What Tennis Cannot Measure: The Void Beneath the Arthur Ashe Roof

What Tennis Cannot Measure: The Void Beneath the Arthur Ashe Roof

1. A night with nobody in the stands

On the evening of 13 September 2026, I walked into Arthur Ashe Stadium through the press gate carrying a paper notebook instead of opening a laptop. Inside a roof that covers 23,771 seats, not a single chair was occupied. Dominic Thiem and Alexander Zverev walked out under the lights, and the sound of rubber soles scraping the hard court was so clear that I could hear the rhythm of their footsteps from the technical area.

Zverev won the first two sets 6-2, 6-4. Thiem reversed the match, took the next two 6-4, 6-3, and forced a fifth-set tiebreak. It was the first US Open final ever decided by a final-set tiebreak, and Thiem became the first man born in the 1990s to win a Grand Slam singles title.

Afterwards I held the official match sheet. It contained everything this industry habitually calls data: first-serve percentage, points won on first serve, points won on second serve, break points created, break points converted, winners, unforced errors, net approaches. There was no line recording that when Thiem collapsed onto the court, the silence was so complete that his breathing became the loudest sound in the building.

When the stands are empty, you hear the breathing of the match more clearly. And precisely because of that, you notice how much of the story the dense statistics sheet leaves out.

An empty stadium does not merely lack noise — it lacks the story being told.

That night is the starting point of this piece. It is not a tribute to Thiem, nor a criticism of Zverev for surrendering a two-set lead. It asks a narrower and more uncomfortable question: if the entire measurement infrastructure of modern tennis — Hawk-Eye, sensors, motion tracking, hundreds of live-updated metrics — still fails to record what actually decides a match, then what are we measuring?

2. The architecture of tennis data

Hawk-Eye first appeared at the US Open in 2026, supporting officials on selected courts, and was deployed across the main show courts in 2026. That was the turning point: for the first time, ball trajectory was captured in three-dimensional coordinates rather than by the human eye. Every serve could now be reconstructed, timed, and measured for spin and landing point.

Twelve years later, in 2026, a 25-second serve clock was trialled at the US Open and quickly became the ATP standard. Another variable was brought inside the frame: time itself.

Meanwhile, real-time analytics populated television broadcasts and mobile apps. Viewers at home could learn which player won more rallies beyond seven shots, who performed better serving wide, who struggled when forced to move to the left before hitting a forehand.

The easy conclusion is that tennis is undergoing a data revolution. But standing on site rather than sitting in an analytics room, you notice a fracture.

In March 2026 the season stopped. Indian Wells was cancelled. On 1 April 2026 Wimbledon was cancelled — the first cancellation since 2026. The US Open went ahead from 31 August to 13 September 2026 without spectators. In 2026 the Tokyo Olympics ran from 23 July to 8 August with virtually every venue closed to fans.

Throughout that period, the cameras kept running. Hawk-Eye kept recording coordinates. Software kept reconstructing rallies. The machines never stopped. But one data column collapsed to zero across the entire system: the sound of the crowd.

That was the moment the industry realised it had overlooked an enormous variable. For decades, every recorded metric had operated under a silent assumption: that the match was played in front of a crowd. When that assumption disappeared, the numbers remained technically correct but ceased to be meaningful.

3. Four things that never reach the stat sheet

3.1 Intention before contact

In 2026 I sat in a corner stand at Nizhny Novgorod for Croatia against Argentina at the World Cup, rather than in the commentary box, because from there I could see Luka Modric's movement without the ball. In the 80th minute he scored. In my notebook I wrote: "Modric is not the fastest runner, but every step he takes carries intention."

Modric is not the fastest runner, but every step he takes carries intention.

Applied to tennis, the line holds almost intact. A returner does not merely need to know where the ball is going; he needs to know where his opponent is standing, which way he is leaning, and whether he is retreating behind the baseline. The decision about where to hit is made before the ball leaves the opponent's strings, based on a chain of signals no camera reduces to a metric.

Football analytics has tried to convert intention into numbers through metrics such as PPDA — passes allowed per defensive action in a third of the pitch. PPDA measures pressure, but it does not measure hesitation. A team that waits half a second before engaging can ruin an entire pressing plan, and no index captures that half second.

Tennis has exactly the same problem. A lob does not always exist to win the point; sometimes it exists only to buy time to recover balance. A serve wide at 0-30 is not aimed at a direct winner but at planting a false assumption in the opponent's mind for a later game. These are purposeful behaviours, and they appear in none of the data feeds handed to audiences.

When I review Grand Slam final footage as a documentary screenwriter, I usually have to reconstruct this layer of intention by hand. I tag each rally, note why I believe the player chose that option, then cross-check against what followed. No software does this for me. It is slow manual work, and that is precisely why it reveals more than any dashboard.

3.2 The gap between points

For most of a tennis match, the ball is not in play. Between points there are roughly 20 to 25 seconds under the clock, plus changeovers and set breaks. Those intervals occupy a large share of total match time and are barely recorded.

I call it the gap between points. What lives there? Breathing. The way a player wipes his face with a towel. Where he looks while the opponent prepares to serve. Whether he chooses new balls or old balls to serve with. Brief exchanges with the umpire, or with a coach where on-court coaching is permitted.

In documentary terms, this is the most valuable material. A player standing still, head down, taking three slow breaths before serving at break point tells you more about pressure than any number. In data terms, that interval is blank space — or worse, a meaningless metric labelled "time between points".

3.3 Pain that never enters the record

Tennis allows a player to call medical staff onto court. History calls it a medical timeout. The event is recorded; the duration is recorded. The degree of pain is not.

There is no metric for a player with a sore back choosing to serve flat rather than kicking, purely to reduce the number of times he must bend. There is no metric for a player with blistered feet altering his lateral movement. There is no metric for a player declining to call the trainer because he fears signalling weakness.

This is the largest blind spot in sports analytics. Every movement is measured; the motive for enduring pain is not. The best matches I have watched all featured one or both players competing far from full health, and the stat sheet never tells the audience.

3.4 The heartbeat of the stands

Football does not live on goals — it lives on the heartbeat of the crowd. Tennis lives on the same thing, compressed into each point.

2026 and 2026 were a natural experiment on that assumption. With empty stands, tennis remained technically itself, but the rhythm of matches changed. There was no noise to draw on in the fourth set. There was no crowd pressure to disrupt an opponent's serving rhythm. Players told me in backstage conversations that they had to generate rhythm internally, and many realised nobody had ever taught them how.

That is a neglected tactical signal. If the crowd is part of the competitive environment, it belongs in the game plan. Its absence is a variable that can be trained for, not merely endured.

4. The Raducanu case: complete data, invisible story

In September 2026, Emma Raducanu won the US Open at 18, coming through qualifying, winning 10 consecutive matches without dropping a set, ranked outside the top 100 when the tournament began. She was the first qualifier to win a Grand Slam singles title, and the first British woman to do so since Virginia Wade in 2026.

Her statistics were strangely modest. She did not own the fastest serve in the draw, the fastest forehand, or the highest winner count. By raw data, she was not the standout player in New York that year.

Which is exactly why her case is the perfect stress test for the limits of tennis data. What Raducanu did best was not speed or power. She shifted her return position between games — sometimes stepping inside the baseline, sometimes retreating deep — forcing opponents to recalculate on every serve. She varied tempo between points within the same game. She kept an almost unchanging facial expression across two weeks, making her emotional state nearly unreadable from across the net.

These are real tactical skills, observable, coachable, and directly linked to results. Yet they barely register in any orthodox index. Analytics can tell you what percentage of second-serve points Raducanu won in the semifinal. It cannot tell you why her opponent lost rhythm in the seventh game.

As a documentary maker, I learned one thing from her run: when data fails to explain a result, that is not evidence the result was luck. It is evidence the data is measuring the wrong thing.

5. The counterintuitive view: the problem is not a lack of data

Here I must say something plainly that may irritate many in the industry.

The prevailing framing — that sport needs more data in order to understand matches better — is a misdiagnosis. We are not short of data. We are drowning in it.

A single ATP-level match now generates thousands of data points. A coach can spend a full week reading pre-match reports and still not absorb them all. The problem is not volume but distortion: the industry measures what is easy to measure rather than what matters.

First-serve percentage is easy to measure. Choosing the right serve type at 30-40 is important but hard to measure. Winner counts are easy to measure. Deciding to hit down the middle to reduce risk in the fifth set is important but hard to measure. When an industry builds its entire evaluation system around what is easy, it gradually redefines excellence in its own image.

The consequences have been visible in football for over a decade. Inverted wingers became the coveted profile because they generate conveniently packaged numbers: goals, assists, touches in the box. The traditional winger, who earns his living by stretching the pitch horizontally and creating space for others, has been marginalised — not because he is less effective, but because his effectiveness does not translate into an index.

Tennis is on the same road. Surfaces have been slowed and brought closer together in character. Net-rushing has all but gone extinct at the highest level. Players with idiosyncratic styles are pushed down the valuation ladder because they produce ugly metrics, even when they win.

There is a second, market-level consequence. When data becomes the sole measure of value, the price of young players is inflated on the basis of quantified potential rather than verified achievement. A player who has never gone deep at a major can be valued on modelled projections. That is the structure of a bubble, and bubbles always end the same way.

6. What remains when the computers switch off

I returned to Arthur Ashe many times after 2026. When the crowds came back, the noise returned, and every metric resumed its ordinary value. But that empty night remains my reference point every time I read a match data sheet.

It taught me something simple: every measurement system has a boundary, and most of the story of sport lies outside it. The writer's job is not to reject data but to stand at that boundary and point to what is being left behind.

If there is one question I would leave with those building the future of sports analytics, it is this: if a Grand Slam final is played without spectators, and nobody records the moment a player generates his own rhythm in the fifth set, are we preserving a body of knowledge about the sport — or only its shell?