The Empty Badminton Data Table: A Lesson in Verification for Sports Reporting

On my desk in Shenzhen sits a nine-part analysis. Each part has a table. Each table has dozens of cells. And almost every cell repeats the same sentence: insufficient information, cannot assess.

I read it three times. The first time I assumed a system error. The second time I assumed the author had been careless. By the third reading I understood: that analysis was more honest than most of what I have read about badminton over the past two weeks.

Sports media does not lack data. It lacks the admission that some regions of data are simply empty. A tournament that has not started. A player who has not published a schedule. A contract with no signing date. Those gaps are always filled with statements delivered in a confident voice, and readers have no way to separate information from inference dressed up as information.

I worked as a sports scientist before I picked up a pen. That job taught me a habit the writing trade dislikes: when there are no numbers, the only honest move is to say there are no numbers.

That nine-part analysis says nothing about badminton. It says one thing about its author: that person refused to fill the blank with a guess. In a week when I received at least four messages a day asking about deals nobody has confirmed, that kind of honesty has value.

Real hierarchies, and a news stream without one

Professional badminton runs on a hierarchy so dry it is almost reassuring. The Badminton World Federation (BWF) tiers the World Tour into Super 1000, Super 750, Super 500, Super 300 and Super 100. Ranking points are allocated by tier, by the round a player reaches, and by how many events count toward the total. Olympic qualification stretches across a four-year cycle with a published points-collection window. The Thomas Cup and Uber Cup run on a two-year rhythm for men's and women's teams. The Sudirman Cup is the mixed-team stage, also on a two-year cycle. The World Championships remain the highest individual stage in the system.

Everything inside that system is measurable. A player knows exactly which round they must reach to defend points. A federation knows exactly which entries will lift its team into a seeding group. A coach knows exactly which week of the year the squad will break under a scheduling clash.

Yet the news stream flowing outside that system has almost no hierarchy at all. On the same day, fans read three categories of information with entirely different reliability, presented in the same tone. An official federation notice. A sourced quote. A writer's hunch. All three sit side by side as if they were in the same weight class.

That is why I keep the habit of naming the source for every number, even when the source is my own notebook from a three-hour video review. Readers deserve to know where a figure came from, under what conditions it was measured, and how many samples stand behind it. Hide that, and you are selling a conclusion with no load-bearing wall.

A player-movement period makes everything systematically worse. When the market opens, the volume of information rises faster than the speed of verification. Rumour holds a structural advantage over fact: it is cheap, it is fast, and it is never penalised when it is wrong. A report confirming a completed deal needs two independent sources and often takes days. A speculative line takes thirty seconds.

Three filters before believing a badminton story

My process for handling an unverified story has three steps, and I run them strictly in this order, even under deadline.

The first filter is the source. Who said it, to whom, and what does that person gain when the sentence spreads? An agent has an incentive to move a price. A coaching staff has an incentive to create internal pressure. A tournament organiser has an incentive to keep tickets selling. None of those incentives makes the information false. They determine which direction the information was filtered. Numbers do not lie. But they are extremely good at selecting which truths to tell. A badminton statistics table can perfectly well tell the story that one player serves far better than an opponent, if the person who built the table selected only the matches that player won in the first game.

The second filter is the sample. Badminton has enormous variance at the level of a single match. A player who wins 21-19 in a deciding game may be only two rallies better than the opponent. If I use the last three matches to conclude something about form, I am describing a feeling, not a trend. The minimum threshold I set myself for a form judgement is ten matches within the same competitive block, or six if the schedule is dense and the opponents are of comparable quality. Below that threshold, I am allowed to describe, not to conclude.

The third filter is structure. A story about personnel movement in team events only means something when placed beside squad structure. How many domestic slots remain, how many foreign slots remain, how much salary budget is left, and at which position in the lineup that player would actually play. Without that structural layer, every figure about money and contracts is noise in careful packaging.

These three filters do not make me write faster. They make me write less, and the smaller output stands up longer.

What badminton data measures, and what it misses

Compared with football, badminton's data ecosystem is narrower on the surface but richer in one key respect: every point has a clear winner, with no drawn-out line disputes.

The indicators that can be measured with reasonable stability include the share of points won on serve, the share won on receive, average rally length, the number of net crossings per rally, unforced error rate, and the distribution of landing points by court zone. In doubles, analysts can also measure the average distance between the two players on a pair and how long they sustain a continuous attacking phase.

One family of metrics is misused more than any other: the ones describing the outcome of a rally while ignoring the entire process that produced it. A fast smash winner is the product of the three shots before it: a deep push into the left corner, a cut that forced the opponent to lift, and an early step to reach the right position. If the table records only the final smash, it credits the harvester rather than the person who created the situation.

In football this is known as the problem with expected goals: it describes the quality of a chance rather than the quality of a decision. Badminton has the same problem on a more granular scale. A player with a high share of points won at the net does not necessarily have better net technique. That player may simply be placed in easier net situations because a partner is generating pressure from the rear court. Setting two players side by side with a raw number is a meaningless comparison when the situational context differs.

This is why I review video at least twice for every situation before it enters an article. The first pass records what happened. The second checks whether the player's position before the decisive shot matches what I remembered. During three weeks of video analysis for a 47-page internal report in Shenzhen, I found my memory was wrong on roughly one in five situations. Wrong in both directions: some moments I assumed were luck turned out to be intent; some I assumed were intent turned out to be reflex.

There is a further source of error few people discuss: measurement conditions. Whether the arena is air-conditioned, which direction the draught runs, and at which speed number the shuttlecocks were set. The same player, the same opponent, the same tactics can produce noticeably different results if shuttle flight speed changes. I once compared two competitive blocks and nearly published a wrong conclusion, purely because one was played indoors and the other in a draughty hall. Since then, whenever I place two data tables side by side, I state the collection context and the error margin directly beneath the table.

In Shenzhen, I have seen data replace intuition. The results are not always prettier. Some coaching decisions cannot be explained by data, and the match later proves the coach right. When that happens, I record it and treat it as an unmeasured variable rather than an error to be stripped out of the model.

The 2026 lesson and the limits of any measurement system

I once believed a good model could predict how a team would play. The 2026 World Cup taught me that every system can be taken apart. A side forecast to press high can walk into a quarter-final with a deep defensive block, concede possession, and win by attacking the space behind the opposing back line. The pre-tournament data was not wrong. It merely described one version of that team, and that version was not the one released in the most important match.

Badminton has an equivalent. A player can attack all season, then arrive at a quarter-final and choose to play long, slow shuttles that drag an opponent into exhausting rallies. A full-season statistics table will label that an anomaly. In reality it is an adjustment. And adjustments always happen before the data catches up.

Since then I have dropped the habit of praising only attacking play. In every article I try to devote at least one passage to dissecting the tactical sacrifice of the defensive side. My two most-used tools are formation length and the gap between lines. In doubles, the distance between two players on a pair, measured over duration, tells you whether they are playing as one block or as two individuals stitched together.

Attacking players such as Viktor Axelsen and attrition-based defenders such as Kunlavut Vitidsarn cannot be judged on the same scale. Setting their attacking-winner rates side by side compares two different competitive philosophies using a ruler suited to only one of them. That is the most common analytical error I see on badminton forums today.

The blind spot is in execution, not in the data

There is a trap I have fallen into before and still have to remind myself about weekly. When the data table is empty, a writer's reflex is to fill it with reasoning. When the data table is full, a writer's reflex is to treat it as truth. Both reflexes fail in the same way: they turn a measuring tool into a conclusion.

The real blind spot in sports analysis is not a lack of metrics. It is that metrics cannot price a decision in the instant of execution. A player knows exactly where the shuttle must go. That player can still miss, because the wrist is tired after the twelfth rally, because of noise behind the back line, because they have lost to this opponent three times in six months. No statistics table holds all three variables at once.

I once sat in a technical room and heard someone say a player's numbers were falling. Nobody in the room asked how many hours that player was training a day, how many hours they slept, or what injury they were managing. Falling numbers are a result, not a cause. In a rushed article, the result is always easier to write than the cause.

Viewers see magic. I see three pressing layers drilled since Tuesday. But I also have to admit those three layers can collapse in the third game for a reason outside the diagram: a lapse in concentration, a disputed call, a slipped foot. The intuition of coaches and players is a measurable variable, it is just that nobody has bothered to measure it.

If I track one coach across three seasons and log every mid-match tactical change, I will have a trend. That trend may forecast better than any statistics table. That is the direction I am testing in recent months.

What is worth tracking during a player-movement period

The player-movement period is when the news stream is loudest, and when filters matter most. Fans do not need more rumours. They need a way to rank a rumour's reliability.

Three things I will track in the coming weeks, in order of priority. Contract structure: length, extension clauses, and how the money is paid year by year. The total figure of a contract usually matters less than how it is broken down. Remaining salary budget at each club: that is the indicator determining whether a third deal is possible. And the injury status of key players, confirmed by medical staff or by whether the player appears on court in training.

All three are verifiable. Everything else I will read and note, but will not put into print until at least two independent sources confirm it.

Process wins a match. Discipline wins a season. That holds for players, and it holds for the people who write about them.

What I want to carry out of that empty nine-part analysis is simpler than any model. When there is no data yet, the most useful thing a writer can do is state clearly that there is no data yet, and then go find some. Badminton fans deserve a news stream they can trace back to a source. Every time a newsroom refuses to fill a blank with a guess, the common standard moves up one notch. Every time a newsroom fills it in to make the deadline, that standard slips, and the cost is not paid by the newsroom. It is paid by the reader.