Psychological experiment

Chance lab

Each trial uses Web Crypto with rejection sampling. Short runs can be uneven; streaks do not predict the next result.

Experiments
Experiments

Chance lab

Test a forecast

n = 0

10—
20—
How it works

Each trial uses Web Crypto with rejection sampling. Short runs can be uneven; streaks do not predict the next result.

Test a forecast

Choose an outcome and probability before a fair trial. Compare with what happens.

Your choice
50%

1/2 ≈ 50%

Chance lab
How it works

Record a testable event before its deadline. Score only resolved outcomes: (p − outcome)². Lower is better; this does not prove causation.

Atlas191

How to install

History stays in the current browser or app storage. Use export/import to move it.

Offline access

Only public pages and assets for the selected language are cached. Personal entries stay separate.

Settings

There is an unsaved result or text. Reload this page?

Atlas191

Menu