When the Data Table Is Empty: The Art of Silence in Tennis Analysis

At 2:47 a.m. in a small apartment in Queens, New York, I opened a data file sent from the statistics system of a Masters 1000 tournament, waiting for the first-serve numbers of a player I had tracked for six straight weeks. The file opened. The column headers were still there — First Serve In, First Serve Won, Return Points Won, Break Points Converted — but beneath them was white space. Not a single row. The quarterfinal was sixteen hours away. My editor texted asking for the piece. I sat there, hands on the keyboard, and a very comfortable escape route appeared in my head: just write from memory, from feeling, from what I thought I had seen on screen the night before. I didn't write. Not because I am noble. But because I have been on the other side of a mistake, and I know its cost. A wrong number in a tennis analysis kills no one. But it erodes the only thing a data recorder can own: the belief that when he says 'about 78 percent,' that number was actually counted. That night I learned something nineteen years in the trade had not fully taught me. When the data is empty, the analyst's real job is not to tell a better story, but to draw a line and stand on this side of it. In tennis, data is not decoration for a piece of writing. It is the spine. Every serve is logged by the Hawk-Eye system, every point classified by depth, direction, speed, and where the player stood at contact. The ATP and WTA publish hundreds of metrics each week, from second-serve points won to winning percentage in rallies lasting nine shots or more. An analyst sitting in New York can see how many centimetres off a Melbourne player's serve has drifted from last season. But that very abundance creates a trap. When there is too much data, people begin to believe there is always a number ready to answer every question. And when that number does not exist — when the file is empty, when the system fails, when the match sample is too small to say anything — the first reflex of many is to invent it. Not blatantly. Invent it subtly, filling the gap with a bias already present in the mind, then calling it 'expert intuition.' I have seen this repeat itself. A player wins five matches in a row, and immediately a piece appears explaining that his backhand has 'transformed.' The data? No one checks. The court conditions? No one asks. Who were the opponents? Three of those five matches were against players ranked outside the top 80. But the story has been told, and the story sounds better than the truth. The recorder's art begins here: distinguishing between story and evidence. A story is always available. Evidence must be sought, and sometimes we must accept that it does not exist. I remember the summer of 2026, when an English football club paid forty-two million euros for a winger from Serie A. The whole newsroom laughed at the deal. They said he was too slight, too slow, unable to handle the Premier League's physicality. I sat tearing apart spreadsheets on top speed, penalty-area entries, and shot conversion inside the box, and found that his metrics sat in the top five percent in Europe among wingers. I wrote that he would score more than thirty goals. He scored thirty-two. When the market mocked Salah, the data nodded quietly. But in the same piece, I also predicted that an attacking midfielder bought for forty-five million pounds would dominate his new club's midfield, and he faded all season. Same method, same care, one hit and one miss. What I overlooked was not in the data. It was in the context: a new role, a new tactical system, the position the manager wanted him to occupy. Since then, I never write a piece based on a single metric. Every analysis must include a section I call the 'role variable' — a detailed account of how the player is used before any quantitative conclusion. In tennis, that variable is even more complex. A player with a high return-points-won rate on hard courts can collapse on grass, where the low, quick bounce turns reaction time into a completely different contest. The same number, two opposite meanings. The summer of 2026 taught me a harder lesson. At a World Cup, after a semifinal, I used expected goals to prove that the winner created only 0.8 expected units while the loser created 2.1, then concluded the finalist 'did not deserve' its place. The community pushed back hard. They said football is not a computer simulation, that the spirit and stamina of a thirty-three-year-old midfielder was what carried the team through. I withdrew from commentary for a month and watched every penalty shootout of that tournament again. And I found something no expected-goals metric captured: the winning team's goalkeeper had a tendency to dive to his right roughly 2.3 times more often than to his left. I built my own penalty-save probability index, and it appeared in no official statistics table. Croatia was no accident. Some index had recorded the story before the ball rolled — it simply was not in the file I had open. Since then I struck the words 'deserved' and 'undeserved' from my vocabulary. They are lazy words. They assign a moral judgment to a sequence of probabilistic events. Instead, I write differently: 'This team won inside a sequence of events with a probability of roughly 18 percent, and this is the part my data still cannot explain.' Every piece I have written since ends with a small section called 'data limitations.' It lists what I do not know. It is honest about the gaps. And strangely, readers trust me more, not less. Because someone who admits his limits is someone who can be trusted in the places where he does not. Let me return to tennis, the sport I actually practise. Over the past three years, I have tracked a trend I believe matters more than any debate about who is the greatest of all time. It is the shift toward first-strike play. Top players increasingly avoid long rallies. They seek to end the point within the first three shots, and their winning rate depends mainly on the serve plus one — the first shot immediately after the serve. When I told a colleague this, he asked for the specific number. I gave him data from a tournament I follow closely. He nodded. Then he asked: 'But how many players does that hold true for?' I went quiet. It holds for the group of big servers. For the rest, those of modest height whose serve is not a weapon, the trend runs the opposite way. They are trying to extend rallies, not shorten them. This is the standing blind spot of data-driven tennis analysis: we often sample from the top group and apply it to the whole tour. The top group is a biased sample. They serve better, move better, and most importantly, they were selected by those very traits. Looking at them to learn how to play tennis is like looking at lottery winners to learn how to manage your finances. I always remind readers of this before giving any number. How a sample is drawn matters no less than the result. A correct metric can become misleading simply because it was measured on exactly the wrong group. Refereeing and video-assist systems are another field where data and people collide. In tennis, electronic line-calling has replaced the human eye at the lines. Technically, this is a clear improvement. Human error in fast-ball situations was a subject of debate for decades. But improving decision accuracy does not equal improving the spectator experience. When a point is overturned by the electronic system, the crowd sees a simulated ball trajectory projected onto the big screen. They see it. But they are not told why. They hear no reason beyond an icon on the screen. Meanwhile, in team sports, video officials routinely explain themselves to camera. This asymmetry creates a gap in perception, and that gap is filled with doubt. The problem with transparency in sport is that it is often just a slogan. Real transparency means that when a decision changes the fate of a match, the viewer must hear the reason. Not a technical line. A reason. Someone must stand up and say: 'We overturned the call for this reason.' An empty stadium does not make the result false; it only strips away our illusions — that human decisions are always fair, that technology is always neutral, that the audience is always respected as a party to the game rather than a forgotten observer. The transfer market is another field where data is used as decoration. There is a popular belief that the transfer fee is the measure of a deal's financial value. That is true in accounting but false in economics. The transfer fee passes through a monitored system, is recorded, amortised across the contract, and scrutinised under financial rules. Meanwhile, the signing bonus for a free agent — money flowing directly to the agent and the player — sits outside that system's view. It is more toxic than a large transfer fee precisely because it is more invisible. Every number in a contract is a confession by the market. If you read closely, you will see what no one wants to say: that some sums are designed never to appear in any public report. I do not have access to the full financial data. I know that. So I tell you this with a probability attached: based on what I observe, roughly 70 percent of major free-agent deals conceal expenditures that regulators cannot trace, and that makes them more dangerous than a publicly listed forty-million-euro transfer. I look at this market through the eyes of someone who has watched it for twenty-eight years. The market forgets nothing; it merely disguises itself as a new summer. Now let me return to the empty file from the start, because that is the real subject of this piece. When data is insufficient, there are three wrong reflexes an analyst can fall into. The first is invention. The second is over-hedging — turning every sentence into 'perhaps,' 'seemingly,' 'in a sense,' until the piece says nothing. The third is blaming the data — 'the numbers say this' — hiding behind it to dodge the responsibility of a judgment. All three reflexes come from the same fear: the fear of being wrong. I understand that fear. It is part of my neural architecture. For years it made me check a number three, four, five times and still not dare publish. I lost sleepless nights not because I did not know what I wanted to write, but because I knew too well where I could be wrong. But I learned that fear is not resolved; it is built. It becomes a defensive architecture. And that architecture has a single principle: set a sufficient threshold before you begin, and stop when you have reached it. My threshold is simple. Three independent sources, or two languages. If I cannot verify an event from three unrelated sources, or from two different languages, I do not assert it. I either record it with a probability, or I do not write about it. And I always write a sentence with a probability level. Not because I like sounding scientific. Because life is already probabilistic, and a judgment without a probability level is a judgment made without accountability. That night, with the empty file, I did something I consider more important than any analysis I have written. I opened another document and typed the first line: 'I do not have this player's serve data for the current tournament. I have data from last season, but the court conditions and opponents differ, so I cannot extrapolate directly. I will not predict this quarterfinal.' Then I wrote a short piece, about four hundred words, explaining why I could not predict. I listed what I knew, what I did not, and what I needed in order to make a judgment. I sent it to my editor. He replied ten minutes later with a single line: 'This is the best piece you've sent me this year.' I do not know whether he meant it or was just encouraging me. I assign roughly a 60 percent probability that he meant it. But either way, it taught me that readers do not need a fake answer. They need an honest person. In tennis there is a paradox I like to mention. The sport can be measured to the centimetre, yet the moments people remember most are the ones that cannot be measured. Forty-five shots in a single point in a final. A thirty-seven-year-old collapsing on the court after winning in four hours. A stadium falling silent before a decisive serve. Data can count the shots. It cannot count the fear. It can measure heart rate after a match, but not the heart rate in the instant before the ball leaves the hand. I know my limits. I do not try to measure what cannot be measured. I only try not to pretend that I have. Fans look with their eyes; I look with a probability distribution. But we are both looking at the same match. The difference is not in what we see, but in what we are willing to admit we have not seen. The truth lies deep beneath the table of numbers, where the headline never reaches. And sometimes that truth is a blank space. An empty data file. A silence the analyst must have the courage to account for, rather than fill. You may think this is a defensive view, even a timidity disguised as morality. I have heard that criticism many times. I assign it about a 35 percent chance of being right. Perhaps some part of me really is timid. But I believe there is a clear line between timidity and discipline, and that line lies here: the timid say nothing at all, while the disciplined say exactly what they know, with its level of confidence attached. In a short piece I once wrote: 'Don't ask for feelings; look at the numbers.' I still stand behind that. But I want to add a second clause I learned after many years: when the numbers are empty, ask me why. And if I answer 'I don't know,' trust me more, not less. Because an analyst who can say 'I don't know' is an analyst you can trust when he says 'I know.' There is a great temptation in this era: the temptation to generate data. Modern tools can produce tables that look real, charts that look professional, prose so fluent it is hard to distinguish from a human's work. And that raises a question I think is the most important of this profession in the coming decade: when data can be generated, what separates a real analyst from a machine? My answer is: the willingness to say 'insufficient grounds.' A machine is designed always to answer. A disciplined human is designed to know when to stay silent. And in a world flooded with information, well-timed silence becomes an asset more valuable than information itself. I do not write about tennis; I merely transcribe scripture from data. But to transcribe that scripture honestly, I must accept that sometimes the blank page is the truest answer. Before every major tournament, I set a threshold. If the data file is empty, I write no prediction. If the file has data but lacks court context and fitness, I write with a warning. Only if the file is complete and consistent do I allow myself a judgment with a probability level. That is the signal I want you to watch for the next round: do not only read an analyst's conclusion. Read their data-limitations section. If they do not have one, ask yourself why. Because someone who never admits a limit is someone who has never truly examined himself. And you — when you read a tennis analysis at two in the morning, and someone tells you a number that sounds very confident — will you ask yourself whether that number was actually counted, or merely filled into a blank space? That question, to me, matters more than all the statistics tables combined.