; Harvest is inconclusive. | Hughes et al. (2018) |
| McKee et al. (2023) | Learning | Whether an intrinsic reputation motivation — the learned analog of indirect reciprocity — produces less territorial, more turn-taking behavior when agents are identifiable versus anonymous. Training currently collapses before cleaning behavior emerges in either condition, so the claim can't yet be tested. | McKee et al. (2023) |