The Cambrian Pickleball Rankings

What started as a group of friends getting together to play pickleball has grown into something genuinely special. Over the last two years, around 40 players have come together across seven tournaments, organized by a small group of volunteers who keep showing up — booking courts, building brackets, recording scores, settling disputes, and somehow still finding time to play.

If you've ever organized one of these, you know how much invisible work it takes. Cambrian's organizers have done it again and again, without compensation, without sponsorship, purely because they love the game and the community around it.

That's the foundation everything else here is built on.

Why we built a ranking system

For most of these two years, we've been doing what every casual league does: showing up, playing, having fun, going home. Brackets were seeded by intuition. Teams were balanced by feel. Everyone has an opinion about who the strongest player is, and everyone is at least partly right.

But as the league grew, a question kept coming up: could we seed tournaments more objectively?

That's where this project began.

The tournaments

The ranking is built from match data across the following Cambrian tournaments, in chronological order:

A small note on completeness:some playoff matches may be missing from a few tournaments — for example, one playoff match in the 8th tournament wasn't recorded. The organizers were busy trying to wrap up the remaining matches at AVAC before closing time 😄. If you remember the scores for any missing playoff matches, share them with the organizers — the rankings will update once that data is added.

How the data was collected

We won't pretend this was elegant.

Two of the tournaments were recorded in Swish (a pickleball scoring app), and the rest were tracked in Google Sheets — sometimes in different formats, sometimes mid-tournament on a phone. Player names were inconsistent. Match orders were ambiguous. Tournament numbering didn't always match chronological order.

We cleaned all of it up. Every match was reviewed, players were reconciled to canonical names, dates were verified, and the entire dataset was normalized into a single consistent format. The cleaned data is being made available to the organizers for cross-validation. If you played in any of these tournaments and want to verify your matches, the organizers have access to the source scoresheets — original and cleaned.

This is important: every score in this ranking can be traced back to an actual match that was actually played. There are no guesses, no estimates, no assumed values.

How the rankings are calculated

The algorithm is Glicko-2 with a hybrid scoring approach.

Glicko-2 is a well-established rating algorithm developed by Mark Glickman, a Harvard statistics professor. It's used widely in competitive games and sports — including World Pickleball Rankings, which uses the same Glicko-2 foundation for their global pickleball ratings. The fact that the world's leading pickleball ranking system relies on Glicko-2 was a strong signal that we were on the right track.

Let me explain how it works without the math.

The basic idea

Every player starts at the same rating with the same level of uncertainty. As they play matches, their rating moves based on three things:

• Whether they won or lost.
• The score margin — a 21-9 win counts more than a 21-19 win.
• The strength of their opponents.

This third one matters more than people realize. Beating a top player gains you significantly more than beating a weaker player — the algorithm expects you to beat weaker opponents, so when you do, it doesn't tell us much. But when you beat someone strong, the algorithm takes notice.

The hybrid scoring formula

We use a 65/35 split: 65% of the match outcome is determined by who won, and 35% by the score margin. This means winning still matters most (a win is a win), but blowouts and close games are weighted differently. A close 21-19 win counts as a real win without being treated as a near-tie, and a 21-5 demolition reflects the dominance it represents.

How recency factors in

Tournaments are processed in chronological order. Recent matches naturally carry more weight in the math — your last tournament impacts your rating more than your tournament from two years ago. Skip tournaments and the system becomes less certain about you. Show up consistently, and your rating sharpens.

How rankings are sorted

This is the part that occasionally surprises people: we don't sort by raw DUPR rating. We sort by Confidence Rating.

The Confidence Rating is your DUPR adjusted for how much data we have about you. A player who has played one tournament and won 8 of 9 matches might have a sky-high DUPR — but we don't actually know if that's their real level. Their Confidence Rating gets a larger penalty until they prove it across more tournaments.

This prevents a one-tournament wonder from topping the rankings, while still giving them credit for what they've done. As they keep playing, their Confidence Rating catches up to their DUPR.

Reliability tiers

Each player has a Reliability score reflecting how confident the system is in their rating:

The first time a new player appears, they're Provisional. Three or four tournaments in, they're Established. The system is honest about what it knows and what it doesn't.

How this compares to DUPR and UTR

A few people have asked: is this DUPR?

No.DUPR is a global rating system calibrated on millions of matches across regions, with their own proprietary formula and adjustments. Players in our league might have a “real” DUPR rating that's calibrated against players worldwide.

What we have is a DUPR-style ranking system, similar in spirit to DUPR (and to UTR in tennis), but calibrated only on Cambrian matches. We use the same kind of mathematical framework that World Pickleball Rankings and many other serious rating systems use, but we anchored our display ratings to feel right for our league — the top established player sits around 4.10 DUPR, with the rest of the pack flowing down from there.

Think of it as a Cambrian-internal ranking. It accurately reflects how you stack up against the players you actually play with. It doesn't claim to predict how you'd do against players outside this group.

If you want a globally-comparable rating, real DUPR is the gold standard. If you want to know who's been the most dominant player in Cambrian over the last two years, this is your answer.

An important disclaimer

This ranking is based purely on tournament performance.

If your ranking feels lower than where you think you should be, please read this part carefully: your true skill level may absolutely be higher than what the data shows. You might be:

• A player who plays great in casual sessions but hasn't had a breakthrough tournament yet.
• Someone who's been paired with weaker partners and has been carrying them.
• A player going through a slump after years of dominance.
• Someone who plays a strategic, close-game style that doesn't show up as blowout wins.
• A player who just had bad luck with matchups.

The algorithm doesn't see any of that. It only sees scores.

Take this positively. The rankings are not a verdict on your ability — they're a snapshot of your tournament results to date. Show up, bring your best, and the system will reward you tournament after tournament. Several players in the current rankings climbed significantly between their first and most recent tournament. The rankings are designed to recognize that improvement.

A bad tournament won't define you. A good tournament will move you up. Sustained good play will move you up a lot.

How we kept it fair

A few principles we held to:

• No bias. Every player goes through the exact same math. There are no manual adjustments, no overrides, no exceptions. The algorithm doesn’t know who anyone is.
• No baseline ratings. Nobody starts higher or lower than anyone else. Everyone begins from the same starting point and earns their position from match results.
• All data, all the time. Every match from every tournament counts. We didn’t drop “bad data” or “outlier matches.” Real tournaments include weird results — and that’s part of the signal.
• Open and verifiable. The cleaned match data and source code will be shared with the organizers. If anyone wants to validate the calculations, they can. We’re not hiding anything — and there are absolutely other reasonable ways to run a ranking algorithm. We picked this one because Glicko-2 has strong academic backing, is used by World Pickleball Rankings, and handles small-league situations gracefully. But this isn’t the only valid approach, and we’re open to feedback as the system evolves.
• Chronological processing. Tournaments are processed in the order they happened, not in tournament-number order.

FAQ

A final thank-you

To the organizers who've kept this going for two years — thank you. To Abhinay for pushing the ranking system into existence. To every player who's shown up, played their heart out, and shared scores with us. To the volunteers who track every match.

This ranking belongs to all of you. It's a small piece of recognition for everything that's been built.

Now let's play some more pickleball.