Badminton in a Data Drought: Why the Prettiest Number Is Usually the Most Suspicious One

The broadcast graphic flickered midway through the third game. The player on the left side of the court had a slightly higher rally-win rate, nearly double the smashes, and by the numbers alone, he was controlling the match. The man sitting next to me in the analysis room — a former national-team coach — just shook his head. He said something I have carried for years: "He is losing from the very rallies that are not being counted."

That sentence captures the entire problem facing badminton right now. The sport produces beautiful numbers, flashed across big screens, shared on social media, and used to pass judgment on a player after a single evening. But most of those numbers omit precisely the thing that decides the match.

Badminton is not short on technology. Hawk-Eye has been installed at events on the BWF World Tour to assist line calls, allowing players to request instant review. Every major court now has high-speed cameras, shuttle-speed sensors, and a substantial volume of tracking data. The problem lies in the gap between what is measured and what is published.

Put it next to football. There, the concept of xG (expected goals) has become common language — publicly defined, debatable as a model, and supported by open data anyone can rerun. In badminton, no equivalent exists. No one publishes "expected points" for each rally; no one releases the full tracking dataset of an entire tournament for outside analysts to verify. Fans receive scattered statistics selected by broadcast producers and quietly treat them as truth.

I began following badminton matches in the 1990s, when Lin Dan was not yet famous and Lee Chong Wei was just a young, fast player. Back then, analysis relied on the eye. We sat and counted by hand how often a player lost the net exchange, how often he was forced deep into the backcourt, how often his attacking clear landed exactly where the opponent was waiting. It was manual labour, and that is precisely why it forced the analyst to understand the mechanism before reaching a conclusion.

When automated data arrived, part of the analytical world changed its method. Instead of asking "what did this smash mean", they asked "how many smashes were there". Quantity replaced quality. The countable number of smashes does not tell you whether that smash opened a winning rally or merely ended one the player had already lost.

This is where we must speak plainly about the nature of combat sports. In badminton, a point is scored when the shuttle touches the opponent's floor. But the match is built on hundreds of rallies, most of which end in errors. Those errors fall into two kinds: forced and unforced. Broadcast stat sheets merge both into a single number called "unforced errors", and that merge creates an illusion.

I once spent an evening breaking down a rally sequence from a top attacking player. The official figure credited him with twenty error situations across a three-game match. Clean. But peeling through each rally, more than half were the consequence of an opponent repeatedly driving the shuttle to the high left corner, forcing him to choose between a risky line and losing the rally. Those were not unforced errors. They were the cost of a defensive system doing its job correctly.

That is why I keep repeating one line to my readers: the prettiest number is the most suspicious one. Not because the number is technically wrong, but because the process that produced it is hidden.

Take shuttle speed. The recorded world record once belonged to a smash exceeding 490 km/h under special measurement conditions. That figure appeared across the press, accompanied by the feeling that badminton is the fastest sport on the planet. It is half true: the shuttle flies fast, but that speed collapses within a few metres, and laboratory measurement conditions differ vastly from match conditions. A pretty number, stripped of context, becomes a false story.

Then there is "point-win rate on serve". This indicator depends on whether the opponent is a strong defensive player, on indoor drift, on whether the server is in the second or third game when stamina has dropped. The same percentage, two different contexts, two opposite meanings. A high serve point-win rate may reflect a weak returner, not a strong server.

Here a cross-border problem appears, one I noticed after years working across markets. Each badminton ecosystem defines and publishes data differently. The Asian confederation, the European confederation, domestic circuits in Indonesia, Malaysia, Denmark, and China — each has its own standard for what is counted and what is ignored. When a player moves from one system to another, his numbers migrate too and change meaning. A careless analyst will directly compare two datasets built on different standards and draw conclusions about progress or decline.

I fell into that trap once. Years ago, I made a judgment about a player based on his point-win rate in one tournament, compared against data from another. Readers spotted the inconsistency in counting methods. I lost a night rechecking, and from then on set a rule for myself: never conclude from a sample smaller than ten matches, and always check the standard under which a number was produced.

But more worrying than the data itself is the public reaction. When a number is published, it carries the credibility of the publisher. Fans have no tool to trace the collection process. They accept the number as confirmed fact, when in reality it is merely an interpretive choice. And those choices, repeated often enough, shape how we judge an entire generation of players.

I once watched a debate stretch over months about whether a former world number one was truly declining. The numbers camp pointed to a falling win rate. The eye camp pointed out that he still controlled the tempo, only that younger opponents had learned to exploit a specific weakness in the mid-court zone. Both sides were right within their own data framework. The problem was that neither would admit the limits of that framework.

That is the biggest blind spot of modern badminton. The sport has enough technology to generate data, but lacks a culture of publishing and verifying it. It lacks an independent community able to rerun the calculations. It lacks globally unified standard definitions. The result is a race to produce glossy stat sheets for television, rather than to understand the match.

The paradox is that the hardest things to measure are the ones that decide. The ability to read an opponent's intent before returning a shot. The ability to control breathing through a long rally. The ability to pick the right moment to accelerate and the right moment to yield tempo. No sensor measures these, and no stat sheet turns them into a percentage.

If current data were accurate in the way it is published, we would live in a world where the player who smashes most is the best player. Anyone who has watched badminton long enough knows that is not true. Viktor Axelsen won Olympic gold not because he smashed more than everyone in every match, but because he chose the right moment, the right position, and held the structure of the match against opponents faster in any single shot. Kento Momota once dominated by slowing his opponents' game, a skill that barely appears on a broadcast stat sheet.

I am not calling for the abandonment of data. The opposite: I want more data, deeper and more open. But I want it accompanied by honesty about its own limits. A number is only valuable when the reader knows where it came from, what it measured, and what it ignored.

In a period when the badminton market is moving sharply — players changing teams, federations restructuring, new tournaments emerging — the need for a trustworthy data language becomes more urgent than ever. Those of us in this profession have a responsibility not to turn a pretty stat sheet into a verdict on a player's career.

When a number is published, the first question should be: who made it, by which standard, and what was left out. Not out of baseless suspicion, but to keep the debate about badminton meaningful. This sport deserves a data system as honest as the long rallies it produces.

A player on court has no stat sheet to lean on. He has only a feel for tempo, for the opponent's weakness, for the moment to accelerate. Perhaps the analyst should start there too — understand the match first, then count.