When Data Goes Silent: The Line Between Analysis and Fabrication in Sport

On a Tuesday morning, in my data room in Sydney, a spreadsheet opened empty. Thirteen columns, not a single number. The data feed from a Challenger-level tennis tournament had cut out at three in the morning, and I had exactly fifteen minutes to decide: wait for the feed, or write with what I had — which was nothing.

That moment was not pretty, but it was honest. And in sports data analysis, honesty is the only asset money cannot buy.

In those fifteen minutes, two paths appeared. One was to tell the truth: "I don't have the data yet." The other was to fill the gap with guesswork, with instinct, with whatever sounded plausible. The second path is always easier. It is also the path that leads straight to a collapsed reputation.

Sports analytics is at the peak of a crisis of trust. There has never been so much data: xG, PPDA, serve speed, points won under high pressure. But there has also never been so much analysis written without a single real data source behind it.

In Australia, where I work for sports broadcasters, this pressure is concrete. Every match ends, and the newsroom needs copy immediately. A tennis match lasts four hours, and you have forty minutes to turn it into a meaningful story. If the data has not arrived, people write from feeling. And feeling, in sport, is a wonderful storyteller that lies very often.

I learned this the hardest way, in 2026.

That year, after the success of my dataset on Aaron Mooy — a player I tracked across 380 matches, finding he covered 12.7 kilometres per match with 87% of passes completed under high pressure — I was confident enough to publish a prediction model for the World Cup. Brazil to win, with a 78% probability. I relied on xG, PPDA and squad rotation, metrics I believed were enough. Croatia reached the final, and my model burned.

The point is not that I was wrong. The point is my first reaction. For a moment, I wanted to write a piece explaining why the model was still right, why Croatia was merely a "statistical outlier." I wanted to fill the gap with a story that sounded reasonable. I nearly did it.

Instead, I did the opposite. I wrote a series of self-criticism pieces, re-analysing Croatia's six matches, and uncovered a metric no one had measured: pressing-transition capacity. Silent data does not mean the truth does not exist; it only means you have not yet found the right place to listen.

That is the lesson every analyst must carve into memory. Numbers never lie, but they can be silent. When they are silent, the best analyst is the one who dares to say "I don't know yet," not the fastest one to invent a fake number to fill the gap.

In tennis, this is even clearer. Players like Novak Djokovic or Carlos Alcaraz are wrapped in grand narratives: iron mentality, nerve in the tie-break, the ability to come back from the brink. But behind every such story there must be a number. First-serve points won in the decisive game. Break-point conversion. Point rhythm when the score is level. If those numbers are not there, the story is just emotion — and emotion cannot predict the next match.

There is a paradox our industry has not dared to name. We reward confidence, not honesty. A piece that dares to say "the data is not enough to conclude" is treated as weak. A piece that offers a decisive number, even one built from thin air, gets shared.

That is why so many models look beautiful on paper but are useless on court. They are built to look convincing, not to be correct. When reality betrays them — as Croatia betrayed me — people do not fix the model, they fix the story.

The irony is that the data gap itself is the most valuable information. When a source goes silent, it can be a signal of a systemic problem, an abnormal match, or something the model was never designed to see. Ignoring the silence is ignoring the "hidden number" — the thing every good analyst hunts for. Every shot leaves a footprint. The best are not those who run the most, but those who leave footprints in the right places.

I once burned my model with Croatia. That was the day I learned to listen to data — even when it says nothing. In an increasingly noisy industry, the ability to stay silent when there is no evidence may be the most valuable skill an analyst can own.

The next match will begin again. And the question is not who wins, but this: when the data speaks, will we be brave enough to hear what it actually says — instead of what we want to hear?