Tool
Balanced Team Splitter
Rate each player, then snake-draft into teams with matched total strength.
Balanced Team Splitter
Rate each player, and the tool builds teams whose total strength is as close as it can make them. Ties within a rating band are broken at random, so the same ratings do not produce the same teams every time.
This is deliberately not a random split, and that is the whole point. A random split gives every player an equal chance of any team and produces a lopsided game surprisingly often; a balanced split produces a good game and gives the strongest player no chance at all of being on the same team as the second strongest.
You cannot have both. Choosing between them is a decision about whether you are optimising for the players or for the match.
How the Balanced Team Splitter works — and why it's fair
Players are grouped into bands by their rating, each band is shuffled, and the bands are then concatenated strongest first. That gives a strength-ordered list where equal ratings appear in a random order — which is the only place randomness enters this tool at all.
That ordered list is dealt into teams in a snake pattern: the first team takes the strongest, the last team takes the next, and the direction reverses on each pass. So with three teams the order runs 1-2-3-3-2-1-1-2-3, and the team that got the weakest pick in one round gets the strongest in the next. It is the same method sports drafts use, for the same reason.
Snake drafting is not optimal partitioning and does not claim to be. A solver could occasionally find a split with a smaller gap between team totals, and for pathological ratings it would find a substantially better one. Snake ordering gets close, runs instantly, and produces a result anyone can verify by reading down the list — which for a group standing on a pitch is worth more than optimality.
The gap between the strongest and weakest team is computed and shown, because that number is the actual output. A gap of one or two on totals in the thirties is a genuinely even game; a gap of eight is not, and no amount of describing the teams as balanced will change how the match plays.
The randomness is confined to ties and cannot move a player between rating bands. A player rated 9 will always be picked before a player rated 8, on every run, which means the same ratings produce the same team strengths every time even though the specific line-ups vary. If that feels wrong, the ratings are the thing to argue about — which is the argument worth having.
When the Balanced Team Splitter is fair — and when it is not
What it does guarantee
- Team totals come out as close together as a snake draft can make them, and the gap is computed and shown.
- Players with identical ratings are ordered randomly, so equal players are not permanently attached to one team.
- The method is fully inspectable — anyone can read down the ordered list and check the deal.
What it does not
- It is not a random split. A strong player has effectively no chance of being teamed with another strong player.
- The output is only as good as the ratings, and ratings are opinions. A wrong rating produces a confidently unbalanced game.
- Snake drafting is a heuristic, not an optimum. A solver could sometimes find a closer split.
Two worked examples
8 players rated 9,8,7,6,5,4,3,2 into 2 teams
- Snake order gives team 1 the ratings 9, 6, 5, 2 for a total of 22.
- Team 2 gets 8, 7, 4, 3 for a total of 22 — a gap of zero.
- A purely random split of the same eight players averages a gap of about 5.
The same 8 players into 3 teams
- Teams of 3, 3 and 2 with totals of roughly 16, 15 and 13.
- The gap of 3 is mostly a consequence of the uneven team sizes, not of the draft.
- Ties within a rating band shuffle each run, so the exact line-ups vary while the totals do not.
When a good game beats a fair draw
Pickup football, basketball and five-a-side are the direct case, and the reason is that a lopsided game is bad for everyone including the winners. A random split of ten players produces a genuinely one-sided match a substantial fraction of the time, and the group notices within ten minutes.
Quiz and board-game leagues use it where the skill spread is wide and known, because a group that contains one specialist and seven casual players will otherwise be decided entirely by which team drew the specialist.
Classroom group work uses it in a slightly different form, distributing the students who need most support rather than concentrating them. The mechanic is identical; only the interpretation of the rating changes.
For a genuinely random split where every player has an equal chance of every team, the team generator is the correct tool and says so. For pairing rather than teams, the pair generator handles the odd-number case explicitly.
Frequently asked questions
Is this random?
Only within rating bands. Players with the same rating are ordered randomly; players with different ratings are always picked in strength order.
Why not just split randomly?
Because a random split of ten players produces a lopsided game a large fraction of the time. If you want the random version, the team generator does exactly that.
What ratings should I use?
Any consistent scale — 1 to 10 is convenient. Only the relative ordering and the totals matter, so the absolute numbers are arbitrary.
What if I leave a rating off?
That player is treated as a 5. That keeps them in the middle of the order rather than dropping them to the bottom.
Is snake drafting the best possible split?
No, it is a good heuristic. A solver could occasionally find a closer split, but snake ordering runs instantly and can be verified by reading the list.
Why do the teams look the same every time?
Because the ratings drive the order. Only ties shuffle, so identical ratings give identical team strengths — the line-ups vary, the balance does not.
What does the gap number mean?
The difference between the highest and lowest team totals. It is the single number that tells you whether the game will be close.
Can it handle uneven team sizes?
Yes. With 8 players in 3 teams you get 3, 3 and 2, and the gap will partly reflect that size difference rather than a draft failure.
Does it consider position or role?
No. It balances a single rating number, so a team can end up with every good defender and no attacker.
Related tools
Team & Group Generator
Split any list of names into fair, balanced teams or groups.
Tournament Bracket Generator
Draw a single-elimination bracket with byes — random draw or classic seeding.
Random Team Name Generator
Combinatorial team names, with an alliteration filter that narrows the space on purpose.