AtlasReasonThe 2-4-6 rule game

Games · free, no sign-up

The 2-4-6 rule game

2, 4, 6 fits a hidden rule. Test your own triples, prove the rule on five new ones, and see how many of your tests were built only to confirm your guess.

In short

The 2-4-6 game hides a rule about triples of numbers. You know one example that fits, 2, 4, 6, and you may test any triple you like before naming the rule. The trap is to test only triples you expect to fit: in this game every hidden rule is broader than the obvious guess, so those triples always come back yes and never show that the guess is too narrow.

The game runs in your browser, and your browser is not running the script for this page. The method and the sources below still apply.

Nothing you type on this page leaves your browser tab: no answer, guess or score is sent anywhere or stored.

How the game works

A rule is picked at random from a set of seven. Every one of them is satisfied by 2, 4, 6 and by any other triple that goes up by two, so the obvious guess is always too narrow rather than wrong. You type a triple and the page says whether it fits. Before each check you can mark whether you expect it to fit; that mark is what the result reads back. When you think you know the rule, the page gives you five new triples to label as fitting or not. Getting all five right is the proof: the five are chosen so that at least two of them are triples where the real rule and the obvious guess, numbers going up by two, give different answers.

What the result shows

The result counts your own tests and how many you expected to fit. Peter Wason introduced this kind of task in 1960 and reported that many people announced a rule after testing only cases that confirmed it. Klayman and Ha later showed that testing cases you expect to fit is often a sensible strategy; it fails in exactly the situation this game sets up, where your guess sits inside a broader rule, because then no confirming test can ever come back no.

What it cannot tell you

One game is one sample of how you tested one guess. It does not measure a general tendency to seek confirmation, and a person who plays several rounds learns the game quickly. Treat the result as a demonstration you can feel, not as a score.

Sources

Text, questions and code on this page are original to MyTestAtlas. The cited studies describe the classic procedure; none of their items is reproduced.