Ultimatum game

Ten points on the table, ten rounds. You play the proposer: each round you pick a split of the 10 points between yourself and a scripted responder, who can only accept — the split happens — or reject, in which case you both get nothing that round. The responder's acceptance rule is on display before you start, so the real question is how much you think a fair-minded counterpart will tolerate. It helps to know the baseline first: in the ultimatum game, one player proposes how to split a sum of money and the other can only accept or reject — rejection means both walk away with nothing. Across hundreds of experiments, proposers most often offer an equal split, average offers sit around 40-50% of the pie, and offers below roughly 20% are rejected so often that acceptance rates approach zero.Convergent / replicated

The strangeness sits on the responder's side of the table. Turning down a low offer costs the responder real money, and textbook game theory predicts people should accept any positive amount while proposers offer the bare minimum. Instead, people routinely reject lowball offers, and brain-imaging work links those rejections to emotional reactions — evidence that responders compare their share to the proposer's and react to unfairness, not just to their own payoff.Supported, with limits As proposer, you are bargaining with that fact: every offer you make is a bet on how much unfairness the other side will swallow before punishing you both.

What might look like a quirk of laboratory students turns out to be far broader — though not identical everywhere. Fairness behavior in this game is a human regularity but not a single universal constant. In experiments across 15 small-scale societies, the pure self-interest model failed everywhere, yet offers and rejections varied across groups more than in earlier research, with differences tracking features of everyday economic life such as market integration; a separate meta-analysis likewise found regional differences, specifically in responder behavior.Supported, with limits The station above gives that regularity a shape you can feel: try the 9–1 split against the default responder and watch what a "rational" lowball actually earns.

One round of this game does not measure a person's generosity or sense of fairness as a stable trait. Offers and rejections shift with the size of the stakes, how choices are collected, players' experience, and whether roles were earned rather than assigned — and even the game's inventor notes that individual behavior in it still cannot be predicted from observable characteristics.Supported, with limits So read your session summary as a record of what you tried against one scripted rule, in one sitting, for points — not as a verdict on what kind of person you are. The site draws no such conclusion.

The ultimatum game was introduced in experiments run at the University of Cologne and published in 1982 by Güth, Schmittberger and Schwarze. The finding that people systematically deviate from the standard money-maximizing prediction influenced a generation of researchers and made this one of the most replicated paradigms in behavioral economics, with thousands of follow-up experiments.Historical / descriptive

One caveat matters especially here. Playing against a disclosed computer opponent is not the same as facing a person. When people knew offers were computer-generated, the bodily stress response that normally accompanies unfair offers — and predicts rejecting them — appeared only for offers from real humans, not from computers. A version of the game with an openly simulated responder can demonstrate the logic, but it should not be expected to evoke or measure the full social reaction.Supported, with limits Treat a session here as a way to explore the game's structure from the proposer's chair; where the research is silent — including exactly how a simulated responder changes the proposer's side of the experience — we make no claims. The evidence base is overwhelmingly laboratory experiments, with the cross-cultural fieldwork described above the notable exception.