Tennis and the Trap of an Empty Data Sheet

In Los Angeles that night, I opened a nine-section analysis of a Grand Slam quarter-final. Every cell was empty. Not a single serve statistic, not a single return-points-won rate, not even the winner's name. Nine pages, and the only thing repeated was one identical line on every row: insufficient information. I sat looking at it for a long while. An analysis engine, carefully designed, running exactly to process, with not one piece of data to speak of. For someone who has read tennis for twenty-five years, from screens and from the stands, that scene stays with me more than any serve.

Tennis entered the data era very early. Hawk-Eye appeared at Wimbledon in 2026 to judge whether a ball was in or out, and quickly became the standard. IBM brought SlamTracker to the Grand Slams from the late 2000s, turning every rally into a searchable data point. Motion-tracking systems measure distance covered, rotation speed, even impact force on the racquet head. The era of Roger Federer, Rafael Nadal and Novak Djokovic overlapped with the period when data became an indispensable part of the sport. Sam Groth once served a ball at 263 km/h at a Challenger in Busan in 2026, and the machines recorded it to the tenth. Technically, tennis has never been measured this closely.

But that is the story of data. The story of people is different.

In more than twenty-five years of watching matches, I have learned that tennis data comes in two kinds. The first recounts what happened: first-serve percentage, points won at net, break-point conversion. The second is what analytics rooms love — built to predict what comes next: pressure indices, form curves, championship probabilities. It is this second kind that breeds the “darlings” of the analytics room. A young player whose numbers spike, scored high by the model, hailed as the heir. Then the tournament begins, and the analytics room's darling must eventually stand on his own two feet.

The problem is that data does not speak for itself. A high return-points-won rate can come from brilliant returning, or from an opponent serving badly for a week. A falling pressure index can reflect fading fitness, or simply a player deliberately slowing down to save energy for a tie-break. If the person reading the numbers is not at the court, does not see how the player stands, does not hear the gasps after a long rally, then those statistics are orphaned. A quiet summer turns records into orphan numbers — I once wrote that line in a year when tournaments had no crowds, and it holds for the prettiest data tables too.

What is worth noting is that tennis analysts are getting better at measuring, but not always at interpreting. We can measure the spin rate of a forehand, but not the nerve to hit a second serve at break point. We can count double faults, but not the fear of double-faulting. A player can win eighty percent of points in practice and only forty percent when the stands are full. That difference lives in no data cell. Data does not speak for itself; the value of analysis lies in the real information base behind it, not in the number of cells filled in. A spreadsheet does not know what desire is, and we should not pretend otherwise.

The crowd often believes more data means more understanding. This profession has taught me the opposite: more data often produces the illusion of understanding. When a model gives a player a seventy percent probability, people forget it rests on assumptions, and assumptions can be wrong. I once predicted Croatia to beat Russia 5-4 in the 2026 World Cup quarter-final shootout, based on Russia practising penalties forty-five minutes a day. The result was 4-3, and I realised I had chosen a safe prediction out of fear of being wrong. The Russian night was scorching, and the only lesson that remained was the silence — the silence of a man who let data cover his instinct.

In tennis, that lesson repeats every tournament. A player rated favourite by the model on a hard court runs into an opponent whose style fully counters him — something no statistics table captures. A seed forecast for a deep run loses in the first round to an unannounced wrist injury. The data is not wrong. It is just not enough. The biggest trap is turning that 'not enough' into a complete conclusion.

If you read a tennis analysis and find every cell empty, do not rush to call it a failure. Sometimes it is more honest than any numbers stuffed in to look good. What I want to see next season is not more complex models, but analysts willing to say 'I don't know' when they truly do not know. Numbers are only seasoning. People are the main dish.