How you govern a shared resource
A commons is anything a group shares and can use up: a pasture, a fishery, a groundwater basin — or the trust, attention and goodwill inside an online platform. Everyone benefits if it lasts; everyone is tempted to take a little more than their share. Here is the whole problem, and the three ideas the lab lets you test.
1 · The tragedy of the commons
If each person acts only for themselves, the sensible-for-me choice is to take a bit more — the cost of overuse is shared by everyone, but the extra catch is all mine. When everyone reasons that way, the resource is stripped faster than it can regrow, and it collapses for all. In the lab, set the quota to the full appetite and turn enforcement off: the pool crashes to zero within a few rounds and cannot recover.
2 · Ostrom’s way out (rules that actually hold)
The economist Elinor Ostrom won a Nobel Prize for showing the tragedy is not inevitable — real communities sustain commons for centuries with a handful of design principles. The lab gives you four of the levers:
- A clear boundary / quota — a known, bounded rule for how much each person may take. The single most important lever: the total harvest must stay under what the resource can regrow.
- Monitoring (enforcement) — the chance an over-quota harvest is actually noticed. Rules nobody checks are just suggestions.
- Graduated sanctions (penalty) — a proportionate cost when someone is caught, aimed at the excess, not a life sentence.
- A reason to cooperate (reward) — stewardship has to pay, or only the takers stay.
Deterrence in the model is exactly enforcement × penalty + reward: raising any of them can only make the greedy less likely to defect — never more — so the pool can only get healthier. Find the combination that keeps a whole season above the collapse floor.
3 · Goodhart’s law (when the metric becomes the target)
“When a measure becomes a target, it ceases to be a good measure.” A platform that rewards one number — raw activity, posts-per-day, harvest-this-round — gets exactly that number, gamed. In the lab, “chase the metric” books a huge first-round harvest (the dashboard looks great) and then razes the commons, so the real goal — value that lasts the whole season — comes out far behind the quieter, stewarding rules. A proxy is not the thing you actually wanted.
Why it’s a MEASUREMENT, not a verdict
Every villager has a hidden greed the simulation draws once and never changes; the numbers you see (harvest, welfare, times caught) are consequences of your rules meeting those appetites. A reputation or karma score in a real system works the same way — it measures how a model of behaviour responds to the rules, not a person’s worth. Design the rules, run the season, read the result, then change one lever and run it again.
Honest note: this teaches the social science of rule design — mechanism design, Ostrom’s commons, Goodhart’s law, agent-based modelling. It is a model of a community, not a claim about real people, and not a claim that playing it changes anyone’s behaviour.