Speed does not buy medals: the data map of an elite badminton point

A final ended 21-11, 21-11. Looking at the scoreboard, people call it a one-sided, even boring match. But when I peel the scoreline away and re-stack the data by structure, it tells a very different story: not a weaker player being crushed, but a system being taken apart joint by joint. At Paris 2026, Viktor Axelsen beat Kunlavut Vitidsarn by identical 21-11 margins in both games. The number on the electronic board is very simple. The number inside the match structure is not.

What drew my attention was not that Axelsen won. It was how he won. That match was not decided by the fastest smashes of the tournament, but by something the eye struggles to name: the depth of his shuttle trajectories, the rhythm of his recovery to the central position, and the number of times his opponent was forced to return from a corner where no corner was safe. Look only at the scoreline and you see domination. Look at the structure and you see a problem being dismantled.

I spent years analyzing football through data, once sitting in a newsroom in Shanghai staring at a team's pressing metrics without understanding why that team won 4-0 and then lost to the bottom club three days later. That stumble taught me that results are never the only evidence. When I moved into badminton — a sport I had followed and commentated on for years — I met the same lesson again, only on a different court.

Badminton is one of the most measurable sports on earth, and also one of the most shallowly analyzed. The court is small, the distances short, each rally lasting a few seconds to a few dozen, and every shuttle contact can be attached to a coordinate. In theory, no sport is cleaner in its data. Yet most of what fans read still revolves around three things: who faulted on serve, who smashed hardest, and who won the title. All three are tier-one data — the tier of the scoreboard.

Since electronic line-calling systems like Hawk-Eye arrived at major events, badminton has held a vast quantity of positional data. Every shuttle landing is recorded to the millimetre, every serve checked against a virtual line. But more data has never meant more understanding. What is missing is not numbers, but a structure that puts numbers in the right place.

The reporting market I work in — China, where I live — has long had a habit: turning smash speed into a headline. A 400 km/h smash becomes a selling point, a short clip, a number to compare on social media. It is attractive, easy to grasp, fast to spread. But it answers the wrong question. Fans do not actually ask "how fast was that smash"; they want to ask "why did that person win". Those two questions are very far apart.

In an elite badminton point, data exists on three tiers, and the tier that decides the result is usually the one most often ignored. The first tier is the scoreboard: who won the point, by how much. The second tier is measurable behavioural data: shuttle speed, number of contacts, trajectory length, landing position. The third tier is rhythm — competitive state, breathing pattern, a player's stubbornness, something that appears in no spreadsheet yet multiplies the two tiers above it.

In other words: the scoreboard is the most readable thing, and also the most deceptive. What I call tier-two data — average rally length, the share of points won in rallies above eight strokes, the win rate in slow, high-trajectory exchanges, is the real map. And tier three is the reason two players with identical tier-two numbers walk into a third game and head in two different directions.

Let us start from what can be measured. In singles, rally length is one of the most important structural indicators. A player who lives on short rallies seeks quick endings: tight serves, early strikes, moving to the net, finishing. A player who lives on long rallies drags the opponent back and forth, waits for errors, and trusts their own stamina. Just by knowing the average rally length each side wants, you already know who must control the tempo.

In the Paris 2026 final I mentioned, what the scoreboard does not say is this: Axelsen did not win by stretching rallies out. Nor did he win by blindly shortening them. He won by mastering the height of the shuttle's path — steering it along a trajectory from which Vitidsarn could not attack. Badminton, in its technical essence, is a sport of controlling height and controlling angles, not a sport of controlling speed. And this is exactly where the public is misdirected.

Imagine a player hitting the fastest smash of the whole tournament. That shot is beautiful; it makes the crowd stand. But if that smash is launched from a position where the opponent is already standing in defence, the energy spent does not match the value returned. Conversely, a drop shot at only 180 km/h landing into a dead corner, where the opponent must take two extra steps, is the highest-probability scoring situation. Speed does not score. Position scores.

In football analysis, I long resisted treating metrics like expected goals as truth. Expected goals measures the quality of a situation, not the outcome. Badminton has a similar version that few people name: the quality of the situation a player creates. A trajectory that forces the opponent to return from a stretched, sideline, or falling-back position is a "high-quality situation" even if it is not a smash. A smash into mid-court where the opponent is already set is a "low-quality situation" even if it looks spectacular.

This leads to the problem of errors. In elite matches, points are usually not won but donated. Across many quarter-finals and semi-finals, the share of points a player gains from an opponent's unforced error is often significantly higher than points won directly with a finishing shot. The higher a player rises, the less they "win by hitting" and the more they "win by forcing the opponent to miss". This is the paradox that makes badminton hard to watch for newcomers: the feeling of explosion does not correspond to scoring value.

Let us move into the physical structure — the variable data often forgets. In football, I once found that after the 70th minute, a player's running distance ceases to be a pure fitness indicator and becomes a deciding one. In badminton, that band is compressed: the third game usually begins after both players have gone through roughly 40 to 60 minutes of collision. And in the third game, the first thing to decline is not power, but movement accuracy when recovering to the central position.

I call this effective movement distance. Not total distance, but the distance a player must cover to return to a ready stance for the next shot. A player who moves a lot but always returns to the right position is efficient. A player who moves less but is often stranded in a corner is eroding their own physical foundation. This is the kind of difference the eye struggles to see within the seven seconds of a rally, yet it becomes obvious in the third game.

This is also where the lesson from the pandemic period returns. With empty stands, matches became a kind of laboratory. Without the pressure of shouting, without the boost of a home crowd, players and teams had to generate energy from within. I once observed that in crowdless matches, football teams pressed higher in the data but with reduced effectiveness, because the mental tempo was no longer lifted by the stands. In badminton, a similar effect occurs in empty arenas: players attack more, but the error rate in decisive rallies rises.

Back to concrete technical structure. When analyzing a singles player, I usually separate four metric groups. First, the service pattern: short or high serve, to which zone, and how often the opponent attacks first off that serve. Second, the opening rally structure: after the first three strokes, who controls the height. Third, the conversion pattern: from attack into points, from defence into counter-attack. Fourth, the error pattern: where errors occur, in which situations, and whether they thicken at a particular stage.

These four groups combined yield a picture nearly impossible to read from the scoreboard. It explains why a player can win a match while losing almost every attacking metric, or why another player scores heavily with smashes yet loses at the key points. Key points are a category of their own. Not every point is worth the same. A point at 19-19 weighs more than one at 5-3, and a champion is usually not the best at every point, but the one who keeps their technical structure at precisely the heaviest points.

In doubles, the picture is even clearer. Doubles is a sport of rotational rhythm. A strong pair is not a pair with the two strongest individuals, but a pair with the smoothest positional transitions. When one player moves to the net to pressure, the other must retreat to the right gap. When one is pushed to the sideline, the other must cover the middle. Coordination errors at this level show up as gaps, and gaps in doubles are punished far faster than in singles.

This is why I do not trust evaluating doubles based on average smash speed. A pair with a lower average smash but correct positioning will always beat a pair with huge smashes but broken rhythm. A smash only has value when it is launched from a structure that has already channelled the opponent into a disadvantaged position. Without that structure, the smash is just a pretty, wasted exchange.

This connects directly to a viewpoint I consider most important: correlation is not causation. People see that champions often have strong smashes, and conclude the smash made the title. But the causal order may well be reversed: precisely because they controlled the match structure, the player had the time and position to launch the smash from a high-quality situation. The smash is the result of control, not its cause. Mistaking the two is the fastest way to misread a player.

I have made exactly that mistake, only in a different sport. In the summer of 2026, I praised a team for a big win without noticing that their pressing metric was actually low — the opponent defended deep, so the problem was never exposed. Three days later, that team lost to the bottom club. Shanghai 2026 is not a scar; it is a map that redrew how I look at numbers. The lesson became a rigid checklist: every analysis must have at least three advanced metrics, and a match result must never be the only evidence.

I applied that checklist to badminton. And the result repeated: scoreboard shocks almost always have their solution at the structural tier. A player rated low who suddenly beats a top seed usually does so not through luck, but by breaking one specific link in the opponent's system — often the service link or the central-recovery link. And the defeats of title contenders rarely come from sudden drops in form; they come from a crack that already existed, waiting for a match heavy enough to expose it.

A system does not collapse in one night; it cracks from the moment I stop questioning the foundation. That holds for both teams and players.

There is something else data cannot measure, and I want to speak plainly about it. The pressing ghost does not appear in the data, yet it still makes opponents step up and then fall apart. In badminton, that "ghost" is the psychological pressure of knowing an opponent can end the rally at any moment. It sits in no statistical cell. It sits in a player choosing a high serve instead of a short one at 18-16, in hitting a safe trajectory instead of the one they wanted, in shoulders tightening slightly in a rally the eye skips past.

This is why I say the third data tier — rhythm — is where I close the story. A number only tells part of the story; the rest I hear with ears once burned by arrogance. A player may post very high attacking numbers, but if their rhythm breaks at the last two points, all those beautiful metrics become noise. Conversely, a player with modest numbers who holds their rhythm through the third game is the one who takes the medal.

Be careful here: I am not saying data is useless. I am saying tier-one data is useless on its own, and tier-two data is meaningless unless read with tier three. The arrogance of a data person is believing everything can be modelled. The cowardice of a data person is abandoning the model entirely upon meeting an unmeasurable variable. Both are errors.

I also want to address the most overlooked dimension in any badminton analysis: how a tournament compresses emotion. In a major-tournament cycle, the pressure does not come from a single match, but from having to hold structure across many consecutive days. That is when questions of squad depth and physical depth matter more than any beautiful exchange. Format and schedule, not just technique, decide who is still standing at the end.

There is an empirical detail I remember well. Reviewing many knockout matches at a major international tournament, I noticed the share of matches whose result diverged from a model based on situation quality spiked in matches extended through added time. Not because the model got the technique wrong, but because it could not account for physical decline and home-field effect. In badminton, when a match stretches to a third game, this logic is even clearer: models based on first-half situation quality become systematically less accurate in the second half of the third game, because they do not grasp each player's recovery speed.

That is why a probability model needs to be challenged by a human. An algorithm can compute the probability that a specific situation leads to a point, but it cannot compute a player's composure at 19-19 after 70 minutes of collision. A player's tactical instinct — what we often call instinct — is really an internal form of data accumulated over thousands of rallies. It does not replace the model. It supplements the model precisely in the blind spots the model cannot cover.

The story of Mikel Merino in football — a stoppage-time goal from a situation the model assigned a very low probability — is proof of this. That header was not a miracle. It was the result of a player holding the right position in a moment when all calculations had forgotten him. Badminton has countless similar moments: an apparently lost retrieval, an apparently harmless trajectory, turning into the deciding point. Those moments do not break the data. They simply show that data always has a boundary.

So when I read a player, I always ask myself three questions. In what context was this metric measured? Does the opponent have a style that counters it? And how does the player react when cornered exactly on their weakness? These three simple questions eliminate most of the flashy yet useless analysis. They are also the filter I use before writing anything.

From a market perspective, storytelling through numbers is not merely academic. It affects how players are valued, how tournaments are sold, and how brands choose their representatives. When the public is taught that smash speed is the measure of worth, a player who controls structure but lacks flashy shots gets valued below their true worth. This is a form of market distortion, and it repeats in every sport.

I have followed how contracts are shaped by headline numbers rather than structural numbers. In some arrangements, rewards never appear in the transfer figure yet last longer and are harder to monitor. This turns a surface metric into an anchor, while foundational quality — hard to package into a headline — is pushed to the margin. Badminton, though smaller in financial scale, follows exactly that path: the smash sells more than the drop shot, even though the drop shot wins more.

I do not mean to dismiss the value of attack. Attack is a way to control. But effective attack is attack built on structure, not attack replacing structure. A player who understands this allocates energy differently: fewer showy shots, more patient rallies, and explosive moments only when structure has cleared the path.

At this point, if I had to pick one number to read a player, I would not pick smash speed. I would pick the share of points won after defensive shots convert into counter-attacks. That is a metric hard to beautify, hard to clip, yet it reflects exactly what decides medals: the ability to turn weakness into opportunity.

So what is the signal for the next cycle? I believe value will shift toward players who master the height of the shuttle and control the central position, while retaining the ability to restore rhythm in the third game. Analysis teams that learn to measure "situation quality" rather than merely counting strokes will gain an edge over rivals still stuck on tier one. And players who understand that medals are awarded for structure, not for speed, will be the ones who last longest at the top.

What remains is a question I do not fully have an answer to, and I will not pretend to. When analytical tools become so common that every professional player has their own dashboard, will the data advantage disappear, and will badminton return to a place where only human tactical instinct decides? If that happens, the final winner will not be the best reader of data, but the one who knows where their data ends — and knows with which sense they must hear the rest.