Guides/Pick randomly without bias

How to pick randomly without bias

A fair pick means every eligible option has the same chance. Humans are bad at “random” — we favour the middle of a list, familiar names, or the last idea we heard.

What “unbiased” means here

• One draw from N distinct options → each should have probability 1/N.
• No silent favourites, no “I’ll skip the ones I don’t like” after promising fairness.
• Duplicates on the list count as separate entries only if you intend weighted picks.

Fair methods

1. Numbered list + random integer: assign 1…N, draw uniformly in that range.
2. Digital picker: paste the list into a tool that samples with equal weight (browser Math.random-style picks are fine for chores and games).
3. Physical: identical slips in a hat; shake; one draw — still watch for uneven slip sizes.

Biases to avoid

• Pointing at a screen — fingers drift to the centre.
• “Eeny meeny” — rhythm length often lands on a preferred spot.
• Re-rolling until you like it — that is preference, not randomness.
• Uneven list cleaning — removing options after seeing the pick.
• Bad shuffle UX — always taking the top item after a weak shuffle can still look patterned if the shuffle is broken; use a dedicated pick, not “sort and grab first.”

Worked examples

1. 5 restaurants: each should be 20%. Drawing “Ava’s” twice in a week is normal — fairness is about process, not short-run balance.
2. Who presents first (6 people): list all six once; one uniform pick. Do not let volunteers go first “unless someone objects.”
3. Weighted pick: if one option should be twice as likely, put it twice on the list (or use explicit weights) — and say so out loud.

Pick fairly online

Use the freeRandom List Picker— paste options, one equal-chance result in your browser. For checklists of tasks (not random order), see theChecklist Generator. Running a prize draw?Fair giveaway list draws.