The Hawk-Eye on the Badminton Court: When a Point Depends on the Completeness of Data

In the deciding game of a Super 1000 quarter-final, the score stood at 19-18. The player unleashed a cross-court smash, and the shuttle buried itself just beside the sideline. The line judge raised a hand to call it in, but no one in the stands could be sure. On the arena's big screen, a three-dimensional model began to reconstruct the shuttle's trajectory, frame by frame, slowly, as if time itself were being stretched. I have sat in the analysis room, rewinding that very frame hundreds of times, and each time a line I tell myself resurfaces: Every rally tells three stories: one of the camera, one of the technology, one of history. The camera's story is what the spectator sees. The technology's story is what the algorithm computes. History's story is what the umpire must write into the record. Those three stories do not always agree, and the gap between them is where every controversy is born. In badminton, where every point carries equal weight and a point at 19-18 matters far more than one at 5-3, that gap can decide an entire match. For more than a decade, professional badminton has shifted from a sport of the human eye to a sport of the machine eye. The Badminton World Federation (BWF) operates the Instant Review System, built on Hawk-Eye technology, allowing players to challenge calls on close line decisions. The mechanism is simple: each player or pair has a limited number of reviews per match; a successful challenge preserves the count, an unsuccessful one costs a review. This is closer to the challenge model in tennis than to a comprehensive review initiated by officials. The wider context lies in the tournament structure. The BWF World Tour is tiered, with Super 1000 as the top group — names like the All England, Indonesia Open and China Open — and Super 750 including the Japan Open, Denmark Open and French Open. The bigger the event, the more cameras, the higher the resolution, and the greater the audience's expectation of accuracy. It is worth recalling that since 2026 badminton has used rally scoring to 21 points, best of three games. That change made every point more precious and turned each officiating decision into an event with real weight. In Japan, where I work, the story carries an extra layer. Fans here are used to a culture of analysis, in which every match is dissected into metrics: smash speed, distance covered, number of sprints, scoring rate on serve. When they see a controversial review, their first reaction is to ask about the input data — whether the cameras had enough angles, what the margin of error was, whether the model lost any crucial frames — rather than to boo. That is a very different mindset from much of the world, and it is also why technical analysis finds a home in this market. The review system operates as a model, and every model depends on the quality of its input data. When the data is complete, the model is nearly infallible. When the data is thin, the model becomes a machine for manufacturing false certainty. Hawk-Eye works by using multiple cameras around the court to capture the shuttle at different instants, then applying interpolation algorithms to rebuild the three-dimensional trajectory and identify the point of contact. With enough cameras, enough angles and enough frames, the result is almost perfectly accurate. But if an angle is missing, if a camera is blocked, if the shuttle is hidden by a player or the net post across a few decisive frames, the algorithm is forced to guess the missing part. And when the algorithm guesses, error appears. This is the point I want to decode. Fans watch the screen and see a smooth curve, a clear contact point, a decisive conclusion. They do not see that behind that curve lie hundreds of calculations, a few lost frames, and a margin of error the organizers must accept. The smoothness of the image creates a feeling of certainty, but that feeling does not always match the true reliability of the data. While following matches at the Super 1000 and Super 750 levels, I have noticed a recurring pattern. The fiercest controversies rarely come from clear-cut situations. They come from rallies where the shuttle lands within a millimetre or two of the line, where both the human eye and the model operate at the edge of their limits. At that distance, the question of in or out stops being a question about truth and becomes a question about how the margin of error is defined. This explains why leading players often seem strangely calm when they challenge. They are not arguing with the umpire. They are betting on a model. And they understand that, at the model's edge, the odds are barely better than a coin toss. Players such as Kento Momota, Viktor Axelsen and Akane Yamaguchi, through years at the highest level, have learned to distinguish between a wrong decision and a correct decision that cannot be proven. The difference is subtle, but it shapes how they react. A word is also due on the metrics the system generates. Smash speed, distance covered, number of sprints — all are attractive figures that impress easily. But they do not automatically equal effectiveness. A player can cover more ground than an opponent and unleash more powerful smashes and still lose, because those figures measure effort rather than outcome. This is the trap any sports-data analyst must guard against: running more does not mean running well. For the Japanese market, this layer is even clearer. Fans here do not merely want to know who won; they want to understand why. They read the data tables, compare matches, and draw their own conclusions. That places a higher demand on analysts: not to pile on numbers, but to identify which figures truly matter and why. A good analysis is the one that helps the reader see something they had never seen, rather than the one with the most numbers. Here I want to push against a popular belief. Many assume that more cameras, higher resolution and higher frame rates will make controversy disappear. I do not believe that. The problem is not the volume of data but how people interpret that data under emotional pressure. An umpire may see exactly the same frame as the model, yet still feel the weight of the crowd, the score, and the memory of past criticism. A player may be convinced the shuttle was out — not because they saw it, but because they need it to be out to protect the point. Emotion does not make the data wrong, but it makes people read the data in their own favour. Data never panics. People blind themselves when they rush into emotion. There is another paradox: the more technology, the higher the expectation, and the greater the disappointment when it fails. When a system is marketed as accurate, people assume it must be right in every case. But no system is right in every case. When it fails, trust collapses far faster than if it had never existed. This is the trap every sport adopting technology must face. I once witnessed a situation in which I myself could not rule decisively. In a group-stage match, the shuttle landed on the line, and the model returned an in result with a margin of less than two millimetres. The organizers announced the call, the crowd applauded, and the match went on. But in the analysis room, we all knew that shifting the camera angle slightly might have changed the outcome. No one was wrong. The data was simply not complete enough for a certain conclusion, and people had to live with that uncertainty. Perhaps what badminton needs is transparency about limits rather than more cameras. An honest system is one willing to state clearly when it is unsure, instead of claiming to be always right. If every review came with a confidence figure, fans would learn to trust data without worshipping it. And who knows — that humility might be the biggest step forward the sport can take.