Empty Stands, Empty Spreadsheets: What Tennis Cannot Measure

On March 8, 2026, I stood in the backstage area of the Indian Wells Tennis Garden and heard a sound that had never existed there before: a tennis ball bouncing on an empty practice court, inside a complex built for tens of thousands of people. That morning the signage was still up. The vendor stalls were already built. Security staff still stood at the gates. The tournament's data package had been finished days earlier: player profiles, hard-court win rates, head-to-head records, finals reached, everything an analytics department needs. Early that afternoon, an announcement came: the tournament was cancelled, three days before the main draw, after a confirmed case in the area. No matches were played. I remember standing there a long time, looking at that data on my phone — a spreadsheet full of information about matches that would never happen. It took me a while to understand what I was looking at: a complete measurement system for something that had disappeared. Still there. Still accurate. Still clean. With nothing left to measure. THE SHUTDOWN Ten days later the entire professional tour stopped. Miami was cancelled. The ATP and WTA suspended the calendar. On April 1, 2026, Wimbledon announced its first cancellation since World War II, ending more than seven decades of unbroken play. Roland Garros was pushed from May to late September, a unilateral decision that caused weeks of friction inside the sport. The US Open went ahead without spectators, and Cincinnati was relocated to Flushing Meadows to share the same bubble. For an analytics department, this was close to ideal conditions. You removed the loudest variable in sport — the crowd. You got matches played in silence, where every contact was audible, where no roar distorted the sense of rhythm. You got a cleaner dataset than anything that had ever existed. That is exactly when the problem started. THE CLEANEST DATASET IN HISTORY, AND THE LEAST REMEMBERED SEASONS Look at Novak Djokovic's 2026 season. He arrived at the US Open 27-0 in Grand Slams that year, having won the Australian Open, Roland Garros and Wimbledon, one match away from something no men's player had done since Rod Laver in 2026. Daniil Medvedev beat him 6-4 6-4 6-4 in front of a full Arthur Ashe Stadium — the largest arena in tennis, nearly twenty-four thousand seats. I am not turning that coincidence into a law. One loss proves nothing about crowds. But I followed Djokovic all season, and what stopped me was the structural contrast: through the months when he was nearly unbeatable, most of his matches were played in spaces with limited human sound. When the stands filled again, the perfect run ended. If a near-perfect statistical season ends with a single defeat in the place with the most people, then the difference was not in the statistics. I have rewatched a great deal of tennis from 2026 and 2026. Technically it was not bad. The quality on show at Roland Garros 2026, played in October cold with strictly limited crowds, or at the US Open 2026, was no lower than any other edition. But I cannot recall a single point from those events. I recall the scoreline. I recall the winner. I cannot recall the feeling. The memory of a match does not live in the data, and the data does not know that. The crowd is not decoration. It is part of the sport at a physical level: it sets the length of the pause between points, it changes a player's breathing before a serve, it creates a pressure no practice session can simulate. A player serving at 40-40 in front of fifteen thousand people standing does not serve the same ball as that player serving in an empty arena. The physiology may look similar. The decision does not. In data models, the crowd is usually treated as noise — a variable to strip out so the signal comes through clean. I think that is the biggest mistake tennis analytics has made in a decade. We removed precisely the variable that makes the sport what it is, then wondered why the models could not predict the big matches. FOUR METRICS AND ONE SWING NOBODY COULD PREDICT Every professional tennis stat sheet revolves around the same family: first-serve percentage, points won on serve, points won on return, break-point conversion, winner-to-unforced-error ratio. These are good metrics. They describe accurately what happened. They describe results, not intentions. A player who misses a first serve at 30-40 in the fourth set is recorded in the first-serve column. The sheet does not record that he aimed at the T because he saw his opponent shift his return position after the previous point, and that the miss was a correct decision executed a few centimetres wrong. The difference between a wrong decision and a right decision executed badly is the entire content of this sport. It has no column. On September 6, 2026, in the fourth round of the US Open, Djokovic faced Pablo Carreño Busta. After losing a point early in the first set, he swung a ball behind him. It struck a line judge in the throat. The match ended there: Djokovic was defaulted, stripped of the prize money he had accumulated and removed from the tournament. It was the first time in the Open era that a top seed had been defaulted from a Grand Slam. No model could have predicted a swing of the racket unrelated to the score, and that swing decided the tournament. A SYSTEM BUILT TO ALWAYS HAVE AN ANSWER In my documentary work I once received a tennis analysis built across nine dimensions: technical and tactical, data and form, tournament systems and scheduling, tour landscape, rules and governance, team management, risk, media narrative, and industry transmission. A complete, rigorous, professionally sound framework. It had no input. No player. No tournament. No data. No dates. Nine dimensions, every cell marked: insufficient information. And still, at the end, the system produced five risk flags, four signals to track, and a four-criteria information-value rating. All at zero. A document with nothing to say had generated something that looked exactly like a finished report. The system cannot say "I don't know" at the global level; it can only say "I don't know" cell by cell. I recognised that logic immediately, because tennis runs on it. The match preview always has five keys. The power rankings publish on days with no play. The television segment still fills thirty seconds of analysis even when the match just played offers nothing to analyse. The format never leaves a blank. When the format becomes the product, emptiness gets presented in the language of completeness. In twenty-five years, the most useful reports I have read were a rare type: the ones willing to state that there was not enough evidence to conclude. That is not a failure of analysis. That is analysis at its most mature. It just does not produce an impressive-looking page. THE PLAYER WHO LEFT AT NUMBER ONE On January 29, 2026, Ashleigh Barty won the Australian Open, beating Danielle Collins to become the first home women's champion since Chris O'Neil in 2026. She was world No. 1. She had held the ranking for 114 consecutive weeks. She was twenty-five. Every projection pointed to years more. On March 23, 2026, she retired. No injury curve predicted it. No ranking projection predicted it. No performance table predicted it. Barty left at the peak, for a reason no sports dataset is designed to record: she had had enough. A spreadsheet can tell you who is leading; it cannot tell you who is finished. The same year, Roger Federer ended his career at the Laver Cup in London on September 23, after a run of knee injuries that every statistical column logged and none could explain in timing. Serena Williams played her final event at the US Open. In November 2026, at the Davis Cup in Malaga, Rafael Nadal stopped too. In under thirty-two months, the four most-documented players in the history of the sport left the court. The data recorded their decline in fragments — ranking slides, withdrawals, retirements mid-match. The decisions came from somewhere else, and no analytics department has a map of that place. A GRAND SLAM ERASED FROM THE SYSTEM THAT MEASURES IT On April 20, 2026, Wimbledon announced a ban on Russian and Belarusian players. A month later, on May 20, the ATP and WTA confirmed the tournament would carry no ranking points. Let that settle. The most prestigious event in tennis would be played with no ranking value. Djokovic won it that year and still slid down the rankings, because his two thousand points from the 2026 title came off with nothing to replace them. Players who won the most matches at Wimbledon 2026 walked away with lower point totals than they arrived with. A tournament can be erased from the very system that measures it — and the winner is still the winner, but the ranking does not know it. It is the sharpest example of something the sport keeps forgetting: data in tennis is not a natural object. It is an agreement. Someone decides what counts, what does not, and in this case an entire tournament was declared not to count. The same holds for integrity. In 2026, two of Jannik Sinner's samples returned positive for clostebol in March. An independent tribunal cleared him in August. WADA appealed to CAS. By February 2026 a settlement produced a three-month suspension running from February 9 to May 4, 2026. The most important integrity data never arrives as a complete table; it arrives in fragments, and readers fill the gaps with faith or suspicion. WHAT YOU HEAR WHEN THE STANDS ARE EMPTY An empty stadium does not just lack noise — it lacks a story being told. But it lets you hear something else. When the stands are empty, you hear the breathing of the match more clearly. I sat in near-empty arenas in 2026 and 2026, and what I remember is not the silence. It is the small sounds. Rubber soles gripping a hard court as a player changes direction. The single bounce before a serve. The exhale of someone who has just covered the whole court and has to get back into position. The chair umpire reading the score so clearly you can hear his breath. Normally the crowd drowns all of it. What was lost is equally clear: shared uncertainty. The real product of a sports match is not victory. It is thousands of people not knowing, at the same moment, what happens next — and holding their breath because of it. Without a crowd, uncertainty still exists, but it is imprisoned inside one person. Spectators are not consumers of the match; they are part of the match. I learned this from another sport. In 2026, in Nizhny Novgorod, I sat in a corner of the stand for Croatia against Argentina, choosing a seat that showed me Luka Modric's movement over a comfortable one for a reporter. The post-match data recorded that he ran more than ten kilometres. It did not record what I watched all evening: that every single step had a specific purpose. Modric does not run the fastest, but every step he runs has intent. Tennis works the same way. A serve at extreme speed is a data point. A slower second serve, placed into an opponent's backhand at 30-40 in the fifth set, is a decision. The system records the first, in great detail, and ignores the second. THE CONTRARIAN ANGLE: THE PROBLEM IS NOT A LACK OF DATA The industry's working belief is that bad decisions happen because we do not yet have enough data. I think the opposite is true: we have too much data, and it has been optimised to look complete. Three forces keep this alive. First, professional sports analytics departments hire people expected to produce reports; a report stating there is insufficient evidence is a career risk. Second, the more dimensions a framework has, the harder it is to challenge — a six-row risk matrix looks more credible than a blank page reading "nothing here," even when they say the same thing. Third, when data arrives in a beautiful interface, people judge it by form rather than content, an effect documented repeatedly in decision science. The deeper counterintuitive point is this: the empty-stadium era, which gave us the cleanest dataset in tennis history, taught us the least about the sport. It was clean because it removed the very variable that creates meaning. The decisive variable in elite tennis is not serve speed, not points-won percentage, not fitness. It is intent — what happens inside a player roughly half a second before the racket meets the ball. There is no camera there. That is why tennis will never be fully measured, however many sensors and ball-tracking systems we install. WHAT TO KEEP The value of tennis analysis over the next decade will not lie in adding more columns to the sheet, but in knowing exactly which rows should be left empty — and having the nerve to leave them empty. A report that says there is not enough evidence is sometimes the most honest document of an entire season. When a framework becomes beautiful enough that nobody questions it any more, it stops describing the match and starts describing itself. And when you read a tennis report in which every cell is filled, you should ask yourself: what did it leave blank in order to look this complete?