The Data Gap on the Tennis Court: When Silence Is a Professional Decision

On Tuesday night I opened the folder holding my analysis file and found every field empty. No tournament name. No player. Not a single figure. All that remained was the "tennis" label and a pre-built template frame waiting for data to pour in — but the data never arrived.

In the trade of writing about officiating, that is the kind of night that tempts you to type a few lines and be done with it. A headline, a couple of plausible-sounding judgments, a little "perhaps" and "it seems", and the piece is ready for the page. I once came close to that. In 2026, as a second-year student, I attributed a yellow card to the wrong player in my report on a university derby. My editor reprimanded me, I had to write a letter of apology, and I spent the following six weeks cataloguing hundreds of card incidents as reference data. The first mistake was not the card I misassigned. It was believing I would never misassign one.

So that Tuesday night, I closed the folder and did not write. It was the most professionally sound decision I had made in months.

Tennis handed judgment to machines long ago, but the boundary between the measuring device and the person reading it has never been clear. In 2026, the US Open quarter-final between Serena Williams and Jennifer Capriati produced a run of erroneous line calls so egregious that the crowd reacted furiously. That case accelerated the adoption of Hawk-Eye in professional tennis, initially at a handful of major events, then across the whole system.

Two decades later the technology has travelled much further. The Grand Slams have replaced line judges one by one with electronic line-calling systems built on the same sensor principle. Yet the more we hand to the machine, the more we forget one thing: every sensor system carries a margin of error. Hawk-Eye has been published as carrying an error of roughly 3.6 millimetres per line calibration, smaller than the diameter of a racket string. Roger Federer publicly criticised the system for not being as flawless as it claimed. He was technically right, even if the way he said it irritated plenty of people.

The biggest controversies in tennis over the past two decades have rarely concerned a sensor being wrong. They have concerned the inconsistent application of sensors: used at one event, not at another; available in one round, dropped in the next; reviewed on one point, ignored on an identical one. Consistency of application is what produces fairness, not the resolution of the device.

The problem in my trade is not how accurate Hawk-Eye is. It is that when the data vanishes entirely, a great many writers keep writing anyway.

In daily work I follow a three-layer ritual. Layer one: verify whether the figure actually exists — which sensor produced it, when it was calibrated, who entered it into the system. Layer two: set it against historical context, measuring how far it deviates from the average for that surface, that phase of the season, that specific player. Layer three: check whether at least two independent sources confirm it.

On Tuesday night, all three layers returned the same result: there was nothing to check. The empty analysis did not come from laziness. It was empty because the input-extraction stage had failed before I even began. All that remained were field labels: player, tournament, timestamp, metric. A skeleton without flesh.

For a writer with weak discipline, that is an opportunity. One can fill the skeleton with general tennis knowledge — the dominance cycle of the new generation, the players on the rise — and call it analysis. The piece will read beautifully. It lacks exactly one thing: the fact that it rests on no source article at all.

That is the greatest risk in the entire sports-content pipeline today. AI does not err on its own. The operator who lets AI write in its place once the data has run dry is the root cause. A tool does not generate its own errors. The operator does. And the gap between those two entities is precisely where my work begins.

A mis-recorded first-serve points-won figure, if it slips into a season-end report, will shape how a player is evaluated for an entire year. It does not sit on the scoreboard. It sits in the reader's head. And once it is in the reader's head, correcting the actual number becomes futile work.

The most counter-intuitive thing in this trade is that "low risk" is not a safe verdict. When the analysis is blank, writing "no issues found" is not neutrality — it is a conclusion that errs toward a false negative. An absence of evidence of fault does not amount to confirmation of cleanliness. In a data-handling room those two things are worlds apart, and only someone who has paid the price can tell them apart.

By the same logic, I never blame the system. Hawk-Eye is not wrong. The Hawk-Eye operator is wrong. A text-extraction system is not wrong. The person who configured it is wrong. When everything runs correctly, nobody mentions the operator. When something fails, people immediately blame "the technology". That is the easiest way to dodge responsibility, and also the way that makes errors repeat.

My not writing on Tuesday night was a control gate. A serious tennis writer needs one hard stop: no headline, no facts, no source — no publication. No exceptions. Because a wrong number repeated three times becomes truth in the season-end report — and nobody traces back to find where it began. I record every card, every minute of stoppage time, for exactly that reason.

I still keep that empty folder on my machine as a reminder: in a sport increasingly dependent on sensors and algorithms, the ability to stay silent at the right moment may be the most important professional skill of all. When data contradicts the eye, trust the data — but never skip checking its provenance. And when the data does not exist, do not invent it. The question the whole industry needs to ask itself: who will be the one brave enough to close the folder?