The Blank Tennis Data Table and the Discipline of Sourcing Every Number

At six in the morning in Sydney, I opened the match dashboard for an ATP 250 tennis tournament and found an entirely blank data frame. First-serve percentage: empty. Return points won: empty. Break-point conversion: empty. Winner-to-unforced-error ratio: empty. Four cells, four silences, and a cold footnote: insufficient information to assess.

In eighteen years of watching and analysing sports data, I have learned that a blank table is never neutral. It is a vacuum, and in a newsroom a vacuum is always filled by the fastest thing available: a story. Someone will rewatch the match, remember one beautiful point, and write that the player won on nerve. Nobody will check whether the number existed, because the number was never created.

Data whispers. Those willing to listen will hear an entire match. But only when the data is actually there.

The data pipeline and the extraction layer

To understand why that gap is dangerous, you have to see how tennis data is made. A match on a hard court in Melbourne or a clay court in Paris does not turn itself into a table. It travels through a two-layer pipeline.

The first layer extracts events: who played, when, at which tournament, the result of every point. Only the second layer analyses: serve trends, pressure in decisive games, the physical drop-off in the third set. The first layer's sources may be the Hawk-Eye system around the court, StatsBomb's shot-tracking data, or official statistics from ATP Media. Each source has its own definition of an unforced error, and that difference alone is enough to change an entire conclusion.

When the first layer fails — a dead link, an article behind a paywall, an image-only report — the second layer does not know what it is analysing. But it runs anyway. It still produces a report full of structure, nine sections, charts. The only difference is that every cell reads “insufficient information”. That is the worst thing that can happen to an analyst: a perfect framework with nothing inside.

Before you trust a number, ask where it was born. I write that line in every piece. It is even truer when the number does not exist.

Based on my experience watching matches, I have found that most data failures do not come from the algorithm. They come from the input stage. A statistics table missing one set, a match tagged with the wrong date, a player assigned the wrong name — any of these is enough for a model to return a wrong conclusion without ever raising an error. Machines do not know how to doubt. People do.

What stands out is that input errors are rarely loud. They do not crash the system. They simply make it answer wrongly, smoothly — which is exactly why they are more dangerous than an obvious fault.

When I nearly wrote about a gap

In 2026, at twenty-five, I wrote a 3,200-word analysis of Melbourne City's pressing metrics for a newly founded Australian football site. Using GPS position data, I showed that midfielder Luke Brattan ran 11.2 kilometres per match but produced only 1.3 successful tackles, meaning the team was pressing in the wrong direction. Fans mocked the piece as too dry. Three weeks later, the coach changed the pressing shape, and Melbourne City won four straight matches.

The lesson was not that I was right. The lesson was that if the GPS vests had failed that day and returned blank data, I would have had nothing to write. And worse — I might have written something that sounded perfectly reasonable.

In 2026, I predicted Croatia would reach the World Cup semi-finals based on xG. Luka Modrić created 2.4 xG per match in the group stage. A group of amateur coaches on Reddit called me a bookworm who did not understand football. Croatia reached the final. After the tournament, a reporter from The Athletic contacted me to ask how I calculated defenders' “defensive xG prevented”. I spent two weeks writing Python, cross-checking against StatsBomb data, and sent back a seventeen-page analysis. In 2026 they laughed at my xG. This year they ask me what xG is.

Then came 2026. When the Bundesliga returned to empty stadiums, I was running a match-result model. My model priced home advantage at 0.45 goals per match. After nine rounds without crowds, the figure fell to 0.08. I turned down a magazine's request to write about crowdless football, because I needed three more weeks of data before I could be sure. When I published, I admitted I had been wrong not to account for the crowd variable.

All three times, I met the same temptation: to fill the gap with something that sounded plausible. And all three times, what saved me was not talent but a simple rule: if there is no number, do not pretend there is one.

The pressure to invent and the correlation trap

This is the part few people discuss. In a newsroom on deadline, a blank data table is not allowed to exist. The editor needs a piece. The reader needs a story. And the analyst — the only person who knows the number is not real — is the easiest person to pressure.

The temptation is called the pressure to invent: filling the empty cell with a plausible value, a familiar average, a percentage someone once saw in another match. What makes it dangerous is that it does not look like fabrication. It looks like inference.

The line between the two is as thin as an offside line. A variable that is valid in one match can be meaningless in another. Misreading a single variable is like losing your bearings for an entire year. Correlation is not causation, and a sample of two matches is not a trend.

I have seen analyses so beautiful they were hard to believe — nine sections, charts, conclusions. Look closely, and every cell is empty. The writer simply filled it with intuition and draped a layer of numerical language over the top.

In tennis, the trap runs deeper. A player who wins three matches in a row can look like they are peaking. But if all three opponents sit outside the top 100, that streak says nothing about title chances. The gap between market expectation and what happens on court usually only appears when we are willing to read the right number — not the number we want.

The discipline of the empty cell

Sports data analytics is learning a lesson medicine and aviation learned long ago: an empty cell must be recorded as empty, and must not be guessed. In statistics we call it a missing value. In a newsroom, we should call it honesty.

Three principles govern every tennis piece I write. First, every number must carry a source: Hawk-Eye, StatsBomb, ATP Media, or my own notes. Second, every conclusion must carry a section on “assumptions that may be wrong”. Third, if the data is blank, I write about the blankness itself — instead of filling it.

I also added a technical safeguard: an automated filter that flags any data table with zero data points. It sounds simple, but it blocks the most dangerous thing of all — an empty result quietly flowing into an article as though it were fact.

Transfer value is a story, but data is the signature. A piece with no signature is just a rumour presented nicely.

A signal for the next round

The match on that dashboard did eventually happen. The data reloaded forty minutes later, and the empty cells filled with real numbers: first-serve percentage, return points won, break points. Looking back, I realised the frightening thing was not the initial gap. The frightening thing was the brief window in which I was ready to write about it as though it had never existed.

A season missing detail is like a match missing stoppage time. Both end somewhere nobody planned.

So when you read a tennis analysis full of beautiful numbers, ask yourself one question: where was that number born, and if it did not exist, what would the writer have written?