Forty-Seven Pages of N/A: The Humility Line of Badminton Data Analysis

The A4 sheet lay on the meeting table of a badminton training center on the central coast one June morning. Forty-seven pages. Every page had a table. Every table had a column of metrics: average strokes per game, rally length, peak smash speed, unforced-error rate, points won after losing serve, movement tempo on court. And in the answer cell of almost every row, the same symbol repeated again and again: N/A.

I read all forty-seven pages in twelve minutes. Not because I read fast. But because there was nothing to read.

A twenty-six-year-old analyst, fresh out of university, had been tasked with compiling match data for a young player about to enter the qualifying rounds of an international tournament on the World Badminton Federation circuit. He followed the process I taught him. He built the frame, set up the tables, named the columns, assigned units. Then he discovered what no one had told him: the data source was empty. No footage sharp enough to break down each rally. No point-by-point score sheet. No physical profile. No systematically recorded head-to-head history.

He had two choices. One was to invent plausible-sounding numbers so the report looked full. The other was to leave N/A.

He chose the second. And so those forty-seven pages became the most honest confession I have ever read about my own profession.

The craft of badminton data analysis in Vietnam has a much shorter history than football. Football has decades of statistics, international data companies recording every pass, a culture of expected-value metrics that has sunk deep into how people argue. Badminton is different. The sport is still often seen as a clash of instinct, of reflexes, of two hands. People remember a smash the way they remember a thunderclap; few remember how many mispositioned movements preceded that thunderclap.

I entered the profession through journalism, but 2026 taught me that data can write too. That year I sat in a meeting room in Nha Trang presenting a report on a football club, and I learned that a number placed in the right spot carries more weight than an editorial. When I moved into badminton, I brought that principle with me. But I also brought a doubt: whether badminton even has enough data to speak through numbers.

Nguyen Tien Minh posed that question to an entire generation. The player born in 2026 once climbed to fifth in the world and won a bronze medal at the 2026 world championships. Those are facts that can be looked up and verified. But behind that medal, how many metrics were fully recorded? How many of his matches were broken down into individual rallies, individual movement rhythms, individual decisions about where to place the shuttle? The honest answer is: not many.

The next generation — players such as Nguyen Thuy Linh in women's singles, or Vu Thi Trang, or Le Duc Phat — grew up in an era where fans can watch every tournament live online. But watching and analyzing are two different things. A match streamed to hundreds of thousands of viewers does not automatically generate a dataset usable for coaching. Between the image and the knowledge there is still a gap. And that gap is exactly where the forty-seven-page report was born.

The first thing to state clearly: N/A is not zero. In the language of data, zero is an observation — it says we measured, and the measured value was empty. N/A is a confession — it says we did not measure, or could not. These two are different in nature, and confusing them is a fatal error in the craft.

When a coach reads a table and sees an unforced-error rate of 0, he might conclude his player played cleanly. But when he sees the unforced-error rate marked N/A, he should understand he is standing before a blind spot. Both cells print the same on paper if one does not read carefully. And in that darkness, human judgment will fill the gap by itself — with feeling, with memory, with bias.

In badminton, the greatest danger of data is not wrong data, but missing data presented as complete data.

There is a psychological mechanism that makes this happen constantly. When we must decide — who makes the team, who rests, who trains harder — we cannot tolerate emptiness. The human brain hates blank space. It will automatically fill it with a story. This player has good spirit. That player is not mature enough. This opponent is awkward. Those labels sound like judgments, but they are really N/A cells wearing the costume of words.

I once watched a national team select players based on three matches seen live, while ignoring twenty other matches only because no footage existed. Those three matches became the entire dataset. And three matches are far too few to say anything about a player. But because that was all there was, people treated it as everything. This is the classic error: using the availability of data as the measure of its value.

So if we want to build a decent badminton data foundation, what should we measure?

The core metrics of a singles player fall into four groups. The first is scoring structure: what percentage of points a player wins when holding serve, and what percentage when the opponent holds serve. The gap between these two numbers says a great deal about style. A player with a high win rate on serve but a low win rate on receive depends on being in control. A player balanced at both ends has a solid foundation.

The second is tempo. Average rally length is one of the most underrated metrics. But it speaks to style and also to stamina. A player averaging rallies under eight seconds usually relies on early attack. A player averaging rallies over fifteen seconds usually relies on endurance and patience. When these two styles meet, rally length becomes a tactical weapon: the short player wants to end early, the long player wants to drag it out and wear the opponent down.

The third is attacking efficiency. Peak smash speed is only the visible part. The submerged part, which matters more, is the ratio of smashes that win points to total smashes. A player with a 400 km/h smash who wins only 40% of his smashes is wasting energy. A player with 350 km/h who wins 60% of his smashes knows when to strike. Speed is the flashy number; efficiency is the real one.

The fourth is errors. Unforced errors split into two kinds: errors caused by an opponent's pressure, and errors caused by lapses in focus. Distinguishing them requires footage and a person to break it down. But once distinguished, it transforms how a player is judged. A player with many errors because he is being pressed is normal. A player with many errors at unimportant scorelines is a concern.

These four groups do not demand expensive technology. They demand one thing: someone to sit down after every match and take notes. That is the dullest job in the profession. And because it is dull, it is often skipped. Every match is a tea session for the data monk — silent, but steeped through. No tea session, no data.

Here I must speak of another trap, one I once fell into myself. In 2026, when tournaments stopped because of the pandemic, I had time to review data from many teams. I found a pattern: teams playing high pressing with very low pressure metrics tended to collapse late, conceding most of their goals in the final ten minutes. I wrote a piece about it and said modern football had bet wrong on running intensity.

But when I applied that thinking to badminton, I realized I needed to be more careful. In badminton, a continuously pressing style does not exist in the football sense. The closest equivalent is a relentless attacking style that pushes the opponent to the back court and seizes control of the net. That style burns terrifying amounts of energy. And the question becomes: is there a stamina threshold beyond which the attacking player starts making more errors?

I believe there is. But I do not yet have enough data to name a specific figure. And this is the point I want to stress: when data is insufficient, the most honest act is to say it is insufficient, not to offer a number that sounds certain. My profession is not the profession of certainty. My profession is the profession of conditional certainty.

There is another aspect badminton data often overlooks: managing competitive load. For years people have said players compete too much, travel too much, and need rest. It sounds very reasonable. But when I look at actual schedules, I see a paradox. Players said to be resting often appear in friendlies, promotional events, exhibition matches. What is called rest is sometimes just a switch from ranking competition to commercial competition.

Load management, if not measured by real data, very easily becomes an excuse to arrange schedules for the organizer's benefit rather than the player's. I am not saying every case is like this. I am saying that without data on playing hours, kilometers traveled, and maximum-effort rallies, every claim about load management is merely belief.

Here I must raise a hot topic in sport: the value of young players. In football, I have argued that the bubble in young-player prices is bursting — paying a hundred million euros for a player who has not played fifty top-flight matches is a naked gamble. Badminton has a milder version of this phenomenon. Academies, sponsors, and training centers sometimes bet very heavily on a seventeen- or eighteen-year-old who has just won a few good results at junior level. But the gap between junior and professional is an abyss. And that gap is barely measured by data.

A young player wins a junior title because opponents of the same age lack stamina. Stepping up to the professional level, stamina is no longer an advantage. Basic technique is no longer enough. What remains is the ability to adapt tactically — and that is extremely hard to measure. But because we cannot measure it, we use results to infer potential. This is a logical error: taking past achievement as evidence of the future while ignoring that the environment has changed.

Good data is not data that makes us more confident. Good data is data that tells us exactly what we should doubt.

The World Badminton Federation ranking system operates on the principle of defending points. Each player must defend points won in the same period the previous year. If they do not compete, the old points expire. This creates a pressure that data often fails to capture: the pressure to take the court even when the body is not ready.

When I analyze a player's schedule, I always draw a two-axis chart: the vertical axis is points to defend by week, the horizontal axis is travel distance between tournaments. These axes often intersect at very bad points — meaning the player must fly half the world to defend a small number of points while stamina is running dry. But to draw that chart, I need data on points, expiry dates, and flight schedules. And that data is usually scattered, with no one compiling it.

In doubles, the data problem is even more serious. In singles at least points can be counted. Doubles is far more complex: there is chemistry, role division, a net specialist, a rear-court specialist. A good pair is not two good individuals added together. It is a system. And a system needs data on interactions, on the gaps the two leave behind, on synchronized movement rhythm.

I once watched a young pair play and saw them lose because shuttles fell exactly between the two players. On the surface, that is a coordination error. But with data on each player's position per rally, one could see a pattern: that gap appeared when one player was dragged to the sideline. That is a system error, not an individual one. But without positional data, people blame chemistry.

The badminton industry runs along a chain. Upstream is youth development and talent supply. Midstream is players and tournaments. Downstream is equipment, broadcasting, and derivative markets. Data flows unevenly through this chain.

Downstream, equipment brands collect data very well — they know exactly which racket models sell, which strings are favored. But that data serves sales, not coaching. Midstream, tournaments have scoring data but lack technical data. Upstream, where data is needed most to nurture talent, data is thinnest.

This is a structural paradox: where data is needed most, there is least of it. Youth training centers lack note-takers, footage, storage systems. Meanwhile a major tournament may have dozens of cameras but uses them only for broadcasting, not analysis.

I remember an evening in Nha Trang, reviewing footage of a match by a young player. I rewound a single rally exactly twenty-three times. It was a rally lasting nineteen seconds, ending in a failed drop shot. If you look only at the result, you would say the player chose the wrong placement. But rewinding slowly, I counted eleven changes of direction in nineteen seconds, and on the eleventh, his foot was about forty centimeters out of position. The failed drop was not a technical error. It was the consequence of a positional error that happened three seconds earlier.

That kind of information only exists when someone is willing to rewind footage twenty-three times. No software does that for free. No algorithm replaces patience. And that is why badminton data in Vietnam remains thin: it demands human time, the most expensive of all resources.

There is a temptation anyone in the data profession must face, and I want to speak plainly about it. When we hold a beautiful dataset, we want it to tell a story. And when we want it to tell a story, we begin to select. We keep the numbers that support our argument, we ignore the numbers against it. This is the moment data turns from a tool for finding truth into a weapon for defending the ego.

Correlation is not causation. A player with a high win rate in long rallies does not mean long rallies make him win. Perhaps he was already good, so opponents were forced to extend rallies, and he still won. Confounding variables are everywhere: court conditions, humidity, the shuttle, the schedule, psychology. An analysis based on only ten matches is an analysis fooling itself.

I have been criticized for using data to defend a losing athlete. People said I was trying to justify failure. But I think differently. Numbers are never in a hurry. We are the ones in a hurry. When a player loses, the crowd rushes to conclude he is weak. Data slowly points out that his reaction speed off the start is among the leaders, that his smash still hits peak speeds, and that the defeat lay in another small detail. That slowness is not excuse-making. It is respect for the truth.

I must also speak of another trap I have seen in many people in this profession. It is using data to create a false superiority. When we have metrics and others do not, we easily fall into contempt. We treat fans as emotional, journalists as ignorant, coaches as backward. But that attitude does not make data stronger. It only makes us lonelier. Data is strongest when placed on the table with people who have no data, to find the truth together, not to prove who is right.

That forty-seven-page report full of N/A, I kept it. I did not tear it up. I left it on the shelf, next to other reports stuffed with numbers. Because it reminds me that my profession has a line: the line between what we know and what we think we know.

When the court is empty and data is abundant, I understood that I follow badminton for people, not only for numbers. Numbers only mean something when they serve a specific person — a player trying, a coach struggling, a twenty-six-year-old analyst willing to leave a blank instead of fabricating.

The next cycle of Vietnamese badminton will not be decided by who has the most data, but by who is honest about their own blanks. That is the signal I am waiting for. And I believe it will come from someone who dares to write N/A.