Nineteen Blank Columns on a Badminton Data Sheet: The Discipline of Writing in a Major Season

On the evening of March 12, 2026, I stayed late in my office in Da Nang with a spreadsheet forty-two columns wide. It tracked a women's singles quarter-final at a BWF World Tour Super 1000 event. Twenty-three columns had numbers. Nineteen were blank. The person who sent it added one line: "Where I could not record, I left it blank — please do not fill it in yourself."

I read that line three times. Then I did something I would not have done ten years ago: I folded the sheet away and wrote nothing about that match for two days.

In this trade, blank columns are the most concealed thing there is. People would rather enter an estimated figure and add a small footnote than leave a gap. A sheet with holes looks like a writer admitting he does not fully understand. At fifty-three, I know this: data is only a map, not the territory. And the badminton map in Vietnam is missing several large pieces.

Who measures, and with what

At Super 1000 and Super 750 events, the Hawk-Eye system is installed only on the television court, usually Court 1. Courts 2, 3 and 4 rely on human recorders. That means every published dataset for a tournament carries a skewed sample: big matches, with strong players and crowds, are measured by machine; early-round matches are measured by human eyes.

When I write for Vietnamese readers, most of what they see is Court 1. And Court 1 is where Nguyen Thuy Linh usually plays. That sounds like an advantage for analysis. It is also a trap: we have a great deal of data about the matches she went deep in, and almost none about the matches she exited early. A file containing only deep runs is not the file of a player. It is the file of scoreboards that happened to have machines attached. For players such as Nguyen Thuy Linh or Le Duc Phat, the number of machine-measured matches is always smaller than the number of matches they actually play.

Prize money at Super 1000 events in the 2026 season sits at around 1.3 million US dollars per tournament, Super 750 at around 850,000, Super 500 at around 420,000, Super 300 at around 240,000. Those figures matter to someone who works the market as I do, because they determine which player can afford another season on Court 2 and which player must rearrange a schedule to protect ranking points. A semi-final slot at a Super 500 can be worth more than a second-round slot at a Super 1000, depending on the points being defended. The ranking does not reward playing well. It rewards playing in the right place.

I have a habit of keeping raw notes by eye and by notebook, dating back to 2026 when I was still being stopped at the door of the press conference room. The official statistic sheet can always be wrong by one rally, one point, one beat. At the edge of the press conference room, I learned something data never records: a person speaking after a win and a person speaking after a loss use the same vocabulary, and differ only in their breathing.

Three groups of columns

Among those forty-two columns, I sort them into three groups by reliability.

Group one is machine-measured, with low error: smash speed in km/h, landing position, rally length counted in strokes. In women's singles, average rally length in Super 1000 quarter-finals and semi-finals in the 2026 season is around 9.4 strokes per rally, against 7.8 in the 2026 season. That figure says something about playing style: elite women's singles is grinding for longer, and the distance between an attacking player and a defensive player is paid for in time.

Group two is human-recorded: net-point win rate, unforced-error rate, quality of service return. This is the most contested group, because two recorders covering the same match can differ by up to 15% on unforced errors. In a match where the total points are only around 70 to 80, a 15% difference is a difference of roughly 6 to 8 points — enough to reverse the conclusion of an entire analysis.

What is notable is that badminton has no public standard for data recording, unlike football, where major data providers have standardised the definitions of a pass, a shot and a duel. In badminton, the definition of an "unforced error" depends on the recorder: a smash that lands out after a twenty-stroke rally can be counted as the attacker's mistake, or credited as the defender's achievement. One point, two records, two opposite conclusions.

Group three has never been measured in Vietnamese badminton, to the point that I left it blank in my own sheet: the technical cost of each decision. When a player chooses a flat smash instead of a placed drop, she changes the next rhythm as well, not just one shot. Football has PPDA to measure pressing intensity; badminton has no equivalent index for pressure at the net. I once spent twelve hours learning how to ask questions from a pressing system, not to find answers. Badminton needs exactly that kind of patience.

So when a Vietnamese player walks into the first round of a Super 500 and someone asks me for a prediction, I usually answer with a list of what I do not know: condition after the flight, the court surface, the scheduled time, and how many minutes the opponent has played in the past seven days. Those four variables explain more than any attacking index we habitually cite.

At the arena of a Challenge-level event in Da Nang, where I live, what I observe does not appear in any table. Young Vietnamese players often win the first rally after the serve and lose the third. Is that a fitness problem or a decision problem? No index answers it, because both leave the same trace on court.

The year 2026 taught me this in the harshest way. When tournaments returned inside empty arenas, the expected-goals index in football dropped by around 18% against the average with crowds, and pressure measurements became meaningless once opponents no longer felt the stands. Badminton is the same. A match without spectators in a provincial arena and a semi-final in a packed hall do not produce the same kind of data. Since then I have always included an "off-court pressure" variable in every analysis I write.

Correlation is not causation

Net-point win rate correlates very strongly with match win rate. But correlation is not causation, and the direction of causation here runs against most people's intuition. People assume that going to the net more means winning. In practice it is usually the player who is already winning who has the time and the position to come to the net. The net approach is a consequence of controlling the rally. An analysis that reads this index as a cause will advise the player to come to the net more, and that advice can lose the very next round.

By the same logic, a single-match sample is a noise sample. Seventy points are not enough to describe a player. I once stood on the wrong side of the data boundary: the Russia–Spain match of 2026, where I leaned on possession and expected goals, and ignored the fact that a deep-defending team can turn defending into a plan. Russia–Spain 2026: I was not wrong, I was just standing on the wrong side of the data boundary. Badminton has its own version: a player who defends stubbornly through a long rally is not a player suffering. She is betting that the opponent will run out of patience before she does.

I do not trust intuition, but I trust what intuition overlooks. And what intuition overlooks most in a major season is the condition of a person who has just flown four thousand kilometres, slept too little, and knows that losing this round means losing a place at the next tournament.

Signals for the next round

If I had to pick one signal to follow in the coming round, I would not pick smash speed. I would pick the distribution of rally length by game. A player who wins the first game with short average rallies, then gets dragged into long ones in the second game, is usually showing the mark of a plan that was not built for the final forty minutes.

At fifty-three, I no longer believe in beautiful predictions. I believe in a spreadsheet with nineteen blank columns. The blank column is the most honest place in a page full of numbers.