Random pick from list

What "pick from list" is good for

It's the simplest tool in the world and yet it's used every day in offices, classrooms, homes and online communities:

• Drawing winners: in a Discord community or a WhatsApp group, paste the names, draw one.
• Deciding what to eat: 10 dishes, pick 1.
• Assigning tasks: 5 people, 5 tasks, shuffle and each takes one.
• Forming groups: 30 people, you want 5 groups of 6. Shuffle and split.
• Picking a movie/book/podcast: your 50-title watchlist, pick one and start.
• Statistical sampling: from 1000 customers, pull 50 at random for a survey.

How it works internally

With "no repeats" on, we apply the Fisher-Yates shuffle algorithm:

for (let i = arr.length - 1; i > 0; i--) {
  const j = Math.floor(Math.random() * (i + 1));
  [arr[i], arr[j]] = [arr[j], arr[i]];
}

Every possible permutation has exactly the same probability of appearing. It's the industry-standard algorithm since 1938 (Ronald Fisher and Frank Yates, hence the name); every shuffle in serious libraries uses it.

The naive shuffle bias

Many people implement shuffle like this: arr.sort(() => Math.random() - 0.5). Incorrect: it produces biased permutations because sort doesn't expect a comparator that returns random values. On large arrays certain elements end up closer to the front than others. If fairness matters (raffles, selection algorithms), use Fisher-Yates.

Best practices for trustworthy raffles

1. Share the list before the draw so everyone can verify their name was on it.
2. Post a screenshot of the result in the group or channel where the draw happened.
3. Run the draw live (stream, Zoom). Eliminates doubts.
4. If there are physical prizes, consider a platform with verifiable timestamp or notary presence. A homemade draw, no matter how well-intentioned, has no legal weight.

Useful variants

The generator supports two modes: pick N items (sample) and shuffle the whole list (return everything in random order). The second mode is ideal for assigning presentation order, pairing people up (shuffle, take in pairs) or reordering a playlist.

Cases where it doesn't work well

• Lists with meaningful duplicates: if you have "John, John, Mary", John has double odds. Want equal-by-name? Dedupe first.
• Weighted selection: if "John" should have more odds than "Mary" (raffles with different ticket counts), this isn't the tool. Multiply entries in the list or use weighted selection.