Splitting a Mixed-Skill Pickleball Group Across Courts

Twelve people turn up. Two of them are pushing 4.5, three started this spring, and the rest are somewhere in between. You have three courts and about ten seconds before someone asks who's playing who.

Most advice on this stops at "split them by level," which skips the part that actually matters. There are two ways to arrange a mixed group, they pull in opposite directions, and knowing which one you are doing is the whole job.

1. The two goals that fight each other

Flighting puts similar players on the same court. Court one takes the strongest four, court two the next four, and so on. Rallies are long because everyone can return the ball, the games are competitive, and nobody spends the evening being aimed at.

Balanced pairing does the opposite. It deliberately mixes levels on every court and evens things out by teaming the strongest player with the weakest. Scores stay close, everyone plays with everyone, and the person who came alone gets to know the group.

Notice that these are opposite instructions. Flighting keeps strong and weak apart; balanced pairing puts them together on purpose. Both produce close scores, and they produce them by completely different means — which is why "just balance the courts" is ambiguous advice that sends two organizers in opposite directions.

The trade is real and worth naming: flighting produces better games, balanced pairing produces a better group. A club night where people came to play should flight. A social where people came to meet each other should pair. Pick the one that matches why these particular people showed up.

2. Flighting: where to put the cut lines

The standard mistake is splitting by headcount — twelve players, three courts, so top four, middle four, bottom four. That is only right by accident. Cut at the gaps in the group, not at equal counts.

Write your twelve names in rough order of level and look at where the real jumps are. A group that runs 3.0, 3.0, 3.0, 3.0, 3.5, 3.5, 3.5, 3.5, 4.5, 4.5, 4.5, 4.5 has its gaps exactly where the headcount split falls, and you got lucky. A group that runs 2.5, 3.0, 3.0, 3.5, 3.5, 3.5, 3.5, 3.5, 3.5, 3.5, 4.0, 4.5 does not — eight of those twelve are the same level, and slicing them into arbitrary thirds creates three courts that all feel wrong.

You are reconciling two constraints that rarely agree: the gaps fall where they fall, and every court needs four players. When they conflict, the fix is not to force the gap or force the count — it is to decide where the boundary player has the better night.

One diagnostic to keep in your pocket: if games are routinely finishing 11-3 or worse, your bands are too wide. Not "someone had an off game" — routinely. Scores are the only honest read on whether your guesses were right, and they arrive within about twenty minutes.

3. Balanced pairing: the arithmetic actually works

If you are mixing levels rather than separating them, there is a formula, and it is exact for a foursome. Rank the four players 1 (strongest) to 4 (weakest) and there are only three possible pairings:

Strongest with weakest is not a rule of thumb, it is the arithmetically balanced answer, and it generalises: with any even group, pair the top-ranked player with the bottom-ranked, the second with the second-from-bottom, and keep working inward. The team totals come out level every time.

Now the part that gets left out. Balanced pairing equalises the scoreboard, not the experience. Put a 4.0 with a 2.5 and the opposition will hit every ball at the 2.5 — that is not unsportsmanlike, it is just how the game works. The score stays close and both players have a strange evening: one covering three quarters of the court, the other being served a diet of pace they cannot return.

So balanced pairing is the right call when mixing is the point and the spread is moderate. Across a genuinely wide spread it produces close scores and two unhappy people, and that is when you should be flighting instead.

4. Seed by guess, re-seed by evidence

However you split them, your first arrangement is a guess built on half-remembered games and whatever people say about their own level — and self-reported ratings run optimistic in one direction and modest in the other, unpredictably.

The fix is to stop treating the first split as final. Run three or four rounds, look at the scores, and re-flight at the break. After twenty minutes of actual play you know more about who belongs where than any pre-session ranking could tell you, and the correction is cheap because everyone expects a break anyway.

Announce it as the plan from the start — "we'll shuffle the courts at the halfway point once we've seen some scores" — and moving someone stops being a judgement and becomes the format working as intended.

5. Or let it sort itself out

There is a third option that does all of this without you ranking anybody: a ladder, where winners move up a court and losers move down after every timed round. Two or three rounds in, the group has sorted itself by level and nobody had to be told where they belonged.

It needs at least three courts and a group that all arrived together, and it is covered in full on the king of the court page. If your session fits it, it is the least awkward answer on this page.

6. Which one for your group

The first row is the one people skip. If your group is genuinely close in level, sorting them adds ceremony, creates a visible hierarchy that did not exist before, and improves nothing. Not sorting is a real answer and often the right one.

7. Awkward counts

Flighting has a hard requirement that balanced pairing does not: every court needs exactly four players. Twelve across three courts works, sixteen across four works, and ten across two does not — you get eight playing and two standing, every single round, which is the worst possible way to distribute a bye.

So the rule is simple. If your player count is not a clean multiple of four per court, do not flight. Run one rotation instead, where a properly built schedule spreads the byes across everybody rather than parking the same two people on the bench all night. The printable round robin charts handle counts from 4 to 16 with the byes already distributed, and the live tool does the same for any number.

If you are set on flighting with an awkward count, the workable version is uneven flights: six in the upper group on one court playing king of the court, and eight in the lower group on two courts running a rotation. Two formats at once is more to explain, but it beats two people watching.

8. How you say it matters as much as how you split it

Skill sorting is socially loaded in a way that court assignments are not, and the framing does most of the work. Lead with the outcome you are buying — closer games, more rallies — rather than the ranking that produces it, and never read out an ordered list of names. There is more on the wording, and on the rest of the run of show, in how to organize a pickleball mixer.

The one thing worth committing to: whatever you decide, decide it before people arrive. Sorting a group by level while twelve of them stand there watching you think is the version that goes badly, no matter which method you land on.

Frequently asked questions

The pairing table uses 4-3-2-1 as stand-in strength values to show the arithmetic; the result — top with bottom, working inward — holds for any evenly spaced ranking. Skill numbers are used loosely here as shorthand for level, not as formal ratings.