Why Monkey Republic needed adjustable worlds

What we tried

The inspiration is Nicky Case's wonderful Evolution of Trust, which turns the prisoner's dilemma into a playable story. In that setup, cooperation creates extra value — and under that assumption, trust-friendly strategies naturally win. Beautiful. But it's one world.

Monkey Republic keeps the playfulness (monkeys, bananas, share or grab) and adds the missing dimension: the world's assumptions are adjustable.

• Four starter worlds. A Win-Win Jungle where sharing grows extra bananas; a Zero-Sum Grove where one monkey's gain is another's loss; a Dark Forest where betrayal hits hard; a Forgiving Village where broken trust can heal.
• One shared brain, eight personalities. Every monkey — the saint, the con artist, the fair-minded copycat, the grudge-holder, the pragmatist, the follower — runs the same decision engine. Personality is just six dials: how much you trust strangers, how fast trust moves, how you forgive, how much gossip matters, how much payoffs teach you, how much you copy the rich. No scripted villains anywhere.
• Kids design a personality, not a speech. Students tune the same six dials for their own monkey, drop it into the society, run dozens or hundreds of rounds, and read what came out. Then they argue, retune, and run again.

What surprised us

The design decision that ended up carrying the whole lesson was making born villains scarce — the cast mirrors what behavioral economists actually find in people: mostly conditional cooperators, a few unconditional givers, a few free-riders, and a slice of pragmatists who go whichever way the wind pays. Which means when a harsh world turns dark, it's not because the thieves multiplied. It's because the pragmatists converted. The world didn't hire more thieves; it retrained them. Kids can watch that happen to a monkey they named.

The other surprise: the moral questions arrive on their own. Nobody has to prompt "if the world is dangerous, should we give up on trust?" — a child asks it, out loud, staring at a Dark Forest run. The simulation doesn't answer it. That's the point.

What children can learn from it

Not "what strategy wins?" but the deeper pair: what kind of world makes this strategy win — and what kind of world do we want to help build? A child who has watched the same monkey flourish in one world and go cold in another owns an idea most adults were never taught with evidence: strategies are shaped by worlds, and worlds are partly made of our choices.

Open questions

• Calibration. We have target behaviors — a long Win-Win run should end trusting, a Dark Forest defensive, a grudge-holder should go permanently cold after one betrayal — and the dials get tuned against real classroom runs, not our guesses.
• How much history to show. The society records everything (share rates, betrayals, repairs); the open design question is which chart makes a ten-year-old gasp, not just nod.