How Weighted Random Selection Works
Use relative weights responsibly and calculate the displayed chance for every option.
Published · Reviewed
How Weighted Random Selection Works
Weighted selection assigns each option a positive share of a common total. An option with weight 3 among weights 3, 1 and 1 receives 3/5 of the probability, or 60%. The numbers are relative and do not have to add to 100.
Recommended process
- Reject zero, negative and non-numeric weights.
- Add all valid weights to obtain the total.
- Display each option’s weight divided by that total.
- Draw one secure fraction and locate its cumulative weight interval.
Worked example
For Walk|2, Museum|1 and Cafe|1, the total is 4. Walk covers half the cumulative range, while Museum and Cafe cover one quarter each. Changing Walk to 4 would make its chance 4/6, not 4%.
What to check before using the result
- Explain why unequal chances are appropriate.
- Show percentages before the spin.
- Do not disguise a preference as an equal draw.
How the random step is handled
Random Picker Tools uses window.crypto.getRandomValues() for browser-based selection. Integer results use rejection sampling to avoid modulo bias, and complete list orders use a Fisher–Yates shuffle. This addresses the software selection step; the organizer still controls eligibility, duplicate rules, consent, accessibility and any legal requirements.
Privacy note
Normal picking and shuffling happen in the browser. Use initials or non-sensitive identifiers when possible, avoid storing unnecessary source data, and create a share link only when temporary server storage is appropriate for the people represented in the list.