How a result is selected
Wheel and Names obtains a random value through Web Crypto in your browser. Rejection sampling maps it to a bounded integer without modulo bias. That integer maps to an entry in proportion to its weight. The result is chosen before the animation; the wheel displays it.
API reference: W3C Web Cryptography API — getRandomValues.
How weights change the odds
An entry’s probability equals its weight divided by the sum of all entry weights. Weights are whole numbers from 1 to 1,000. Weights of 1, 1 and 2 give probabilities of 25%, 25% and 50%. Weights do not guarantee a specific number of wins in a short series.
Try different weights
This example shows probabilities for one pick. It does not change your wheels.
| Entry | Weight | Chance |
|---|---|---|
| Anna | 25% | |
| Boris | 25% | |
| Sofia | 50% |
Total weight: 4
What repeated names mean
Each line is a separate entry in an ordinary spin. In an equal-weight list of “Anna, Anna, Boris”, Anna has a combined probability of 2/3 and Boris 1/3. If two people share a name, use distinct labels such as “Anna K.” and “Anna M.”. Labels differing in case are treated as different.
How multiple winners are selected
The multiple-winner mode combines exact matching labels and adds their weights. After each pick, that label is removed from the temporary pool before the next pick. A label cannot repeat within one draw. The next selection uses probabilities recalculated from the remaining weights.
Why a result can repeat
Ordinary spins are independent when the entries and weights stay the same. Several Yes results in a row do not increase the chance of No on the next spin. Removal changes the list: an ordinary spin removes the selected line, while a multiple-winner draw removes all lines bearing the selected labels.
Where lists and results are stored
Lists and settings use localStorage in this browser, with the latest 100 results per wheel. There is no account or cross-device synchronization. A share link contains labels and weights and opens an independent copy without history. JSON export keeps a wheel; CSV export keeps result history.
What you can verify
You can inspect the weights, list and result. Wheel and Names does not provide independent auditing, a server draw record or proof that entries were unchanged before the draw. A cryptographic randomness source alone does not establish that a particular giveaway followed its rules.