The Badminton Data Gap: Anatomy of a Sport That Refuses to Count

During a World Badminton Championship semifinal, a statistic on the television broadcast made me stop in the middle of my workspace in Chengdu. The winning player was credited with twelve direct points. I rewound the recording, counted every rally by hand, and got nineteen.

A seven-point discrepancy, on a figure the broadcaster used to explain why the match unfolded the way it did. Viewers at home trusted that statistic, because they had no other way to verify it. That night I understood something about the sport I have followed for years: badminton does not lack data. Badminton lacks a system for verifying data. A sport is running on faith in figures that have never been cross-checked.

I came to badminton by a winding path. At seventeen, I sat in front of a screen watching France play Argentina in the 2026 World Cup knockout round. Instead of shouting when Mbappé scored, I opened a spreadsheet. I recorded touches, top speed, distance covered. That night I learned that emotion needs to be verified, that a match can be read again through data without losing its beauty.

But football gave me something badminton does not: an open data ecosystem. In football I have expected goals, I have passes per defensive action, I have progressive passes, I have per-minute distance. I can buy data from multiple independent providers, cross-check them, find the errors. In badminton I have essentially one source, and that source lets no one check it.

Why I record every rally by hand

At twenty, I defended my master's thesis on a result-prediction model based on expected goals, and I thought I would stay in football forever. But an opportunity in 2026 put me in the broadcast booth of the Sudirman Cup, the largest mixed-team badminton event on the planet. Behind the scenes of a world-class badminton event, I noticed a paradox: this is a sport whose rallies move faster than any other racket sport, with rallies lasting over two minutes and smashes exceeding four hundred kilometres per hour. Yet its public data trove is thin as paper.

I started keeping my own notebook. Every match, I logged rally duration, the number of times a player was pulled to the two flanks, direct smashes, unforced errors before and after the thirtieth point. I logged the moments a player lost balance, the rhythm of breathing during interval breaks. When the pandemic suspended global tournaments in 2026, I stayed home and built a health ranking for every national badminton system, based on domestic tournament structure, the number of players in the world top fifty, and federation payrolls. My way of seeing a sport changed from then on. When the court went silent, I built a ranking of resources to understand why a system collapsed.

The three layers of one problem

Badminton's data problem splits into three overlapping layers, and each is difficult.

The first layer is collection. At major events, the tracking system for shuttle flight is installed to serve the umpires, not to publish data for the public. It knows whether the shuttle lands in or out, but it is rarely configured to export touch counts, distance covered, or change-of-direction counts. At smaller events, there is no system at all. Journalists sit courtside and count by eye.

The second layer is standardisation. There is no unified definition of a "direct point." Depending on who is counting, a winning smash may count as a direct point, or it may count as an opponent's error if the umpire deems the return failed. The same rally, two counters, two results. And when two broadcasters count in two different ways, viewers believe both.

The third layer is access. Data that is collected and standardised then sits in the hands of a small group. Independent researchers have no access to challenge it. Journalists have no raw dataset to cross-check. Fans only see the final figure on screen, filtered through three layers, without knowing which layer distorted it.

Together these three layers produce a dangerous consequence. When data has only one source and no one can check that source, every analysis becomes a story. I no longer shout at the screen; I record every rally, and I have discovered that most of what I once believed about the top players was built on figures no one ever recounted.

The illusion of smash speed

No metric is abused more than smash speed. A smash reaching four hundred and twenty kilometres per hour becomes a headline across the press. Fans assume the hardest smasher is the most dangerous player.

When I cross-checked smash-speed data against match results across more than two hundred matches I had recorded, the correlation was almost zero. The fastest smash usually falls in a rally where the player is already comfortable, the opponent is deep, and the defence is set. It makes a beautiful image on television, but its scoring probability is no higher than a cross-court smash at only three hundred and thirty kilometres per hour placed precisely into the opponent's blind spot.

Smash speed is chosen for broadcast because it is easy to measure. Measuring it needs only one sensor. Measuring placement, movement rhythm, how long an opponent takes to change direction, the space the winner creates, all of that is far harder. And the sports media industry, after all, prefers cheap metrics. We are measuring what is easy to measure, then telling the story as if what is easy to measure is what matters.

What the notebook shows me

I will take one concrete example. Viktor Axelsen, the Danish player with towering height, is known worldwide as one of the hardest smashers in history. But what I recorded across many of his matches was not smash speed. It was the number of times he forced his opponent to run. On average, in every rally he won, he made his opponent change direction at least three times before finishing. The player across the net did not lose to a smash; he lost to attrition.

Then Kunlavut Vitidsarn, the far smaller Thai player. In the matches I tracked, he won not through power but through the ability to sustain long rallies and place the shuttle into the four corners in an almost programmed sequence. My notebook gave him a ratio I called the long-rally rate. In his wins, that ratio usually exceeded sixty percent. In his losses, it fell below forty. No official statistic carries such a metric. No one broadcasts it.

An Se-young, the Korean women's player, taught me something else again. What made her dominant was not the winning smash but the ability to turn defence into counter-attack. I once recorded a match in which she trailed after the first game, yet her direct counter-attacking points from defensive positions rose steadily with each game. Reading her match is a chain of controlled adjustments, not a lucky shot. The key metric here is adjustment speed, something no tournament stat sheet measures.

Three players, three different ways of winning, and the official metric system does not accurately describe the way any of them wins.

My system and its limits

I built my own badminton metric, which I called the Attrition and Control Coefficient. It is not complicated. It sums three components: the number of times an opponent is forced to change direction per rally, the controlled long-rally rate, and the rate of successful counter-attacks after being pushed into a defensive position. I adjust the weights of the three components each season, based on which component predicts results better in my database.

This metric is not perfect. It rests on one person's eyes, not on sensors. It cannot measure psychological pressure, nor the fatigue hidden behind a return that is half a beat slow. But it is transparent. Anyone can open my notebook, cross-check every rally, and find where I counted wrong. That is what the broadcast statistic does not permit.

Data is like scripture: reading much is not for believing, but for questioning. A published metric without its raw data is a statement of faith, not a piece of evidence. And I refuse to accept faith in place of evidence.

Why football manages it and badminton does not

The question many colleagues ask me is: if football built an open data ecosystem, why does badminton not follow?

The first answer is market scale. A top football match draws an audience dozens of times larger than a top badminton match. Football data revenue is large enough that many independent companies compete to buy sensors and hire courtside recorders. In badminton, data margins are too thin to create competition. Without competition, there is no cross-check, no verification.

The second answer is more structural. Football has hundreds of independent national leagues, each selling its own data. Badminton has one global federation managing almost the entire professional competition system. When a single organisation runs the events, holds the data, and issues the official statistics, no pressure forces it to publish raw data. A single source always has an incentive to control how the story is told.

The third answer lies in the sport itself. Badminton moves too fast for manual recording. An average rally lasts seconds, yet in those seconds there can be five changes of direction, three smashes, and two net approaches. Capturing all of that by eye is nearly impossible, even for the experienced. That makes badminton a sport needing more technology than football, yet possessing less data technology than football.

The injury and workload void

There is one area where the data shortage has the gravest consequences: injury.

Football publishes relatively broad information on injuries, recovery timelines, and match load. Badminton does not. A player withdraws from a tournament with a vague statement, and no one learns whether it was a knee, ankle, or shoulder injury. No one learns how many matches that player played in how many days before collapsing.

I once built a tracker for the schedules of leading players across a season, based only on public calendars. Some players competed continuously on three continents within five weeks, with recovery windows between events of under seven days. When a body is pushed into that cycle, injury is not a risk; it is an inevitable result. But because injury data is not published, the cycle repeats, and every time a player collapses, the public calls it misfortune.

The way we treat players returning from injury reflects this same ignorance. We demand they prove themselves in their very first match back, as if their bodies could forget months away from competition through willpower alone. That standard does not come from sports science. It comes from the habit of telling stories about miraculous comebacks, built on data that does not exist.

A contrarian truth

I have spent most of this piece criticising badminton's data shortage. But going against myself, I must admit one thing: more data does not automatically make this sport better.

Look at what happened to football. When expected goals became widespread, the way people talked about football changed. Players began to be measured by abstract numbers rather than by moments. The debate over who is better drifted off the pitch and into the spreadsheet. A sport can lose its intuition when everything is reduced to metrics.

If badminton follows that path, it must pay a similar price. The beauty of a rally is not in the number of direction changes. It is in the moment a player knows the opponent's return in advance, reading the intention before the shuttle leaves the racket. No sensor captures that, and if we try to capture it, we may destroy the very thing we mean to preserve.

But I still want data, for another reason. I want data not to replace intuition, but to expose lies. When a federation says the schedule is reasonable, I want a number to check it. When a stat sheet says a player scored twelve direct points, I want the raw data to recount. When a player says his body is fine, I want an injury record to cross-check. Data does not make a sport more beautiful. It makes a sport more honest. And honesty, in the end, is the minimum a professional sport owes its audience.

Signals to watch

In the coming major season, I will watch three specific signals.

First, whether the organisers of major events begin publishing raw data alongside summary statistics, or keep the old habit of only releasing the final figure. If raw data appears, that is a sign the system is starting to open.

Second, whether any independent provider is bold enough to collect badminton data without the federation's permission, accepting legal risk to create a second reference source. The emergence of a second source would change the entire game.

Third, whether injury information is published more transparently, so that the debate over match load shifts from sentiment to evidence. If that happens, players pushed into brutal competitive cycles will have a chance to be protected by data, not only by goodwill.

I will still sit courtside with my notebook, counting every rally. Not because I distrust technology. But because until there is a trustworthy second data source, the notebook of one careful counter remains the most honest mirror for every broadcast statistic. One match does not make a truth. A thousand matches, recounted, might. And if this sport agrees to count, fans will no longer have to believe. They will get to know.