You get $10 and your partner gets nothing. Whatever you send them is tripled on the way, and they decide how much to send back. They are under no obligation to send anything.
TrustRisk toleranceFairness
Unofficial practice. Not the real Pymetrics game and not connected to
Pymetrics, Harver or any employer.
Untimed,
unlimited retries, nothing saved and nothing sent anywhere.
How the Pymetrics money exchange 1 game works
You start each round with $10. Your partner starts with nothing.
Choose how much of your $10 to send, anywhere from nothing to all of it.
What you send is tripled before it arrives. Send $4 and they receive $12.
Your partner then sends back however much they want, which can be nothing.
You rate how fair they were, on a scale of 1 to 10.
Three rounds, a different partner each time. Nobody sees how the last round went.
There is no clock. Keyboard: arrow keys to set the amount, Enter to send.
What the money exchange 1 game measures
The headline is simple: how much of your $10 you hand to a stranger who has no obligation to give any of it back. Your first send matters most of all, because it is the only decision you make before seeing anybody behave. After that it watches whether you move. Sending less after a stingy partner is the ordinary response, sending the same regardless is a different profile, and so is opening up after generosity. The fairness rating is not decoration either. Getting $6 back on a $4 send is half the pot and half again on what you risked, and whether you call that fair depends on which of the two you are measuring against.
Or play all 12 in one sitting
This page is the money exchange 1 game on its own, practice only.
The full run is the rehearsal. All 12 back to back in the order the real
assessment most commonly uses them, timed, one attempt each, and no feedback until the end.
About 35 minutes, and nothing is saved, so it needs one sitting.
Answered straight, including the parts people would rather hear differently.
There is no correct amount, and any practice site that gives you a number is teaching you to fail the real thing. Employers do not score this against a right answer. They compare your pattern against a model built from people already doing the job, so a large send is a match for a role built on collaboration and a poor one for a role built on guarding resources. Play it the way you would actually play it.
Not reliably. In the original experiment this game comes from, people returned slightly less on average than they had been given, which means trusting usually costs a little. Sending nothing is often the better-paying choice and still not the better answer, because the assessment is not reading your balance.
No, and they are not in the real assessment either, which describes an AI player. Here the three partners are a generous one, an average one and a stingy one, dealt in a random order so a replay is not measuring what you remembered. Which one you meet first is luck, not judgement.
Because otherwise there would be no reason to send anything. Tripling is what turns the decision into a real one: every dollar you send creates two that did not exist, so the pair are better off together than apart, but only you take the risk of finding out.
Both, and it cannot separate them on its own. Sending money to a stranger is a gamble as well as an act of trust. The real assessment pulls the two apart by watching the balloon and cards games alongside this one, which is why the twelve games are scored together rather than one at a time.
This is the Trust Game from Berg, Dickhaut and McCabe, published in 1995 and replicated ever since, which is what the real Pymetrics version is built on. The $10, the tripling and the 1 to 10 fairness scale are documented. The two things nobody publishes are how many exchanges you play and how the partner behaves. Three rounds is our reading of a game every source puts at two to three minutes, and the partner's behaviour is taken from the original experiment's own results rather than invented.
Pymetrics and Harver are trademarks of their respective owners and are not affiliated with
OpenThinkr. Practice results are for interest and self-reflection, not assessment.
