The Annual Badminton Season: When Long Rallies Become a Weapon and Smash Speed Hits Its Ceiling

In the deciding game of a Japan Open quarter-final in Yokohama, at 19-19, I clocked a rally and counted 31 racket contacts. Thirty-one touches, lasting 42 seconds, ending with a cross-court push that landed tight to the sideline. The arena stood up. I sat down, reopened my dataset, and realised that rally was not an outlier. It was the pattern.

Across four consecutive weeks of charting at BWF World Tour level, over 46 matches I timed by hand, average rally length rose from 8.4 strokes to 11.2. The share of points won in rallies longer than 15 strokes climbed from 21% to 34%. The number of smashes per game fell by roughly 9%. When a sport slows down at the data layer, the cause is rarely that players have become more passive. It is usually that somebody has worked out how to make speed worthless.

Method: three crude metrics and one notebook

People read badminton with their eyes. A smash above 400 km/h is called power; a thirty-stroke rally is called quality. The eye retains feeling but not frequency, which is why I work with three crude metric groups anyone with a laptop could rebuild: rally length, point distribution by court zone, and unforced error rate.

The current cycle is the annual season. No single tournament is big enough to define the year. The BWF World Tour runs from Super 1000 down to Super 100, with more than thirty events per season. A player inside the world's top twenty typically plays 16 to 20 events, plus national team ties and ranking qualification rounds. There is no final to save everything for, and no off-season long enough for full recovery. The real currency of the annual season is therefore not technique. It is the ability to allocate stamina.

Based on my experience tracking matches over nine years, the annual season has a clear rhythm. In the early months, players compete with technique and freshness. By October and November, when the body has accumulated fatigue and the ranking table is close to locking, matches change shape: less risk, more rallies, and points saved like assets.

I chart on site, by hand. No software pays me for data at the edge of a Super 500 or Super 300 court, so I still work the old way: a notebook, a pen, a stopwatch. Because I chart by hand, I am forced to be selective about what I measure. What I choose to measure is racket contacts, shuttle landing zone, and the nature of lost points.

Rally length: a tax levied on career longevity

In my 46-match dataset, game one averages 9.1 strokes per rally. Game three averages 13.4. The share of rallies exceeding 15 strokes is 18% in game one and 41% in game three. This is not a quirk of this season. It is the structure of the sport when two players of equal level know each other too well.

What caught my attention was not the rise itself but its consequence. I timed recovery between points. After a rally under 8 strokes, the average interval is 14 seconds and heart rate returns close to baseline. After a rally over 15 strokes, the interval is still 14 seconds — the rules permit no more — but the breathing has not settled before the shuttle is served. Long rallies do not win points directly; they win by pushing an opponent into a state of stamina debt, and collecting that debt in the following game.

Kodai Naraoka is the clearest example of this mechanism. His 2026 World Championships final against Kunlavut Vitidsarn lasted 109 minutes, and notably Naraoka won the opening game after a string of long rallies, then collapsed in the third by a wide margin. Commentators said he ran out of gas. I say he paid a debt he had created himself.

Smash speed hits its ceiling

In men's doubles, peak smash speed passed 400 km/h years ago, and pushing it a few kilometres per hour higher no longer changes outcomes. Across the 46 matches I tracked, the rate of winning a point directly from a smash was around 28%. Nearly three quarters of smashes do not end the point.

Worse, I isolated the group of smashes that were returned. After a smash is blocked or driven back, the attacking player has roughly 0.7 seconds to recover position. In that window, more than half of the next points were scored against the player who had just smashed. The smash is no longer a finishing weapon; it has become a high-risk loan, and the interest is charged in abandoned court space.

That is why elite players now smash less but more selectively. They no longer smash to score. They smash to force a lift, then use that very rally to re-establish the pattern. The 9% drop in smash volume in my data is not a sign of timidity. It is a sign of calculation.

Unforced errors: the discriminating metric

I separated out one category of lost point: the unforced error, meaning a shuttle into the net, out of bounds, or missed entirely in a situation where the player was not under direct pressure. This is the only metric in my trio that correlated clearly with results.

Across 46 matches, the winner had a lower unforced error rate than the opponent in 39. That is 84.8%. A higher net-point win rate than the opponent predicted the correct result in 41 matches. A higher smash speed than the opponent predicted correctly in only 25.

In other words, my data does not say the hardest hitter wins. It says the player who makes the fewest mistakes inside the longest rallies advances. And because the annual season stretches matches out, the value of not erring rises week by week.

The other side: correlation is not causation

Here I have to refute myself before someone else does.

There is one technical variable I cannot control, and it may explain much of the rise in rally length: the shuttle speed grade selected for each tournament. Organisers and the referee adjust shuttle speed class to the temperature and humidity of the hall. Hot, humid air makes a shuttle fly faster, so a slower shuttle is chosen to compensate. Cold, dry air makes it fly slower, so a faster shuttle is used.

Part of the longer rallies in the window I charted may simply be a consequence of sitting in different halls, in different weeks, with different shuttle grades. I do not have on-site humidity verification for every session. That is a hole in my model, and I write it down rather than cover it up.

The second variable is the calendar. When I observe four consecutive weeks, I am observing a window in which many players have already contested their third or fourth event of the month. Fatigue reduces smash volume and increases long rallies. That is a calendar effect, not a tactical one.

The third variable is sample size. Forty-six matches is a small sample, concentrated in one region and one period. I do not have enough data to declare this a trend for the whole sport. I only have enough to say that within my observation window the trend exists and deserves tracking.

One more thing needs stating plainly: home court was never an advantage, only noise encoded into points. In Yokohama the home crowd rose for every long rally. But in my data, the rate of winning the point immediately after a long rally was no higher for Japanese home players than at neutral venues. Noise shakes the stands; it does not make the shuttle land more accurately.

Signals for the next cycle

People need belief to place a bet; I need data to be certain. For the coming rounds I will track three things.

First, net-point win rate. If this metric keeps predicting results better than smash speed, I have grounds to say the sport is shifting from a contest of power to a contest of position.

Second, unforced error rate in third games. This is where stamina becomes error, and where my model could be overturned if a player proves they can hold accuracy to the final stroke.

Third, hall conditions. If I can record temperature and humidity for each session, I can separate the tactical component from the physical one. Until then, I must concede that part of the rise in rally length may not be human at all. It may be the air.

I do not remember matches; I remember the heat map of the match. Every rally is a statement, every number a confession.

In that corner of the Yokohama hall, after the 31-stroke rally, the winner did not shout. He bent down, picked up the shuttle, handed it to the umpire, and walked slowly back to the service line. That is the behaviour of someone who knows he has collected something more valuable than a point. If the annual season keeps following this trajectory, the most expensive thing on court will no longer be the fastest smash. It will be the capacity to endure a long rally without missing.