Skip to content
WheelSpinny
Menu

How WheelSpinny picks a winner

What happens when you press Spin, what “fair” means here, and what we don’t claim.

The short version

Step by step

  1. Your list becomes slices. Each non-blank line is one slice. Repeated lines are separate slices, which is how you make an option more likely on purpose.
  2. A random number is drawn. We ask your browser for a random number using crypto.getRandomValues, the cryptographically secure generator built into every modern browser.
  3. The number becomes a slice, without bias. Turning a large random number into “one of 7 slices” with simple division would favor some slices very slightly. We use rejection sampling instead: the rare numbers that would cause that bias are thrown away and a new one is drawn, so every slice is exactly equally likely.
  4. The wheel is aimed at the result. We calculate a stopping angle that puts the chosen slice under the pointer, landing somewhere inside the middle 70% of the slice so it never looks like a borderline call. The wheel then spins forward several full turns and slows to that angle.
  5. The result is revealed. The winning slice is highlighted and the result pops up in words over the wheel, and screen readers announce it.

Things that don’t affect the result

Run the fairness test

This runs the exact function the wheel uses to pick a winner, thousands of times, here in your browser, and counts how often each slice wins. Nothing is simulated or sent anywhere.

Spins

Reading the test

A fair wheel doesn’t give every slice exactly the same number of wins. Over 10,000 spins on 6 slices, a gap of 50 or so from the expected 1,667 is normal. What matters is whether the gaps are the size chance predicts, and that’s what the chi-square test measures.

The percentage it reports is how often a perfectly fair wheel would spread its wins at least as unevenly as your run did. Anything above 5% is unremarkable. Even a perfect wheel dips below 5% in about 1 run in 20, so a single low result proves nothing; a real bias would show up run after run, and more clearly with more spins. The test shows the wheel is fair over many spins. It can’t make any single spin “lucky” or “unlucky”.

How each tool picks

Every tool on WheelSpinny draws from the same secure random numbers, with the same bias-free method:

Where a tool animates (a spinning wheel, rows gliding into a new order), the result is always chosen first and the animation is aimed at it.

What we don’t claim

Being clear about the limits matters as much as the method: