When the Spreadsheet Is Empty: A Lesson in Honesty in Tennis Analytics

On the second monitor, my familiar spreadsheet sits open with twenty-four columns: first-serve percentage, points won on first serve, points won on second serve, return points won, break-point conversion, break points saved, winners, unforced errors, tie-break efficiency, decisive-point win rate, and a string of derived metrics. That Tuesday morning in Brisbane, all twenty-four columns were empty. I had a match on the calendar, a player profile open on the screen, an assignment from the newsroom, and not a single line of reliable data with which to write a conclusion.

That moment reminded me of the oldest lesson in the trade. Most of an analyst's time is not spent finding answers, but recognizing when there isn't enough material to answer at all. In tennis, where a single point can swing a set and a single set can swing a career, the line between analysis and fabrication is thin enough that one overly confident sentence is enough to cross it. Data does not lie; it is the people reading it who make excuses.

An empty profile is not a bad profile

I grew up in this work with one non-negotiable rule: every tactical claim must be paired with at least two quantitative indicators, and every conclusion must be cross-checked between what the eye sees on court and what the numbers whisper from the spreadsheet. That rule was forged during the 2026-18 Premier League season, when I was a sixteen-year-old writing for a Manchester City fan site. That day I used pressing data to argue that Pep Guardiola's side was far from fragile: an expected-goals figure of 1.8 against the opponent's 0.4 in a game of near-total control. The piece ran two thousand words, but what kept it alive over time was not the prose — it was the Excel sheet tracking every one of the twenty teams each matchweek, a habit I kept for months afterward.

That habit followed me into tennis. Before I took on the empty-data assignment described above, I had spent nearly a year building quantitative profiles for every player inside the ATP and WTA top 50 that I regularly cover: not merely ranking tables, but week-by-week protected-point structures, win rates by surface, tie-break efficiency, and the gap between ranking by points and ranking by actual form. An empty profile is not a poor one — it is simply a profile that cannot yet answer the question the newsroom has asked.

In the market where I work — Australia, alongside a Vietnamese tennis community watching from afar — the gap between enthusiasm and evidence is especially visible. The Australian Open unfolds in front of me every January, and I see fans in Melbourne and Saigon react the same way to a player who suddenly shines: they want to believe they are witnessing the birth of a legend. That emotion is legitimate, but it usually runs months ahead of the data.

Tennis scores point by point

What makes tennis different from football — and what makes it most misunderstood — lies in its basic unit of measurement. Football measures expected goals across a frame; tennis measures binary points, won or lost, on every rally. A player can win 55% of total points in a set and still lose it, because a set is decided only at specific moments. Confidence intervals are not a technical detail to decorate an article — they are the entire story.

I often remind readers of a figure anyone following professional tennis should memorize: the probability of winning a point on a successful first serve typically sits around 70-75% at the men's elite level, while on a second serve it drops to roughly 50-55%. That twenty-percentage-point gap is where every tennis match is decided. A player who lands 60% of first serves and wins 74% of them, while landing 40% of second serves and winning only 52%, can win one match easily or collapse in another depending entirely on which set his first-serve percentage lands in.

In 2026, I learned that a 95% probability still has a 5% that laughs. Ahead of the World Cup in Russia, I built a prediction model from six major tournaments of historical data, using Elo ratings and qualifying records. The model ranked Brazil as the number-one contender at 23.4% championship probability and pushed France down to fourth at 11.2%. Brazil went out in the quarterfinals; France lifted the trophy. I realized my model was missing two variables Elo cannot capture: squad depth and the mental state of key players. Within a month of the tournament I tore the model down and rebuilt it from scratch, adding club minutes played before the tournament for every player.

That lesson applies directly to tennis. Here, the forgotten variable is usually the schedule. A player entering the third round after two three-set matches, while his opponent has just won two straight matches in tidy two-set wins, represents two entirely different physical states even if their ranking tables look identical on paper. The tie-break win rate, in turn, reflects nerve at the exact moment when psychology decides more than technique. None of the twenty-four columns measures fatigue directly, but the number of sets, games, and minutes on court over the past seven days can approximate it.

The hot-streak trap

A six-match winning streak over three weeks reads loudly in a headline. But when I break that streak down match by match, the story usually changes tone. Three of those six wins might have come against opponents outside the top 50, two others after an opponent retired injured mid-match, and only one was a genuine test of class. The small sample is tennis's most common trap, because unlike football — where a season has thirty-eight rounds — an individual tennis season may compress into just a few dozen matches per player.

I have sat long enough with a young player's data to know that a hot streak can come from three entirely different sources: genuine technical upgrade, luck in tie-breaks, or a temporary dip in the opponents' standard. Of those three, only the first is sustainable. To tell them apart, I track indicators less volatile than results: return points won, break points saved, and the winner-to-unforced-error ratio. If a winning streak comes with a steadily rising return-points-won rate, it signals a player who has found structure. If the streak comes only with a surging first-serve percentage, I keep my skepticism.

A major-tournament season has its own logic. When the calendar compresses around a few Grand Slams and a few Masters 1000 events, player psychology is not the same as on ordinary weeks. Fans ride flags and stories, and that is entirely right for elite sport. The analyst's job is not to dampen that emotion, but to keep the analysis anchored to what happens on court rather than to the gaps in the story being told.

Protected points and the ranking cliff

A complete tennis profile cannot ignore the protected-points structure. In the professional ranking system, each player carries a block of points earned over the previous fifty-two weeks, and when the week of an old tournament returns, that block must be defended or it evaporates. Some players hold very high positions thanks to a few explosive tournaments, yet face a cliff over the next three months when nearly half their points come due at once.

Reading that structure lets me make calls that the visible results never reveal. A player who just won a Masters 1000 may still be sitting on a block about to collapse; conversely, a player who just lost early may be entering the lightest stretch of the season. This is information the eye cannot see but the spreadsheet sees clearly — and it is also the kind of analysis most prone to fallacy if the writer confuses a data fact with a forecast about the future.

Looking at recent history, the generational handover shows this best. The era of Novak Djokovic, Rafael Nadal, and Roger Federer lasted so long that prediction models habitually assigned every major title to that group, until Carlos Alcaraz and Jannik Sinner forced those models to be rewritten. Rafael Nadal holds the record of fourteen French Open titles, a figure almost impossible to repeat on clay, while Novak Djokovic holds the record of twenty-four Grand Slam titles. Those milestones are not data for predicting the future — they are data for understanding that long eras can end faster than people expect.

Which court, which surface, which answer

Surface is the most underrated variable in tennis conversation. A player excellent on hard courts in Melbourne or New York can become thoroughly ordinary on the clay of Paris, where the ball bounces slower and points run longer, or on the grass of Wimbledon, where the ball bounces low and speed rewards big servers. The same technical profile, three surfaces, three different answers.

I split each player's data by surface, and when a season moves from hard to clay to grass, I treat that as three separate tests rather than one continuous run. This sounds obvious, but the consequence is not: a general season win rate can mask two entirely opposite profiles inside it. A player winning 70% on hard courts but only 50% on clay is not a steady 60% player — he is two different players depending on where the ball bounces. Technical adaptation between surfaces, with adjustments to serving position and decision-making tempo, often takes weeks to months, and that is the window in which an analyst must be especially cautious about any rising form streak.

This is also why a player's results at Roland Garros are hard to infer from results at the Australian Open. Point structures on clay are longer, rallies are more extended, and big servers often lose part of their weapon. Iga Swiatek's clay record is a textbook case of a technical profile bound to a surface: her spin rate and lateral movement peak in high-bouncing, slow conditions, and any analysis of her must start from that fact.

When reputation outruns the data

In any individual sport, reputation and data rarely move at the same speed. A player who has won multiple Grand Slams may be priced by media and sponsorship markets on what he has done, while current form indicators tell a less glittering story. The gap between reputation and data is where misjudgments are born.

Transfers are where people pay hundreds of millions to buy one row in a spreadsheet. In tennis, the equivalent lives in sponsorship deals and wildcards. A tournament awarding a wildcard, or a brand signing a young player on a short hot streak, is betting on that streak continuing — in many cases, betting on a data row with too small a sample to trust. That is not wrong. It should simply be called by its right name: a bet on the future, not a conclusion about the present.

A player shining at a major does not prove that every one of his indicators has reached championship level. It proves only that, over two specific weeks, on one specific surface, in one specific draw, he won seven matches. Those seven matches could be the start of an era, or just a fine summer in which the opponents' standard made it easier. Data cannot distinguish the two possibilities immediately. Only time can.

Contrarian is not always true

Going against the crowd gave me my most memorable pieces. In 2026, during the Euro held across Europe, I wrote a rebuttal to the wave of criticism aimed at Denmark's coach after a losing opener against Finland. I used pressing data and chance-creation figures to show the team was far from poor — it had simply lacked luck. My editor cut the piece for going against the common feeling. A week later, Denmark reached the semifinals, and my article became the most-read piece of the month. The first data rebellion was never meant to overthrow anyone — only to prove the numbers deserved to be heard.

But precisely because I once won that way, I must guard against myself most of all. An analyst can easily turn contrarianism into a reflex rather than a conclusion. Whenever a counter-view surfaces, the first thing I ask myself is not whether it is right or wrong, but whether the phenomenon repeats. Correlation is not causation, and an outlier is not a model. I only call something a signal when it reappears across multiple matches, surfaces, or tournaments.

Likewise, I force myself to say out loud what I cannot measure. Existing data cannot answer questions about a player's state of mind after a painful loss, the effect of a bad practice session, or how an incompletely healed injury will show up in the scoreline. Shoulder pain, confidence in a forehand, fear before a decisive serve — none of the twenty-four columns measures those things. For a data addict, admitting that stings as much as admitting more sample is needed. But that sting is precisely the line between an analyst and a fabricator.

The temptation of an empty spreadsheet

Back to that Tuesday morning in Brisbane. I had everything but data. The temptation was real: with one click I could have stitched together a flowery passage about a player's character, built a story about fighting spirit, and filed on time. Nobody would notice immediately. The piece could spread, be shared, and gather tens of thousands of reads. But such a piece would not survive a second reading — and it is on the second reading that credibility is built.

What I increasingly believe is that the best tennis writing is sometimes the writing that refuses to conclude. When data is insufficient, the honest answer is to say it cannot yet be assessed, rather than stuffing readers with a firm opinion so they feel satisfied. Tennis content has been flooded in recent years with absolute claims: who is certain to win, who is certainly finished, who is certainly the greatest. Most of those claims are written in the voice of someone who has seen the truth to its end. But nobody sees the truth of a sport in which every point depends on a serve landing in or out.

I do not want to become the joy-killer — the label my readers occasionally pin on me when I say a single result does not prove a long-term conclusion. I do this work because I love tennis, and precisely because I love it, I do not want to tell its story falsely. How fascinating is a sport when you know that behind every forehand of a top-10 player lie hundreds of hours of video analysis, thousands of practice sessions, and a point-selection system calculated to the second? That truth is far more interesting than a fabricated prophecy.

For the tennis community in Vietnam, where Ly Hoang Nam and Nguyen Thuy Linh are among the few names to appear on professional rankings, data discipline matters even more. When the level gap remains wide, writers easily fall into two extremes: overpraising a small result, or concluding too soon that the road ahead is closed. Both are conclusions reached before the data, and neither helps readers who genuinely want to understand what is happening.

Signals to watch

There is a paradox I carry from the empty stadiums of 2026: with no crowd noise, you hear the match breathe more clearly. I spent weeks comparing hundreds of matches before and after football returned without fans, and found that losing the stands changed how teams played — pressing dropped, sides grew more cautious, set pieces were taken differently. The no-crowd season was the cleanest laboratory football has ever had. Tennis underwent a similar test during its no-spectator period, and I hold to one conclusion from then: when the competitive context changes, historical data loses part of its explanatory power.

That is why, in the current major-tournament season, I will track specific signals rather than guess at results. I will watch the break-points-saved rate of top-10 players in three-set matches; if it rises late in a tournament, it signals that fitness is overriding technique. I will also look at tie-break efficiency for players who have played at least two straight three-set matches, where mental fatigue shows most clearly. And I will pay special attention to the gap between a player's first-serve and second-serve points-won rates, because a suddenly narrowed gap often warns of a weakening second serve for some reason the stat sheet does not mention.

Those signals will not tell me who wins. They only tell me where to look during a match, and what to doubt when a story looks too good. In a sport where a single point can swing an entire week, knowing what you don't know is sometimes the biggest advantage.

I will still open two monitors every morning, still keep the habit of updating the spreadsheet after every round, and still publish the limitations of my model at the end of each piece. The major-tournament season is near, emotions will rise, and I understand why fans want a firm promise about who will lift the trophy. I have no such promise to give. What I have is a set of indicators, a tracking system, and an uncomfortable honesty with myself: if the spreadsheet is empty, I will say it is empty, rather than fill it with my imagination. When the season begins, I will sit down again beside those two monitors, update every column, and wait for what the data wants to say — even when what it wants to say is silence.