Last updated 3 September 2026
How teams stay even
The full list is shuffled once, then dealt out in rotation — first person to team one, second to team two, and so on, wrapping back around. That keeps every team's size within one person of every other team, even when the total number of names doesn't divide evenly by the number of teams.
How a fair shuffle works
Picking teams properly means every possible arrangement should be equally likely, and that is harder to achieve than it sounds. The naive approach — sorting a list by a random value — produces a biased result with many sorting implementations, subtly favouring certain orderings. This tool uses a Fisher-Yates shuffle instead, which walks the list once from the end, swapping each entry with a randomly chosen earlier one. Every permutation comes out equally likely, which is exactly the property you want when the outcome is a team assignment someone might dispute.
Uneven groups and repeated draws
Names rarely divide evenly. Eleven people across three teams gives 4, 4 and 3, and the sensible convention is to distribute the remainder one at a time from the top rather than leaving one group short by several. That way group sizes never differ by more than one, which is what most people mean by "split them evenly".
If you are picking teams for the same group repeatedly — a weekly game, a rotating pair programming schedule, a classroom — remember that independent random draws will happily put the same two people together three weeks running. That is not a fault in the shuffle. If you need genuine rotation rather than genuine randomness, keep a record of previous pairings and shuffle within the constraint, because randomness alone will not enforce variety.
Common questions
How are teams made even?
The full list is shuffled first, then dealt out one at a time in rotation — person 1 to team 1, person 2 to team 2, and so on, wrapping back to team 1. That keeps every team within one person of every other team, even when the total doesn't divide evenly.
Can the same name win twice when picking winners?
No — winners are drawn from the shuffled list without replacement, so each name can only be picked once no matter how many winners you ask for (up to the size of the list).
Is the shuffle actually fair?
It uses a Fisher-Yates shuffle, the standard unbiased method where every possible ordering of the list is equally likely — not a naive approach that can favor certain positions.
Why not just sort the list by random numbers?
Because that shortcut is not evenly random. Handing a sort function a random comparison means the same pair can be compared more than once and get a different answer each time, which pushes the result towards certain orderings. Dealing the list out after a Fisher-Yates pass avoids that entirely.
What happens when the names do not divide evenly?
The remainder is distributed one person at a time across the groups, so sizes never differ by more than one. Eleven names across three teams gives 4, 4 and 3 rather than leaving one group several people short.
Why did the same people end up together again?
Independent random draws have no memory of previous rounds, so repeat pairings are expected rather than a fault. If you want guaranteed rotation for a recurring group, you need to track past pairings and exclude them — randomness alone will not produce variety.