Random team generator

The product here is teams, not a winner

Every other wheel on this site answers "who?". This one answers "how do we split?", and that is a different output: a set of teams, all at once, rather than one name.

The wheel is still here for drawing an individual — who goes first, who picks — but the splitter is the part that does this page's job. Give it a number of teams, or a number of people per team, and it deals everyone out.

Deal, do not slice

Nine people into four teams is the case that separates a good split from a lazy one.

The lazy way is to shuffle and then cut the list into four pieces, which gives you 3, 2, 2, 2 — the remainder bunched at one end. The way this does it is to shuffle and then deal round-robin, like cards, so the difference between the biggest and smallest team is never more than one person. That is what anyone splitting a class for a game means by "even", and it is worth doing properly because the alternative is visibly unfair in a way people notice immediately.

When it does not divide evenly, the page says so and says by how much, rather than letting you count.

By count or by size

Which one you want depends on what is fixed.

If you have four pitches, you need four teams and the size follows. If the game needs pairs, you need teams of two and the count follows. Both are here because picking the wrong one means doing arithmetic in your head in front of a group that is waiting.

The shuffle is a real shuffle

Teams are dealt from a shuffled list, and the shuffle uses the same cryptographic source as every draw on this site — not the ordinary random function most code reaches for.

That matters less for teams than for a raffle, but it costs nothing and it keeps one thing true across the whole product: there is exactly one source of randomness here, it is documented, and nothing quietly uses a weaker one.

Doing it again

Splitting again reshuffles from scratch. There is no memory of who was with whom last time, which is the honest behaviour but not always the wanted one — if you are avoiding repeat pairings across a session, that is still something to track yourself.