Badminton's Instant Review System: The Real Cost Is Not the 90 Seconds

I start the stopwatch. Fourteen matches I logged myself during the 2026 season, thirty-eight challenges, and I recorded every second. The longest took 2 minutes 47 seconds, counted from the moment the player raised a hand until the umpire read the result. The shortest was 34 seconds. The average was 1 minute 22 seconds. Through all of it, nobody played badminton. The stands went quiet, the coach folded his arms, the player on the other side walked to the back of the court, drank water, wiped his face, and finished thinking about the next point.

Those thirty-eight reviews add up to 51 minutes and 56 seconds. I carried that number into a press room in Da Nang, where people argued for thirty minutes about one shuttle landing on a line, and nobody mentioned that the player had just spent nearly two minutes standing still. From the margins of the press room, I learned something the data never records.

The Instant Review System is the most argued-about thing in modern badminton, but its real cost does not sit in the seconds it consumes. It sits in the decision the system forces a player to make across the next fifteen points.

The rest of this piece is how I arrived there, and where I am still unsure.

Context: a system that answers exactly one question

The Badminton World Federation (BWF) brought the Instant Review System into official competition in 2026, with Hawk-Eye first appearing at the World Championships in Copenhagen. Before that year, disputes over a shuttle landing near a line were settled by the human eye and by the authority of a line judge. After that year, they were settled by a three-dimensional rendering.

The mechanism is tidy: each side gets two challenges per match. A successful challenge is retained; an unsuccessful one is spent. Technically, the system answers only a narrow set of questions — whether the shuttle landed in or out, whether it touched a player's body, and a few fault situations visible in a reconstructed image. It cannot answer the larger question: whether the decision honoured the spirit of the law.

Badminton has used rally scoring since 2026, replacing the old service-based system. Each game runs to 21 points, with an interval at 11, and in a deciding game the ends change at 11 as well. The structural consequence is concrete: a game contains only twenty to thirty genuinely decisive rallies, each lasting under ten seconds on average, and the whole game usually closes inside thirty-five to forty minutes. Removing one minute and twenty seconds from a structure like that carries weight. But I do not think that is the point. The point is that the system takes time, then takes the right to decide, and hands both back slowly.

The core: thirty-eight challenges, and what they leave behind

Of the fourteen matches I logged, twenty-four challenges came in singles and fourteen in doubles. The success rate in singles was nine out of twenty-four; in doubles, eleven out of fourteen. I do not rush to conclude from those two figures, because in doubles the shuttle travels faster and the sidelines are more often obscured, so a higher rate of line-judge error is predictable. But it opens a more interesting question: if error rates differ by discipline, challenge allowances should differ by discipline, and today they are identical.

What held me longest was the distribution across time. Twenty-nine of the thirty-eight challenges were first challenges. Only nine were second challenges. Put differently, across those fourteen matches the system effectively granted each side one decision, while the second was hoarded like an asset nobody wanted to spend. Players understand this precisely: the second challenge is worth more at the business end of a game, so they save it. A system designed with two challenges is operating as a system with one.

Then comes the part I cannot be certain about. I split the thirty-eight reviews by duration: those lasting over ninety seconds and those under sixty. In the over-ninety group, the challenging side won the very next rally thirteen times out of twenty-two. In the under-sixty group, the rate was six out of sixteen. I raise it because it shapes the question, not because it answers it. Thirty-eight reviews is far too small a sample, and I once taught myself an expensive lesson about believing beautiful correlations.

That lesson has a date. Russia versus Spain, 2026: I was not wrong, I was standing on the wrong side of the boundary of the data. I read attacking metrics closely and ignored that a team can choose to defend as a strategy rather than as a surrender. Since then, whenever I see a handsome number, I ask what it is hiding. With thirty-eight challenges, the number is hiding something simple: when a review runs long, the challenging side gets more rest, and in a sport where breathing rhythm is part of technique, more rest is a real advantage.

That leads to another observation, and here I differ from most commentary I read. People say the review system breaks a match's rhythm. It does break it. But it breaks it twice. First, during the waiting period. Second — and more heavily — inside the head of the opposing player, who just struck a shuttle he believes was in, and must now stand and wait to learn whether he will be believed. In a sport where psychological edges are often no wider than one correct call, hanging that call on a screen for nearly two minutes is a tactical intervention, even if nobody calls it one.

The counterintuitive part: one machine, two products

I used to think the review system was a technical problem to be fixed, like a slow clock that needs oiling. Then I looked at that 51 minutes and 56 seconds differently. If an average match runs about seventy-five minutes, then three and a half minutes of review is under five percent of total duration. Five percent is not enough to ruin a television product. It arguably improves it: a slowed-down rally, with graphics and a clean conclusion, is exactly what a director wants before a break.

In the end, the review system exists because television needs certainty. And the calendar exists because television needs inventory. The two are products of the same machine, and both take the same thing from players: their own time.

A World Tour season spans Super 1000, 750, 500, 300 and 100 tiers, plus the World Tour Finals, the World Championships, continental events and the Olympic Games. A player inside the top group of the rankings is bound to a minimum number of events each season, and the Super 1000 tier sits in the near-mandatory band. Added together, a year for a leading player can reach twenty tournaments. I have added that figure up repeatedly, because I do not trust a feeling.

An Se-young won the women's singles gold medal at the Paris 2026 Olympic Games. She walked into that final with a knee that had troubled her since 2026, and after taking gold she said plainly that her injury had not been handled correctly. Carolina Marin — who has endured three major knee injuries within five years — collapsed in the Paris 2026 semifinal, and He Bingjiao reached the final because Marin could not continue. Kento Momota, one of the finest controllers of match tempo I have ever filmed, was in a road accident on the way to the airport in Malaysia in January 2026, after winning the Malaysia Masters; he returned to the court, and never recovered that tempo.

Three cases, three different causes, and I have no intention of building them into a neat chain of causation. But they share a structural feature: all three happened to players carrying the densest schedules, and all three were managed inside a system where the tournament is the unit and the player's body sits outside the standings. People love to talk about load management. But load management, at most national teams, is a negotiation between the calendar, sponsorship contracts and ranking position. When those three parties disagree, the knee is always the weaker side.

And this is what I want to say clearly: the review system and the calendar belong to a single story. One revenue source, one product logic, one assumption that a player's body is a resource to be extracted until it runs dry. We spend one minute and twenty seconds establishing whether a shuttle weighing under five and a half grams landed inside or outside a line. We spend no minutes asking why that knee was still on court in the eighteenth week of the season.

At fifty-three, I know this: data is a map, not the territory. The map tells me how long thirty-eight challenges took. It does not tell me what a player was thinking while they ran. And I do not trust intuition, but I trust what intuition leaves out.

What to watch next season

A maximum duration for each review is the first signal I will write in the notebook. If the BWF sets a hard ceiling, a natural experiment begins: the same system, one different constraint, and we will learn how much of the cost comes from the technology and how much from the process.

The second signal is quieter: the count of retirements and mid-tournament withdrawals among the top ten players. It is an indirect measure, but far steadier than any claim about fitness.

And the third signal, the one I have waited longest for: whether a leading player dares to skip a Super 1000 event to protect a knee, and whether the price of doing so is lower than the price of a surgery. If the answer is yes, the negotiation has turned. At that point, badminton's next revolution will not come from a better camera. It will come from someone willing to delete a date from the calendar.