Group organizer
Random Pair Generator
Paste a list of names and split everyone into random pairs — with handling for odd numbers and a repeat-avoidance search across rounds.
Where Partner Rounds Come Up
Lab pairs, reading buddies
Classroom partners
Assign lab partners, reading buddies, or project pairs without accusations of favoritism — and without the same two friends always ending up together, thanks to repeat-avoidance across rounds. The adult running the room chooses the odd-number rule first, then generates in front of the group.One rotation per sprint
Pair programming and peer review
Rotate programming partners or reviewer assignments each sprint. The round history keeps rotations genuinely rotating instead of collapsing back into habitual pairs, as long as the roster stays the same between rounds.Sparring, chess, debate
Practice matches and drills
Sparring partners, tennis or chess practice matches, debate drills — generate a fresh round of matchups every session, with a clear rule for the odd person out.Several short rounds
Icebreakers and networking
Pair up strangers for one-on-one conversations at meetups, workshops, or team offsites — run several short rounds so everyone meets several new people.
5 Rounds With 6 Students, Round by Round
Settings
- Roster
- Priya, Marco, Aisha, Tom, Lena, Sam
- Group size
- 6
- Avoid repeating partners
- On
Rounds run by the tool’s own search
- Round 1: Lena & Marco · Sam & Aisha · Priya & Tom — no repeated partners.
- Round 2: Tom & Lena · Marco & Sam · Aisha & Priya — no repeated partners.
- Round 3: Tom & Aisha · Sam & Lena · Priya & Marco — no repeated partners.
- Round 4: Sam & Tom · Marco & Aisha · Priya & Lena — no repeated partners.
- Round 5: Marco & Tom · Aisha & Lena · Sam & Priya — no repeated partners.
After all rounds
- Rounds with zero repeats
- 5 of 5
- Different partnerships seen
- 15 of 15
This is one fixed-seed run of the same shuffle and 300-attempt search the tool uses, so the page shows the same rounds on every visit; your own rounds will differ. Once all 5 rounds are clean, every student has met each of the other 5 exactly once, which is why a further round could not avoid a repeat. With 7 students, one pair each round becomes a trio, and the trio members count as having partnered with each other for repeat-avoidance. With “One person sits out” selected, one clearly-labeled student skips each round instead — useful when the activity genuinely cannot work with three, like a two-player game.
How Fast Pairings Multiply, and How Often Clean Schedules Survive
| Group | Distinct pairings | Clean rounds possible | Sessions finishing all of them |
|---|---|---|---|
| 4 people | 3 | 3 | 100 of 100 |
| 6 people | 15 | 5 | 75 of 100 |
| 8 people | 105 | 7 | 75 of 100 |
| 10 people | 945 | 9 | 13 of 100 |
| 12 people | 10,395 | 11 | 0 of 100 |
How Repeat Avoidance Really Works
The feature that separates this from a basic shuffle is round memory. With "Avoid repeating partners" enabled, the tool records every pairing you generate in the current session. Each time you click "New Round," it tries up to 300 fresh shuffles looking for an arrangement where nobody is matched with a previous partner, and picks the best one it finds. It stops early the moment a clean arrangement turns up.
There is a hard mathematical ceiling: an even group of n people has at most n − 1 repeat-free rounds, the same structure as a round-robin tournament schedule. The tool does not build such a schedule, though. It chooses every round at random and cannot look ahead, so an early round can leave the group in a position where no clean round remains. The table above shows the effect: groups of four always get through and groups of six or eight usually do, while for 10 people only 13 of 100 simulated sessions completed every clean round. In practice, expect a run of clean rounds first and repeats later, and expect repeats sooner as the group grows.
When a round does contain repeats, the tool does not fail silently: it shows exactly how many repeated pairings the round contains, so you can decide whether to accept them or reset the history and start fresh. The pool of possible pairings grows quickly, from 3 for four people to 654,729,075 for 20, so if the same matchup does come up twice in a row for a small group, that is ordinary probability at work, not a broken shuffle.
Odd Groups: Trio or Sit-Out
- Group of three is the right default for most collaborative activities — discussions, projects, peer review — where a trio works nearly as well as a pair and nobody is excluded.
- Sit out fits strictly two-person formats: chess matches, debate rounds, interviews. The person who sits out is simply the last name of that round’s shuffle, so the choice is unbiased, but it is a fresh draw every round: nothing rotates the role, and with 7 people the chance that the same person sits out twice in a row is 14.3%. If a fixed rotation matters, note who sat out and remove them from the next round’s list.
- For recurring sessions, consider alternating modes: trio mode for discussion days, sit-out mode for testing days. The label on each result card always tells you which group is the three or who sat out, so nothing is ambiguous when you export.
Pairing Rules and Export Details
- Best for
- Assigning partners for lab work, peer review, practice matches, and conversation rounds, especially when the same activity repeats and partners should change.
- What you enter
- A text box of names separated by new lines or commas, a duplicate-removal checkbox (off by default, ignores capitalization), an "Avoid repeating partners" checkbox (on by default), and an odd-number choice (group of three by default).
- What you get
- One card per pair, a labelled trio or a dashed "Sits out this round" card when the group is odd, a round counter, and a repeat-count note when a clean arrangement was not found.
- How pairs are formed
- A Fisher–Yates shuffle, then names are taken two at a time from the shuffled order. With repeat-avoidance on, rounds after the first try up to 300 shuffles. Randomness comes from the browser’s Math.random(). It is suitable for entertainment and everyday, low-stakes decisions, but it is not cryptographically secure or independently audited.
- Time to use
- A round appears instantly; pasting the roster is the only slow part.
- Group size
- Two names are the minimum. There is no fixed maximum in the tool.
- Round memory
- Every round’s pairings are recorded until you press “Reset Rounds” or reload the page. The history is kept even while the avoid-repeats box is off, and it does not update when you edit the name list.
- Copy format
- “Pair 1: A & B” on one line per pair, plus a “Sits out this round” line when someone is left out.
- CSV export
- A “Pair, Name” header row, then one row per person. A trio shares one Pair label across three rows, and a sitter appears as “Sits out”. The file is named pairs-round-, the round number, and the date.
- Keyboard & screen reader
- The name box has a visible label, the checkboxes and odd-number radio buttons are native controls, and the Generate and export controls are reachable with Tab and triggered with Enter or Space. Generate stays disabled until two names are detected. Each new round is announced through a labeled
aria-live="polite"status region that spells out the round number, every pair, and who sits out, so the outcome doesn't depend on seeing the layout.
Four Moves From Roster to Rotation
Paste your names
Enter names one per line or separated by commas.Pick an odd-number rule
Choose whether an odd group makes one trio or one person sits out per round.Generate pairs
Click "Generate Pairs" — the list is shuffled and split into random pairs of two.Run more rounds
Click "New Round" for fresh pairs. With repeat-avoidance on, the tool steers away from partners used in earlier rounds.
What a Pair Generator Is, and Is Not, Compared With Teams
A random pair generator takes a list of people and matches everyone with exactly one partner, chosen at random. It answers a very specific and very common question: "who works with whom?" — for lab partners, peer review, pair programming, practice matches, dance partners, study buddies, or icebreaker conversations.
Pairing is a different job from team-making. A team generator divides a group into a few larger teams; a pair generator produces many small units of exactly two. The two tasks also fail differently: a slightly uneven team of 5 vs. 6 is fine, but a "pair" of one person is not a pair at all — which is why odd-number handling matters so much more here, and why this tool makes it an explicit choice instead of a silent decision.
Our Random Team Generator splits a group into a chosen number of larger teams — ask for 3 teams from 15 people and you get teams of five. The Pair Generator instead fixes the unit size at two and produces as many pairs as the group allows. Use teams when you need a few groups competing or collaborating; use pairs when the activity itself is built around partners. If you just need a single winner rather than groupings at all, a Random Name Picker is the right tool, and for reordering a whole list without grouping it, use the List Randomizer.
How One Round Is Shuffled and Paired
Under the hood, the tool runs a Fisher-Yates shuffle — the standard unbiased shuffling algorithm — on your list, then walks the shuffled order two names at a time. Because the shuffle gives every ordering of the list an equal chance, every possible set of pairs is equally likely too for a single shuffle. The randomness comes from the browser's built-in pseudorandom generator (Math.random()); it is fair for classroom and everyday use, but it is not cryptographically secure or audited, so don't use it for gambling or legally binding draws.
The number of possible pairings grows surprisingly fast: 15 ways for six people, 105 for eight, 945 for ten. Those counts also tell you what to expect from a first round: with 6 people, any specific pair meets with probability 1 in 5, because each person is equally likely to land with any of the other 5.
Habits That Keep Pairings Fair and Useful
- Check the detected count before generating. The tool shows how many names it parsed — if two names merged onto one line while pasting, you will catch it there.
- Decide the odd-number rule before you show the group. Announcing the rule first, then generating in front of everyone, removes any suspicion the process was steered.
- Reset rounds when the group changes. If people join or leave between activities, click "Reset Rounds" — the old pairing history no longer describes the group, and keeping it makes clean arrangements harder to find.
- Export each round if you need records. Round history lives only in the open page. For grading, tournaments, or attendance, download the CSV per round.
- Give paired strangers a starting point. For networking or classroom icebreakers, pair the tool with a few icebreaker questions or a round of Two Truths and a Lie so the first minute of every new pairing is easy.
Pairing, rounds and odd-number questions
It takes a list of names and splits everyone into random pairs of two. Unlike a team generator, which divides people into a few larger groups, a pair generator matches everyone with exactly one partner — the format used for lab partners, peer review, practice matches, and pair programming.
Last updated October 11, 2026. Something wrong with this tool, or a feature you need? Tell us — reported problems are looked into within 24 hours (corrections policy).