Rally Chains and Real Value: Vietnam's Badminton Through Point-by-Point Data

Game Three, 18-18

In the deciding game, at 18-18, I started a stopwatch on a single rally: twenty-seven racket touches, fourteen changes of direction, ending with Nguyen Thuy Linh's cross-court smash landing exactly on the sideline. The arena stood up. Thirty seconds later, on the very next point, the rally lasted five racket touches and ended with a short serve error by the same player.

Two points. One player. One score. Two different sports.

I stayed behind after the match with my notebook and my rally-coding sheet. Across 31 matches featuring Vietnamese players that I watched live or reviewed in full over the past fourteen months, one pattern repeats with almost uncomfortable regularity: the performance of Vietnamese players changes not with the name of the opponent, but with the length of the rally. The same player, the same tournament, the same round - yet their point-win rate falls steadily once a rally passes fifteen racket touches.

People call that character. I call it a data curve nobody has drawn yet.

How I Built This Dataset

I work as a transfer market administrator in Jakarta, but my real job is coding. Every match I watch, I code each rally into a row: rally length in racket touches, server, receiver, point winner, point-ending type (smash winner, net error, out-of-bounds error, service error, dead shuttle after the opponent lost balance, and those points I file under 'not technically explicable').

The dataset covers 2,846 rallies over fourteen months, 47 players from nine countries and territories, including seven Vietnamese players. I split rally length into five bands: one to five touches, six to ten, eleven to fifteen, sixteen to twenty, and twenty-one plus. I split matches into three time frames: from the first point to eleven, from eleven to fifteen, and from fifteen onward.

Every number I publish has a footprint. And I can point you to that footprint.

Three rules have governed my work since 2026, after a dossier on a winger whose agent published a dribbling index double the real figure. First, every metric must carry its sample size. Second, every metric must carry its unit and time frame. Third, if the data cannot support a conclusion, I write that no conclusion is possible - and I accept losing readers for it.

I do not average whole-match rally totals, because that throws context in the bin. A rally at 3-2 and a rally at 19-19 can have identical length and entirely different weight. I count them separately.

Three Time Frames and the Death of Game Three

The first thing the data showed me: across 2,846 rallies, 38 percent ended within five racket touches. The six-to-ten band accounted for 29 percent. Eleven to fifteen took 18 percent. Rallies of sixteen touches or more accounted for only 15 percent. Nearly two thirds of modern badminton points are settled before a rally becomes a long tactical contest.

That is the general picture. The Vietnamese-specific picture is what matters.

In the one-to-five band, Vietnamese players in my dataset won 61 percent of points, above the 56 percent average of the Southeast Asian comparison group. In the six-to-ten band, their win rate fell to 54 percent. In the eleven-to-fifteen band, it was 47 percent. Beyond sixteen touches, it collapsed to 39 percent.

That curve slopes down almost linearly. It says something the naked eye cannot easily see: Vietnamese players do not lose because they lack basic technique. They lose because their point system is designed to end rallies early, and when an opponent is good enough to extend rallies, that system loses power.

By match time frame: from the first point to eleven, Vietnamese players won 55 percent of points. From eleven to fifteen, 51 percent. From fifteen onward, 44 percent.

Game three tells the story most clearly. Across nineteen matches that reached a deciding game, their point-win rate from fifteen onward in game three was 41 percent, against 49 percent in the same time frame in game one. Eight percentage points sounds small. In a badminton endgame, eight points of win rate is a weakness written into the opponent's notebook.

Some things look like luck. They are actually an equation.

After 15-15, the Match Changes Hands

I isolated rallies played at 15-15 or higher. In 412 such rallies, Vietnamese players won 44 percent of points. Of the 56 percent lost, 52 percent came from unforced errors rather than opponents' winners. I define an unforced error as losing a point without direct pressure from a high-quality shot: a shuttle out of the sideline while not being pushed, a shuttle into the net from a balanced position, a service error, or a smash into the net from a controlled situation.

Below 15-15, unforced errors made up only 38 percent of lost points. Under pressure, that share rises fourteen percentage points. Their opponents do not get stronger. They get weaker.

Against a control group of eleven top-20 players coded to the same standard, the unforced-error share of lost points at 15-15 and above was 41 percent - eleven points lower. That gap does not live in the wrist. It lives in shot selection under a fraction of a second.

This is where I regularly disagree with television commentators. They talk about character and big-match experience. I do not deny either exists. But if character cannot be expressed as a measurable number, it is just a pleasant way of speaking. And a pleasant way of speaking does not help anyone train better the next morning.

The Smash Is Not the Weapon

One of the most common beliefs in Vietnamese badminton is that players need harder smashes. I understand why: a fast cross-court smash makes the arena stand up, and the arena records emotion while ignoring data.

My data says otherwise. Across 2,846 rallies I counted 6,109 smashes. Vietnamese players averaged 7.2 smash winners per match but also 9.8 smash errors. The winner-to-error ratio was 0.73. For the top-20 group, the figures were 8.1 and 7.4, a ratio of 1.09.

The difference lies in the denominator, not the numerator. Vietnamese players do not smash less. They generate more power per smash but choose worse moments, and the price is 1.36 smash errors for every smash winner.

Split by court position, the picture sharpens. Smashes from deep positions, when a player has been pushed back and is on the defensive, win only 28 percent of points but account for 61 percent of all Vietnamese smashes in long rallies. They smash most at the moment smashing helps least. The top-20 group does the opposite: in long rallies they use drop shots, clears, and drives to restore balance, smashing only from the front court or when the opponent has lost the centre. Deep-court smashes make up just 34 percent of their total.

Data does not carry the roar of the crowd. It carries the truth.

The Net, Where Matches Are Actually Decided

If I had to pick one figure to summarise this entire dataset, it would be the share of points decided at the net.

Of 2,846 rallies, 1,146 points were settled in the front half of the court, roughly within two metres of the net. That is 40 percent of all points in a fraction of the court's area. Elite badminton is a game of the half-metre around the net.

For Vietnamese players I recorded 5.4 net winners per match and 7.1 net errors, a ratio of 0.76. For the top-20 group: 6.8 and 5.2, a ratio of 1.31.

One nuance matters. Net errors are not only shuttles pushed into the net. They include safe net choices - lifting the shuttle high to avoid risk and handing the opponent time to smash. Those rallies are not logged as errors in the official record. They are logged as opponents' winners, and the Vietnamese player walks back to the service line not knowing they just lost a point to a decision made two seconds earlier.

I counted this group separately. At 15-15 and above, across 412 rallies, there were 38 rallies where a Vietnamese player chose a high lift at the net from a balanced position. They lost 29 of those 38. A 76 percent loss rate from a decision considered safe.

Safety at the elite level is a conditional concept. It is safe if the opponent cannot punish it. It is self-destruction if the opponent is world class.

Schedule Density and the Injury Invoice

Based on my years of match observation, I have reached a conclusion I have defended repeatedly in front of team managers: schedule density is the single biggest cause of injury. No medical team rescues a player who plays two matches a week for nine months.

In my dataset I flagged 14 players with dense schedules - three or more tournaments in eight consecutive weeks at World Tour level. Nine of them withdrew with an injury within the following twelve months. That is 64 percent. In a comparison group of 19 players with sparser schedules - at most two tournaments in eight weeks - only five withdrew in the same window. That is 26 percent.

The 38-point gap is the largest single figure in the dataset. It exceeds every technical gap I measured between player groups. If you want to improve a Vietnamese player's results, the first move is not hiring another fitness coach. It is cutting their tournament count.

Vietnamese players in the dataset show a worrying pattern: they enter many consecutive events to accumulate ranking points, because the World Federation ranking counts a player's best results over a 52-week cycle. It is a structural trap. Players need points to enter big events, but chasing points exhausts them before they get there.

In four cases I tracked in enough detail, Vietnamese players arrived at their target tournament with measurable declines in lateral movement speed, with average reaction time in direction-change rallies roughly 0.08 seconds slower than three months earlier. Eight hundredths of a second sounds minor. At elite level, that gap is the difference between a winning push and a push the opponent reads early.

Valuing a Player with Data

This is the real work I do, and the part I am least often asked about.

Football has transfer fees. Badminton has a market too, but a dimmer one: personal sponsorship deals, national team budgets, overseas training slots, and arrangements both sides prefer to keep quiet.

When a federation or club asks me to value a player, I do not look at trophy counts. I look at four data groups: performance at 15-15 and above, unforced-error rate by time frame, age and remaining prime years, and schedule density over the last thirty-six months.

I put endgame performance first for a simple reason. That is where the money is paid. A player winning 55 percent of points in that zone goes deep at major events, earns entry to high-prize tournaments, and carries higher commercial value. A player winning 60 percent in the first twenty minutes but only 42 percent in the deciding zone will never reach a semifinal, however elegant the fundamentals.

I once applied this to a specific regional case. A young player was offered support worth four years of training costs. I asked for her 15-15 data over twelve months. It came back at 39 percent, with an unforced-error share of 49 percent of lost points. The headline numbers used to sell her were movement speed and smash-point rate.

I advised against signing. The other side signed. Three years later she had not once passed the quarterfinals at continental level.

People call that a market shock. I call it a re-examination of real value.

The point is not that I read people well. The point is that the badminton market, especially in Southeast Asia, prices players on the most visible metrics and ignores the most decisive ones. Speed, jumping power, and smash force show up in thirty seconds of highlights. The choice of shot at 17-16 shows up in no highlight reel at all.

A Lottery Ticket and Broken Families

There is a part of this story data cannot measure, and I must address it differently.

Scouting networks in developing countries, Vietnam included, both find geniuses and manufacture lottery tickets. A twelve-year-old is spotted in a rural province, moved to a big training centre, and a whole family sees a route out of poverty. That is the lottery ticket.

Most tickets do not win. Among roughly 500 young players I have indirect data on through regional training centres, only a very small fraction reach national team level, and a smaller fraction still earn a living as professionals. The rest return with something hard to reverse: a childhood converted into ten years of training, and an interrupted education.

I have no solution. But I refuse to call this process 'youth talent development' without a number for the recovery rate. If a centre claims to produce champions, my question is: out of how many, and where did the others go.

This is why I always ask for sample size. Not because I like numbers. Because I want people to see those who are not counted.

The Regional Map

To understand Vietnamese badminton you have to place it on the regional map, because football and badminton in Southeast Asia operate as linked ecosystems.

Indonesia is a badminton power with dense club structures and a domestic market large enough to sustain professionals. Thailand built a world-class women's singles generation over fifteen years, backed by a domestic tour and large corporate support. Malaysia runs a strongly centralised national academy model with closed training and state funding.

Vietnam sits elsewhere. Its leading players mostly come through centres in Ho Chi Minh City and Hanoi, with fewer resources and fewer regular international opportunities. That produces a measurable effect: Vietnamese players experience far fewer 15-15 rallies in a year than same-age Indonesian or Thai peers.

In my dataset, Vietnamese players average 62 rallies at 15-15 or above per year of international play. The Indonesian group averages 134; the Thai group 148. Vietnamese players face less than half as many decisive situations.

This reverses the conventional reading. When people say Vietnamese players are weak in the decisive zone, they usually mean a psychological or character problem. My data points first to an opportunity problem. You cannot learn to play at 18-18 if you meet 18-18 sixty-two times a year while your rival meets it a hundred and forty times.

Endgame skill is an accumulation skill, built through repetitions. No repetitions, no skill. That has nothing to do with courage.

The Contrarian Angle: Correlation Is Not Causation

Here I must argue against myself, because analysis without self-critique is advertising.

My data shows Vietnamese players win less in long rallies and at high scores. The easiest explanation is fitness. It is convenient because it points to a clear remedy: train fitness harder. It is also wrong.

I tested the fitness hypothesis. If fitness were the issue, win rates in long rallies should decline steadily with match time. The data showed the opposite in a small subset: in long rallies during the first ten minutes of game one, the Vietnamese win rate was 43 percent; in long rallies during the last ten minutes of game three, it was 40 percent. A three-point gap.

If fitness were the main cause, that gap would be far larger, because the difference in energy reserves between minute five and minute seventy is enormous. Three points tells me the problem is not in the lungs. It is in shot selection - and shot selection is shaped by another variable: how many times a player has been in that situation before.

I must be careful with language. I am not saying fitness does not matter. I am saying that within this specific dataset and sample size, fitness does not explain most of the variance. I could be wrong. Give me finer movement data tied to match minutes and I will revise.

That is the difference between analysis and belief. Analysis can be corrected. Belief cannot.

A second contrarian point. People often say Vietnamese players need overseas training to improve. My data does not strongly support that. Of seven Vietnamese players I coded, four had extended overseas training spells. Comparing endgame performance before and after, I found clear improvement in only one case, a seven-point gain. The other three moved within statistical noise, under three points.

This does not mean overseas training is useless. It means it does not automatically produce results, and whether it does depends on something nobody measures: the quality of training partners, the number of live-pressure minutes, and whether the player is ever placed in decisive situations during practice.

Training where the opponents are strong but you never play at 18-18 is the same as training at home.

Signals for the Next Cycle

I do not predict. When the data is insufficient, I write that no conclusion is possible, and I accept that this makes my work less entertaining.

So this section is about signals, not forecasts.

Signal one: the count of rallies at 15-15 or above. If over the next twenty-four months Vietnamese players lift the 62-per-year figure above 100, I expect the decisive-zone win rate to follow - and that gain will come from choosing different tournaments, not from training more. It is externally observable. You only have to count.

Signal two: the unforced-error share of net rallies at level scores. My current figure is a 76 percent loss rate across the 38 specific rallies I counted. If that falls below 60 percent, it will show that coaching is changing behaviour, not just technique.

Signal three, the one I care about most: the number of tournaments a Vietnamese player enters within eight consecutive weeks. If it stays at three or higher, the injury invoice will keep being paid - and it will be paid precisely during the phase of a career when the player's market value is highest.

After twenty years of working with sports data, I have learned one thing: people always want to know who will win, and they almost never want to know why.

My job is answering the second question. The answers are long, dry, and usually skimmed. But they have one advantage no prophecy has: they can be checked, and they can be corrected.

I do not need to watch a match to know who ran more. Data does not sleep.

And if someone one day shows that my rally curve was drawn wrongly, I will redraw it. That is the entire difference between counting and believing.