Rocket League is the cleanest competitive game there is to rate. No roles, no characters, no abilities, no economy, no map pool worth arguing about, no draft. Two teams, one ball, identical starting conditions. Every other post on this blog spends half its length on confounders — autofill, map veto, character counterpicks, comp diversity — and in Rocket League none of them exist. The only variable is the players.
Then the playlists ruin it.
A community that plays 1v1, 2v2 and 3v3 is not playing one game at three difficulties. It's playing three games that happen to share a control scheme, and that has a specific and awkward consequence for rating — one that most leagues resolve by splitting into three ladders, which is correct in principle and usually a mistake in practice.
Playlists Are Not Tiers
The obvious model is that 1v1 is the pure test and the team playlists add coordination on top: same skill, more noise. If that were true, a 1v1 rating would be your best estimate of everything, and 3v3 would just be 1v1 plus teamwork.
Ask any Rocket League player and they'll tell you it doesn't work like that, and the reason isn't teamwork — it's that the two playlists reward partly opposed habits.
1v1 is a possession game. There's no one to rotate with, so the winning pattern is conservative: shadow defence, don't over-commit, take the backboard, wait for the mistake. Challenging aggressively when you're the last player back is simply losing.
3v3 punishes exactly that. If you hold back with two teammates behind you, you clog the rotation, starve your team of boost and turn a three-man press into a traffic jam. The winning pattern is to commit, get out, and trust that someone is behind you.
So a player can grind 1v1 into a genuinely strong possession game and carry a habit into 3v3 that actively costs their team. The skills overlap in mechanics — aerials, car control, and touch transfer completely — and diverge in decision-making. That's why 1v1 specialists are a recognised type in the community rather than just "good players who haven't played 3s yet".
For a rating system this means something precise: a single cross-playlist rating doesn't just average noisily, it averages two things that trade off against each other. A player who's 1400 at 1v1 and 1100 at 3v3 isn't a 1250 at anything. There's no game where that number is their strength.
So Split The Ratings — Which Is Where The Trouble Starts
The fix is obvious and correct: rate each playlist separately. Because Rocket League's playlists are exactly team sizes — 1, 2 and 3 — this maps cleanly onto a by-team-size rating rather than needing anything custom. A 3v3 result updates the 3v3 rating and nothing else.
Then you find out what splitting costs.
A rating needs somewhere around 25 matches per player before it means much. Here's a 30-player league playing 20 matches a week across the whole community — a healthy, active Discord league.
One ladder, everything is 3v3:
| Playlist | Matches/week | Per player/week | Weeks to 25 |
|---|---|---|---|
| 3v3 | 20.0 | 4.00 | 6.2 |
Six weeks and the board means something. Now split it across playlists in roughly the proportions Rocket League communities actually play them — 3v3 heaviest, 1v1 a niche:
| Playlist | Matches/week | Per player/week | Weeks to 25 |
|---|---|---|---|
| 3v3 (60%) | 12.0 | 2.40 | 10.4 |
| 2v2 (30%) | 6.0 | 0.80 | 31.2 |
| 1v1 (10%) | 2.0 | 0.13 | 187.5 — 3.6 years |
The 3v3 ladder survives the split. The 2v2 ladder takes most of a year. The 1v1 ladder never converges at all — the season will end, twice, before those numbers describe anybody.
And the reason is arithmetic rather than popularity: a 3v3 match generates six player-results, a 1v1 generates two. So the least-played playlist is also the least efficient at producing data, and the two effects multiply. A 1v1 match is worth one third of a 3v3 match in ratings evidence before you account for it being played a tenth as often.
Work out your own number
Don't take my example league — the arithmetic is simple enough to run on yours. For any playlist:
weeks to converge = 25 × (active players)
─────────────────────────────────────────
(matches per week in that playlist)
× (players per match)
players per match is 2 for 1v1, 4 for 2v2, 6 for 3v3. Use your real per-playlist match counts rather than a total, because the split is exactly what you're testing.
If a playlist comes out longer than your season, it does not have a ladder — it has a number that will still be mostly its starting value when the season ends. That's the whole test, and it's worth running before you create the board rather than after.
The Real Question
So the question was never "should I split by playlist?" — the answer to that is yes, they're different games. The question is how many ladders your league can afford to feed, and for most Rocket League Discords the honest answer is one, occasionally two.
What that looks like in practice:
Pick a flagship playlist and rate it properly. For most communities that's 3v3, because it's what the game's competitive scene is built around and it's what people show up for. That ladder gets the season, the standings, the tier roles, the tournament seeding — everything.
Run the others as unranked or casual. They still happen, they're still fun, results still get recorded for stats and history, but they don't feed a rating that will never converge. This is not a downgrade; a ladder with 0.13 matches per player per week is worse than no ladder, because it displays a number people will believe.
Add a second ladder when the volume justifies it, not before. The test is simple and worth writing down: can that playlist alone give each active player ~25 matches within one season? If your 2v2 traffic doubles, it's earned a board.
If you genuinely want all three, run them as separate seasons — a 1v1 month, a 2v2 month — so the volume concentrates instead of spreading. Three thin ladders in parallel is the worst configuration available; three sequential ones is a real format.
The general shape here is the reverse of the usual advice. Narrower ratings are more meaningful per result and slower to converge, which the attribution post covers as a general trade. Rocket League is where that trade is at its most brutal, because the playlist split is both mandatory (opposed skills) and expensive (thin volume), and those pull in opposite directions.
Three Things That Are Smaller Than They Look
Worth covering, but each one is a paragraph rather than a section — they're either well-trodden or less RL-specific than they seem.
Solo versus party. Two friends on comms beat two strangers of equal rating, and in a 2v2 that gap is amplified simply because coordination is half the team — there's no third player and no role structure to absorb it. But the fixes are the same as anywhere else and the LoL inhouse post works through them properly: split parties across teams where the lobby allows, or track it as a match stat and check whether it actually correlates in your group before legislating.
The stat line. Rocket League's box score — goals, assists, saves, shots — is famously bad at describing contribution. The player who spends a match pressuring, taking boost and forcing bad clears often finishes with a blank line, and demo/possession/pressure aren't in the scoreboard at all. So Rocket League is the game where rating on performance stats goes wrong fastest. Track them for interest, rate on results, per the attribution argument.
Seasonal resets. Rocket League's own soft reset each season means your community already expects ratings to compress toward the middle and be re-earned, which is a genuine cultural advantage — you don't have to sell the idea. Import the expectation, not the mechanic; how to run a competitive season covers the carryover modes and what each does to a community.
Setting It Up
- Rate by team size. 1v1, 2v2 and 3v3 get separate rating tracks; a 3v3 result never touches the 1v1 number.
- Choose one flagship playlist and give it the season, the roles and the tournament seeding.
- Run the other playlists unranked until their own volume clears roughly 25 matches per active player per season.
- Don't publish a cross-playlist combined rating. It averages skills that trade off, so it describes nobody.
- Consider sequential playlist seasons if all three matter to your community — concentrated volume beats parallel thin ladders.
- Rate on results, track the box score separately.
Frequently Asked Questions
Why not just use one rating for everything?
Because 1v1 and 3v3 reward partly opposed decision-making, so a combined number sits between two values that don't average meaningfully. A player who is strong at one and weak at the other gets a middling rating that predicts their performance in neither playlist — and it'll be wrong in a different direction depending on which queue they're in.
Can I convert someone's 3v3 rating into a 1v1 starting rating?
As a rough seed, yes — mechanics transfer, so a strong 3v3 player is rarely a weak 1v1 player from a standing start. Just don't treat it as a measurement. Seed the new ladder from the old one to skip the worst of the cold start, mark those ratings provisional, and let real 1v1 results take over.
What about 4v4 or Rumble and the extra modes?
Same logic, weaker case. Casual chaos modes select for different things again, so they'd technically deserve their own ratings — but they're almost never played with enough volume to support one. Record them for stats and history and leave them unranked. The volume test decides this, not the principle.
Is 2v2 or 3v3 the better flagship?
3v3 for most communities, because it's what the competitive scene is organised around and it produces six player-results per match instead of four, so it converges fastest for the same number of games. Pick 2v2 only if it's genuinely what your community plays more of — the flagship should be the playlist with the volume, not the prestige.
How do I stop the unranked playlists feeling like an afterthought?
Give them things a rating isn't needed for: stat leaderboards, tournaments with their own seeding, monthly events. A playlist without a ladder can still have standings, records and a champion — the ladder is one feature, not the whole competitive structure.
Running a Rocket League community on Discord? Team Up keeps separate rating tracks per team size automatically, so 1v1, 2v2 and 3v3 results never contaminate each other, and lets you run any playlist unranked while still recording its stats and history. Start with the leaderboard setup guide, or the matchmaking queue guide to get the games flowing first.