When Tennis Coverage Has No Data: Lessons From the Jannik Sinner Case

It is 6 a.m. in Brisbane. I open my laptop and find a 1,200-word tennis story about one of the world's leading players. I read it to the end and notice something: across the entire piece, there is not a single concrete number. Not one first-serve percentage, not one timestamp, not one named source. Only feeling, only inference, and a headline heavy enough to earn the click.

I know that feeling. Years ago, as a schoolboy writing analytical blog posts for a Manchester City fan site, I confidently published a long piece before the 2026 World Cup declaring that "the data has identified the champion." My model ranked Brazil as the number-one contender with a 23.4 percent title probability. France sat fourth at 11.2 percent. France won. In 2026 I learned that a 95 percent probability still has a 5 percent laugh. And I learned something more uncomfortable: a piece written without verifiable data is not analysis. It is belief, packaged and labelled as analysis.

Tennis runs on an information ecosystem with a different structure from football, but the same disease. In football I track PPDA, xG, touches in the box, counter-attacks that produce a shot. In tennis the metric set is narrower but far more transparent: first-serve percentage, points won on first serve, points won on second serve, return points won, break-point conversion, and the metric I value most — points won at the decisive moments inside tiebreaks and deciding sets.

The ATP and the WTA publish this data immediately after each match. The ITF and Grand Slam organisers publish their own sets to their own standards. Independent trackers such as Jeff Sackmann's Tennis Abstract archive tens of thousands of matches, queryable by player, by surface, by date range. In other words, the raw material for a decent tennis analysis is available, free, within hours of the final ball.

So why do so many stories still appear without a single number?

The answer is speed. The Australian market I serve has a brutal time-zone reality: a match finishing in Melbourne at 11 p.m. local is 8 p.m. in Europe and midday in the United States. The window to publish before readers move on is a few hours, sometimes less. Inside that window, a writer has two choices: wait for official data, or run ahead on instinct. Many choose the second. Nobody is ever punished for choosing the second.

I understand that pressure firsthand. In 2026, working remotely for an Australian sports outlet during the European Championship, I wrote a rebuttal to veteran reporters who had attacked Denmark coach Kasper Hjulmand after a 0-1 opening loss to Finland. I used pressing data and expected goals to show Denmark had generated the highest group-stage xG total, 3.6 across three matches, behind only France and Spain. The chief editor spiked the piece for going "against the general feeling." A week later Denmark reached the semi-finals. The article eventually ran and became the most-read piece of the month with 45,000 views.

Data does not lie; it is the person reading the data who makes excuses. But when there is no data at all, the writer has nothing to read — and that is where danger begins. Because a void has gravity. A void always gets filled, and it gets filled with the cheapest available material: guesswork.

The Jannik Sinner case is the cleanest example I have tracked of how information distorts as it passes through layer after layer of intermediaries.

In March 2026, a sample taken from Sinner at Indian Wells returned a positive result for clostebol, a substance prohibited by the World Anti-Doping Agency. That is the original fact. A second sample soon returned the same result. Sinner received a provisional suspension, appealed, and was allowed to keep playing while the process ran.

Let me stop here, because this is where most coverage went wrong. The original fact contains only four propositions: a positive sample; a named substance; a provisional suspension procedure; an appeal. Four propositions, all verifiable, all traceable to an official document. But as the news spread, dozens of further propositions were added — about intent, about morality, about how "he must have known," about "some cover-up somewhere." None of those had data behind them. They were neither entirely false nor entirely true. They sat outside the zone of verification, and that is precisely where vigilance is required.

In August 2026, an independent tribunal accepted Sinner's explanation: the clostebol entered via cross-contamination from physiotherapist Giacomo Naldi, who applied the spray Trofodermin to a cut on his own finger before massaging the player. The tribunal found "no fault or negligence" on Sinner's part. The consequences were financial and ranking-based: he forfeited the prize money and ranking points earned at Indian Wells, roughly 325,000 US dollars and 400 points.

In September 2026, WADA appealed to the Court of Arbitration for Sport. In February 2026, WADA and Sinner reached a settlement: a three-month suspension running from 9 February to 4 May 2026. Sinner missed several events and returned in Rome.

Now count the facts in those paragraphs. There are specific dates. There is a named substance. There are named people. There are named governing bodies. There is a sum of money and a number of points. Every element traces back to an official document from the ITIA, WADA or CAS. Not one detail requires feeling, "apparently," or "reportedly."

And this is the part that made me sit back longest. Across nearly a year, I read hundreds of articles about Sinner. Most did not contain even a third of the facts I just listed. They contained emotion. They contained "apparently," "seemingly," "multiple sources suggest," "the public is asking questions." Those articles were empty data packages with a heavy headline attached.

I once built a small spreadsheet to test this. I took 50 international articles on the Sinner case published between March and August 2026 and counted how many verifiable propositions each contained — meaning a proposition that could be checked directly against an official document or a public data record. The average was 2.4 such propositions per article. Strip out material quoted from the ITIA statement itself, and the figure fell below one. In other words, the bulk of the content produced did not come from data. It came from the gaps between data points.

That structure is familiar, and I have seen it in football many times. When a team loses, people do not read the metrics. They read the scoreline. The scoreline is a fact, but it is the poorest fact of all. A 0-1 defeat with xG of 2.4 against 0.6 tells a completely different story from a 0-1 defeat with xG of 0.3 against 1.1. Same score. Different nature. And the reader only receives the score.

In tennis the structure is even clearer, because the sport contains many individual points but very few high-weight points. One player wins 6-4, 7-6 because he served better on exactly two important points. Another player wins 6-4, 7-6 because his opponent shot himself in the foot in the tiebreak. Identical scorelines. Distant realities. Only second-serve points won and decisive-point conversion separate the two.

This is why I never publish a tennis judgement without at least two supporting metrics. Not because I like numbers. Because numbers are the only thing that keeps a writer from deceiving himself, and keeps a reader from being deceived politely.

Back to Sinner. When the three-month suspension was announced in February 2026, I saw two kinds of reaction. The first said three months is far too light, this is star privilege. The second said if there was no fault, there should be no sanction at all. Neither reaction requires data — each requires only a position. And both skip a harder question: which process produced this outcome, and is that process consistent with other cases?

That question is answerable with data. It demands comparing sanctions in other clostebol cases, reading the full CAS award, cross-referencing the ITF Anti-Doping Programme. That is heavy work. Nobody clicks it. An angry headline earns 40,000 views. A penalty comparison table earns 400.

To place the case in competitive context, remember that 2026-2026 was the clearest generational handover in men's tennis in two decades. Carlos Alcaraz won Roland Garros and Wimbledon in 2026. Novak Djokovic won Olympic gold in Paris 2026. Sinner won the Australian Open 2026 and the US Open 2026, then the Australian Open 2026 and Wimbledon 2026. A world number one absent for three months in the middle of that cycle is an enormous variable for any probability model — far larger than a single defeat or a minor upset.

I tried to estimate the ranking-points effect of that gap. Three months out, for a player who normally goes deep everywhere, equates to losing the chance to score at roughly three or four major events, while also forfeiting the opportunity to defend points at events won the previous season. That is a measurable gap, not a sentimental one.

And here is where I want to go against my own tribe — the tribe that believes data will save us.

The truth is that data does not automatically produce truth. It only produces the capacity for verification. The distance between those two things is where the entire sports media industry lives, and most of us live there comfortably.

I have made this mistake repeatedly. In 2026 I had a model full of variables and I believed it. I believed it enough to write a declarative headline. I did not lack data — I lacked humility. After the 2026 World Cup I removed the word "certain" from my analytical vocabulary entirely. That same year, when the pandemic arrived, I had a chance to re-test my method on a nearly clean dataset: 100 Premier League matches before the shutdown and 50 after the restart in empty stadiums.

The results stayed with me. Average pressing intensity fell from 9.8 to 11.6 — teams played slower and more cautiously without crowd pressure. Expected goals from set pieces dropped 14 percent, while penalty conversion rose 18 percent. I did not conclude that crowds cause any of this. I concluded only that when the crowd variable is removed, a different set of behaviours becomes visible. That is a controlled comparison.

The Sinner case gives me the opposite sensation. Here the problem is not too much data being misread — it is too little data being read. People do not read the dozens of pages of the CAS award. They read the 30-character headline. And that headline, however honest, cannot hold a legal process inside it.

So when someone asks me whether Sinner is guilty, I answer that it is the wrong question. The right question is: does the anti-doping system operate consistently? To answer that, we need data on all cases, not just Sinner's. We need to know how many "no fault" findings have been accepted in the past decade, what WADA's appeal success rate is, and what the average time from positive sample to final ruling looks like.

I have tried to assemble those numbers. They are scattered, hard to find, and never published as a complete dataset. That is the story worth writing: the asymmetry between how much data governing bodies hold and how much the public can reach.

And I have to say one thing plainly that few in this industry want to hear. The same match-data feed we use for analysis is also supplied to betting companies. I have worked with live data feeds and I know how they flow. Part of the data serves readers. Part serves the betting market. The line between those parts is much thinner than the industry admits. In a case like Sinner's — where information arrives in waves, each one shifting market value — the information gap itself becomes a tradable asset.

I say that not to frighten anyone. I say it to point out that when data is left blank, someone always profits from the blank. The issue is not whether a specific person exploited it. The issue is that the structure allows it, and that structure remains intact now the Sinner case has closed.

What I am tracking over the coming months is not match results. It is three things.

First, how governing bodies publish case data. If the ITIA and WADA begin releasing a standardised anti-doping dataset — number of cases, processing time, outcomes, sanctions — that will be a real signal. If not, the void will keep being filled with emotion, and emotion is always free.

Second, how tennis data platforms handle unusual gaps in playing schedules. Three months out for a world number one is a large variable in any probability model. Whoever adjusts correctly gains insight. Whoever ignores it will be surprised when season results diverge from expectations.

Third, and most important to me as someone reporting for the Australian market: whether newsrooms are willing to slow down. A story filed two hours later with three verified sources always beats a story filed two hours earlier with none. I have tested that across my own career, and I am still testing it every week.

I still open the spreadsheet every morning. I still check on-court results against model expectations. I still write a "model limitations" section at the end of every analysis — even when nobody reads that far.

Transfers are where people pay hundreds of millions for a single row in a spreadsheet — and an empty data package is always free, until it costs someone money, reputation, or an entire career. In tennis, where every point is recorded and every sample is stored, we no longer have any excuse to keep writing by feel.

The only question left is who will be the first to slow down and read the data all the way to the end.