Team Splitter
Shuffle a list of people into balanced groups.
Shuffle, then deal
The reliable way to split a list into groups is to shuffle it properly and then deal cards round-robin: first person to team one, second to team two, and so on. This guarantees sizes differ by at most one and that every possible arrangement is equally likely. The naive alternative — repeatedly picking a random person and assigning them to a random team — produces uneven group sizes and needs awkward correction at the end.
The shuffle itself matters. Sorting by a random comparator, a widely copied one-liner, does not produce a uniform permutation and visibly favours certain arrangements. A Fisher–Yates shuffle, which this tool uses, does.
Random is not the same as balanced
A random split balances nothing except size. If your group contains four experienced players and eight beginners, a uniform shuffle will regularly put three of the four experienced players on one team. That is not a flaw in the randomness — it is randomness doing precisely what it does. If skill, seniority, or any other attribute needs to be spread evenly, split the list into tiers first and shuffle within each tier, then deal one person from each tier to each team.
The same applies to keeping people apart or together. Randomness has no knowledge of who worked together last week or who should not be paired; those constraints have to be applied by hand before or after the shuffle.
Uneven remainders and re-rolls
When the group does not divide evenly, some teams get one extra member. Deciding in advance which teams absorb the remainder — or that the extra person rotates as substitute — avoids an argument at exactly the moment people are keen to start. Re-rolling deserves the same treatment: agreeing beforehand whether a re-roll is allowed, and on what grounds, prevents the split from being quietly re-run until someone likes it, which destroys the fairness the shuffle provided.
How to use it
- Paste one name per line and choose the number of teams.
- If skill needs balancing, tier the list first and split each tier separately.
- Agree in advance who absorbs an uneven remainder.
- Accept the first result, or set re-roll rules before you draw rather than after.
Before you draw
Will the teams be balanced by skill?
No. A random split balances only size. To spread ability, sort people into tiers first and split each tier across the teams separately, then combine.
What happens when the group does not divide evenly?
The extra people are dealt out one per team, so sizes differ by at most one. Deciding beforehand which teams take the extra member avoids a dispute after the split.
Why not just sort by a random value?
Sorting with a random comparator does not produce a uniform shuffle — some arrangements come up far more often than others. This tool uses a Fisher–Yates shuffle, which gives every arrangement equal probability.
Can I keep certain people apart?
Not automatically — randomness has no knowledge of your constraints. Split the group manually into sub-pools first and shuffle within each, or adjust the result by hand afterwards.
Disclaimer: These tools are for games, classrooms, and informal group decisions. They are not audited for regulated lotteries, gambling, or any draw with legal requirements, and should not be used to make financial, medical, legal, or safety decisions.