MERITOCRAT.ORG / RESEARCH
Experimental Research Tool · Under Construction

The Sandbox

A proposed many-agent virtual economy for asking the question that paper models cannot answer alone: what happens when the parameters start interacting?

Candidate tool · design open for public input

Plow the model. Sow the parameters. Watch what grows.

The Sandbox is envisioned as an experimental companion to PBS/MPS research: a GUI-controlled virtual world populated by artificial economic agents. Researchers would be able to vary assumptions, run competing rules side by side, and observe the resulting history as a compact narrative event stream rather than an expensive graphical world.

Animated under-construction illustration of a farmer plowing and sowing a field
Research status: proposed interface and experiment environment. This page does not claim that the simulator, behavioral assumptions, or PBS/MPS itself has been empirically validated.
Feature ballot

What should the first public build contain?

These are proposed feature blocks, not promises. Consider them a living ballot for economists, developers, modelers, and curious critics.

Up for vote

World Parameters

Population, resources, productivity, wealth, scarcity, market depth, mobility, shocks, institutional rules, inheritance, taxation, transaction costs, and the initial distribution of opportunity.

Up for vote

Agent Endowments

Innate strengths, learned skill, accumulated knowledge, starting resources, inherited advantage, risk tolerance, patience, social ties, and changing economic status.

Up for vote

Behavior & Incentives

Cooperation, competition, reciprocity, opportunism, trust, moral commitments, incentive sensitivity, bounded choice, and stochastic agency. “Free will” would be an explicit modeling assumption, not a claim about human nature.

Up for vote

Bias & Fairness Stress Tests

Researchers could introduce disparate-treatment, inherited-advantage, and protected-attribute bias scenarios as controlled adversarial variables, then measure system responses. These controls are for detecting and studying unequal treatment, never for asserting racial hierarchy or intrinsic group worth.

Up for vote

Provenance Ledger

A simulated evidentiary history of contribution, identity, timing, context, transformation, and derivative relationships. The point is to test what can be reconstructed, what gets lost, and where false precision begins.

Up for vote

Performance Oracle

Adjustable outcome metrics, measurement noise, validation strength, benchmark drift, oracle disagreement, and gaming pressure. Different worlds can disagree about what “performance” even means.

Up for vote

Contribution Decomposition

Experiment with marginal contribution, persistence, interaction effects, Shapley-style allocation, approximate attribution, counterfactual rules, and uncertainty. Compare allocations instead of treating one estimator as truth.

Up for vote

PBS Value Laboratory

A parameter workbench for exploring candidate forms of V(PBS), including provenance depth, provenance reliability, output yield, output consistency, demand signals, liquidity assumptions, and contribution-allocation layers.

Up for vote

Institution & Market Lab

Switch between royalty, licensing, cooperative, profit-sharing, bargaining, market, and other candidate arrangements. The Sandbox should be able to test PBS/MPS against simpler alternatives rather than assume PBS/MPS wins.

Up for vote

Adversarial Mode

Contribution stuffing, collusion, metric gaming, oracle capture, strategic fragmentation, evidence gaps, identity attacks, and institutional capture become experiments instead of footnotes.

Up for vote

Narrative World Console

A Zork-style event stream would translate simulation state into readable economic history: firms form, contributors cooperate, prices move, shocks arrive, disputes appear, allocations change, and institutions adapt.

Up for vote

Experiment Package

Planned download area for scenario files, parameter presets, experiment manifests, reproducible seeds, run summaries, and comparison outputs so a result can be challenged instead of merely admired.

Why developers may care

A laboratory for failure, not a demo built to win

The useful version of this sandbox should make it easy to produce worlds where PBS/MPS performs well, poorly, or not at all. Developers could use it to regression-test attribution rules, economists could compare institutions under identical starting conditions, and critics could publish parameter sets that break the model.

Sensitivity analysisCounterfactual comparisonMonte Carlo runsAgent heterogeneityEconomic shocksDistributional outcomesAnti-gamingReplication

Suggested output

YEAR 12 · NORTH VALLEY

Three cooperatives merge after drought losses. A high-persistence irrigation heuristic created in Year 3 remains active in 61% of surviving farms. The current allocation changes materially when inherited-capital advantage is removed from the counterfactual. The run is flagged for sensitivity review.

Public design process

Economists: please argue with the knobs.

The parameter list, behavioral assumptions, outcome measures, and comparison institutions should be contestable in public. If a variable is badly defined, missing, ethically inappropriate, or impossible to identify, that is useful information. The goal is not to hard-code an ideology. It is to make assumptions visible enough to challenge.

Get involved

When you leave a message, tell us what you would change first. If you found the breadcrumb, include TURN THE SOIL. It will tell us you read past the sliders.

Download area

Sandbox package

Under construction. No simulation package is being represented as available on this placeholder page. The eventual public package is proposed to include documented parameters, scenario presets, reproducible experiment definitions, and result exports.

Research and proposed framework. Not legal, financial, or investment advice. Simulated agents and simulated outcomes do not establish real-world causation, moral desert, legal rights, or investment value.