Commons Collective / interactive research
Simulation Lab.
Different institutions. Shared constraints.
Explore what changes and what stays the same.
Synthetic & uncalibrated. These are illustrative assumptions, not empirical estimates, forecasts, or evidence that a collective will succeed. Seeded results are conditional on the selected model. No real people, money, or AI services are involved.
Loading the local simulation controls. If this message persists, reload the page or check that JavaScript and module workers are supported.
Plan the institution / protect the member
Capital before scale.
A staffed startup plan, from three people to ten. What must be funded before membership fees can carry the institution?
Planning calculator, not a completed agent simulation. The member ramp is an explicit assumption, not demand generated by spending or evidence that investment guarantees growth. Defaults below are a deterministic cohort projection, calculated locally. No Monte Carlo trials run until you request them.
Read the funding study, assumptions, and example plans · 10,000 design trials and 4,000 independent-seed validation trials; conditional results, not a market forecast.
View the published trial tables, distributions and downloadable records.
Operating-agent stress simulation below A separate experiment; its inputs and outcomes do not inherit this finance plan.
2. Inspect funding & member value
Deterministic assumption-based projectionCapital needed before fees arrive
Protects the full selected planning horizon, not only the target month: the worst pre-revenue cash trough, plus a rolling operating reserve and contingency, not just the final cash balance. Committed capital is a source of cash, never operating revenue.
Different meanings of break-even
Cash break-even starts the first six consecutive nonnegative cash-profit months. Meeting the target also requires the final six months to stay nonnegative. Run-rate profit removes expansion acquisition/onboarding and hiring, smooths events, and retains replacement acquisition/onboarding.
Planned cash & funding gaps
Read the monthly tableBelow zero means a funding gap, not a business allowed to keep spending negative cash. After the first shortfall, subsequent activity is an unfunded hypothetical plan.
Sources & uses through the target month
Investor capital bridges the net cash gap; it does not pay every cost alone. Member fees also finance uses, including the commons share. Provider payments are part of the member price, not institution revenue or an institution expense here. Reserve and contingency are funding buffers, not additional expenses in this table.
Compare uses through projected break-even & the full horizon
These secondary windows may differ from the chosen target. “Not reached” means no durable operating cash break-even occurs in the selected horizon; it is not a zero-cost outcome.
Monthly plan: membership, staffing, income, costs, cash & savings
All monetary columns are USD. Member price includes the membership fee. Member saving compares that all-in price with the individual subscription assumption. “Unfunded plan” rows do not represent executed activity. Month 0 contains external launch setup; first-month hiring and payroll are separate.
Read the assumptions & accounting boundary
A paid cohort starts only when the minimum launch size and after-fee saving requirement are met. Otherwise membership and fees are zero while the staffing and fixed-cost plan continues. This price check assumes equivalent service, no member-paid switching cost, and a provider able to serve the cohort. It does not test retention behavior, service quality, willingness to pay, or real customer acquisition.
Payroll is fully loaded, not take-home pay. Setup excludes employee payroll. The projection omits taxes, debt, interest, investor-return requirements, and a financing transaction schedule. Potential capital recovery is a cash-capacity check, not a scheduled distribution or guaranteed investor repayment.
3. Stress the assumptions, not the promise
Optional seeded Monte CarloTrials sample triangular member-scale, correlated-cost, and churn assumptions, plus a uniform integer launch delay. A conditional share is not a real-world probability of success. Capital ON/OFF never changes the underlying growth or spending plan.
Module workers are unavailable. The deterministic planning calculator and plan download still work; Monte Carlo requires worker support.
Not run. No Monte Carlo results.
No backend or API keys. Explicit money-view sharing uses one-time storage in this tab, removed at the destination; nothing is uploaded. Exported monetary results use integer cents; inputs use USD dollars. Percentiles can interpolate between outcomes.
Maximum 1 MiB. Loads a compatible finance projection’s validated configuration and recalculates it locally; saved result numbers are never trusted. Loading clears trials and restores the trial count to 100, without automatically running Monte Carlo. Operating scenarios and runs belong in the separate operating importer below.
Observe a trajectory
Not startedNo model loaded
Run starts a fresh trajectory. Step starts or advances a paused trajectory by one tick. Editing a scenario stops the worker and clears its results. No simulation starts on model selection.
Local engine identity pending.
Outcomes, not predictions
Committed worker snapshotsYour results begin here. Run or step the selected scenario; no outcomes are calculated in advance.
Trajectory
Equivalent chart data table
| Tick | Operator cash (USD) | Quality (%) | Participants | Commons work |
|---|
Domain metrics & accounting balances
| Metric | Value | Unit |
|---|
Keep money, rights, and quality separate
Operator cash is not member welfare. The purchasing cost difference compares modeled purchasing costs with direct purchasing and includes outstanding buyer liabilities, not just paid cash. Positive values mean lower modeled cost; late payments do not create savings. Restricted commons funds and contributor agency balances are not unrestricted operating income. Permission is a gate, not a price: more budget does not create contribution or repair authority.
Quality is the observed fulfilled/attempted task rate, not a promise. Read it beside denied and unmet demand: lower spending can accompany reduced service, and cost differences alone do not establish greater benefit. A missing quality rate means no valid denominator, not zero quality.
Compare the arrangements
A small local experiment, not the full research study. Compare A0–A3 within the selected domain and F/P choices. Paired arrangements use the same seed for each replication. No jobs start until you press a comparison button.
Larger populations, horizons, and sweeps can take several minutes. Progress and Cancel remain available; reduce the horizon or population for a quick first experiment.
2–30; default 5. Small samples give unstable Monte Carlo estimates. These errors do not measure uncertainty about the real world.
One configuration
Hold the assumptions fixed and compare all four arrangements.
5 replications × 4 arrangements = 20 trajectories
One-dial sweep
Choose a parameter to see its units and allowed range.
3 settings × 5 replications × 4 arrangements = 60 trajectories
No comparison results. Each comparison replaces the active trajectory and uses a fresh worker.
Inspect the event trail
Latest 40 eventsNo committed events yet.
| Tick | Type | Event |
|---|---|---|
| Run or step a scenario to inspect its events. | ||
Save, load & verify
Versioned JSON · local filesDownloads are independent records and survive model resets. There is no automatic scenario persistence. Explicit financial-view sharing uses one-time storage in this tab. Save before switching or leaving.
The summary replays and verifies the completed source, then aggregates the selected ledger window. Explicit sharing temporarily stores the source in this tab and removes it at the destination. No data is uploaded.
A scenario saves inputs only. A completed-run export is available only after a successful trajectory; it is never synthesized from partial or batch results.
No file selected.
Import validates file type, version, and identity before replacing the whole scenario. Nothing executes automatically.
Replay regenerates the seeded world in a worker and checks its final digest against the saved run. Pause and step are available during replay. Run scenario instead reruns only the imported inputs, without claiming replay verification. Unsupported scenario/engine versions and completed-run build mismatches are rejected rather than silently migrated.