Générateur de binômes

What the random pair generator does

This tool takes a list of names and matches them into pairs. It shuffles the full list once using a Fisher-Yates shuffle backed by the browser's cryptographic random source, then walks through the shuffled list two names at a time to form pairs. If the list has an odd number of people, the last leftover name is either grouped into a trio with one of the pairs or flagged on its own, depending on which option you choose.

When to use it

Pairing tools are useful for coffee-chat matchmaking at a company, assigning lab or study partners in a class, setting up peer code review buddies, or matching mentors with mentees for a session. If you actually need larger groups instead of two-person pairs, use the random team generator, which supports any group size. If your goal is closer to a gift exchange where each person is anonymously assigned to give to another, read the caveats below before assuming this tool covers that case, and consider the general fairness background in how random number generators work.

Handling an odd number of people

With 15 people, for instance, a straightforward pairing produces seven pairs and one person left over. This tool handles that in one of two ways: it can either place the leftover person into one pair to make a trio of three, or it can list them separately as "no partner this round" so you can manually pair them with a facilitator or skip them for one cycle. Neither approach is more "correct" than the other — pick whichever matches your event. For a trio, remember the group now has three-way dynamics rather than a clean one-to-one match, which matters if the activity assumes exactly two participants (like a two-person interview exercise).

Worked example

Take a team of 9 for a coffee-chat rotation: Alex, Bianca, Carlos, Dana, Elliot, Farah, Grant, Hana, and Ivan. Shuffling all nine and pairing sequentially might yield: (Farah, Alex), (Ivan, Dana), (Bianca, Grant), (Hana, Carlos), with Elliot left over. Choosing the "trio" option would instead fold Elliot into one existing pair, for example making it (Farah, Alex, Elliot), so that all nine people are included in exactly one group this round.

Secret-Santa-style limitations

It's worth being explicit that a simple pairing is not the same as a Secret Santa-style derangement. A derangement guarantees a one-directional assignment where nobody is matched to themselves and, depending on the rules, nobody gives to the person who gives to them. This tool's pairing is symmetric and undirected — it just groups two names together — so it does not by itself prevent the two people in a pair from being a "mutual pair" (both directions matched to each other), and it has no concept of self-pairing risk since names are drawn from the same shuffled list without replacement. If you specifically need directional gift-exchange assignments with no self-matches, that requires a dedicated derangement algorithm, which this tool does not perform; treat this generator as a plain, symmetric matchmaking tool instead.

Rotating pairs over multiple weeks

1. Run the shuffle and record the resulting pairs for week one, exporting them as a .csv so you have a record.
2. Before week two, if you want to avoid repeat pairings, manually remove or note previous partners; the tool does not automatically track pairing history across sessions.
3. Re-run the shuffle for week two on the same list — since each run is independent, a repeat pairing is possible unless you actively exclude previous matches yourself.

Fairness note

Every possible pairing of the list is equally likely to occur because the list is shuffled once with Fisher-Yates driven by crypto.getRandomValues() before being split into twos. This does not mean any single run is "100% random" in some absolute sense — it means the process does not favor any name or position, so no participant is more likely than another to end up with a particular partner.