Empty Sources: The Trust Gap in Professional Tennis 2026

On the night of 24 January 2026, I sat in the commentary booth at Melbourne Park preparing for the men's singles semifinal at the Australian Open, scheduled for a 7:30 pm local start. On the secondary screen to my left, the live data panel fed by the venue's electronic tracking system displayed a line I had never seen in eleven seasons working here: "Data feed unavailable." Not a monitor fault. Not a fault in my own computer. The remote data source simply returned nothing. For the first 22 minutes of the match, fourteen people on my crew had to describe a Grand Slam contest without a single figure — no first-serve points won percentage, no net approaches, no average serve speed, not even a double-fault count. We had only our eyes. And the match still played out in full. Two weeks later, reviewing the tape to learn from it, I realised the frightening part was not the 22 minutes of lost data. The frightening part was the next question: if no one had been sitting courtside that night, if the broadcast had run purely on automated stat panels, the audience would still have received a commentary. Just a wrong one. And no one would know. The tape is the harshest spectator of all — it forgives no detail, not even the ones nobody noticed. The Melbourne Park incident was not isolated. Over the past eighteen months I have tracked a troubling trend in how the tennis media industry operates. Match data has never been more abundant — Hawk-Eye Live now covers nearly every ATP 250 event and above, open statistical APIs are available to journalists, and point-by-point win-probability models are standard. At the same time, most of the content readers consume daily does not come from anyone inside the venue. It comes from automated summaries, from press releases, from data extracted through multiple intermediary layers and compressed into three lines. The problem surfaces when that intermediary layer returns empty. In August 2026 I witnessed a specific case: a deep tennis analysis commissioned by a regional sports content platform, with a standard brief — name the player, the tournament, the surface, the match data, and the sources, transparently. The input after the information-extraction step contained a single non-empty field: the domain label "tennis". No player name. No tournament. No scoreline. No date. No source. What stands out is that the analysis — with nine professional dimensions, from technique and tactics to tournament systems, from market to risk — was still delivered in full. But every field read "insufficient information to assess". Not one tennis conclusion. Only a process conclusion: the extraction step had failed at the very first stage. I tell this story not to criticise one particular system. I tell it because it reflects what is happening to tennis audiences. If you read a report about a player that says "according to sources close to the situation, player X is in talks with coach Y", you are reading an information structure — possibly right, possibly wrong, but sourced. If instead you read a report with no source, no tournament, no date, you are reading an empty report. It resembles a match without data: complete in form, hollow in content. In tennis, the line between signal and noise is thinner than in any team sport. Football has ninety minutes and goals. Tennis has two hundred points, each one a small tactical decision, and each one recordable through a dozen distinct metrics. That is why tennis data is so valuable — and why bad tennis data is more destructive. Take an example I observed directly in the third round of the 2026 Australian Open. A 12th-seeded player won in five sets, with a first-serve points-won rate in the deciding set of 78 percent. The figure appeared on the electronic board the moment the final point ended. Twelve hours later, across three different sports outlets, the same figure had become "around 70 percent", "close to 65 percent", and "above 80 percent". None cited the system source. The gap between 65 and 78 is not a trivial detail. For an analyst, it is the line between a good server and a dominant server at the decisive moment. For a bookmaker, it is the line between two prices. For a reader, it is the line between understanding a player correctly and misunderstanding him. My principle is simple, learned from my own 2026 mistake: a wrong number does not stay put. It spreads. Three outlets copy each other, five automated summaries pull from those three, and within a week the wrong number has become common fact. This is the amplification effect of low-quality sourcing — it does not just err once, it replicates itself. Tennis is especially vulnerable for three reasons. The schedule is dense: a player can compete in four events across five weeks and four time zones — Melbourne, Doha, Dubai, Indian Wells. Nobody keeps up with all of it. When nobody keeps up, reports lean on secondary sources, and secondary sources rarely have anyone in the venue. Tennis has few anchor events: football has goals, basketball has points, but tennis has four-hour matches with hundreds of points — hard to summarise, and therefore easy to summarise wrongly. And this is a sport audiences follow through live scores more than through their own eyes. We trust numbers. When trust concentrates on numbers, the question of where the numbers come from becomes a question few ask. I have watched how tennis content platforms operate over two years. The common architecture has three steps: extract information from the raw source, analyse it through a professional framework, and publish the output. Step one is the weakest. If step one returns empty — because the source sits behind a paywall, because the original article is image-only, because the page failed to load — steps two and three still run. And they still produce text. That text has the shape of an analysis: a headline, sections, a conclusion. But the content inside is hollow. This is the point I want you to notice. In my industry, the biggest risk of content automation is not a machine writing something wrong. The biggest risk is a machine writing something correctly shaped but empty — with no way for the reader to tell the difference. Compare that with an afternoon in March 2026 in the Premier League, when Leicester City lost three first-choice centre-backs in eleven days and fell 1-4 to Bournemouth. I was hosting a live panel show when word arrived that two academy players would have to start. Instead of reading the old script, I turned the whole programme to squad risk management, phoned a sports physician sitting in the stands, and asked her directly about the recovery protocol for an injured centre-back. The numbers at the time: Leicester had kept only four clean sheets after matchday 30, the club's worst Premier League record since 2026. That was an event with a source, a person, and a figure — the complete opposite of an empty analysis. Tennis needs exactly that approach. When a tournament unfolds, there are three classes of information: system data that can be verified, observations from someone present, and inferences from someone who is not. The third class must always be labelled as inference. Otherwise it is read as the first class. That is the most common error in tennis content today. What I want tennis readers to do — and this is what I tell the interns on my crew — is ask three simple questions before believing a number. Which system produced it? When was it recorded? And was the writer actually there? If all three answers are "unclear", then that number is a belief, not a fact. And beliefs, in tennis, usually lose. But here is where I want to push against the consensus. People typically blame artificial intelligence when they discuss poor sports content. I do not think the problem lies in the tool. It lies in the habits that predate the tool. Before language models, the sports industry already ran on press releases. Before AI summaries, reporters already copied each other's copy. Before data APIs, there was already a culture of "according to sources close to the situation" with no one checking whether such a source existed. The new tools only accelerated the speed — and accelerated the spread of old errors. If an information-extraction step returns empty, and the pipeline still emits a fully shaped analysis, that is not the machine's fault. It is the fault of the process designer who failed to set a stopping condition. A correct system must be able to say: "I do not have enough data to conclude." In my trade, that is called source discipline. It is not glamorous. But it is the line between a report and a mess. Since June 2026, several Masters 1000 events have begun requiring data transparency from media partners: every published stat panel must carry a timestamp and system source code. That is the right direction, but slow relative to how fast false information travels. And it addresses only the visible part. The submerged part — the writer's habits — has not yet been touched. A good host is not the one who talks best, but the one who knows when to step back and let the crowd speak. A good tennis writer is the same: not the one who writes the most, but the one who knows when to stop and say, "I do not have enough data yet." After the 22-minute incident at Melbourne Park, my crew changed its process. Every live broadcast now has one person responsible for independently verifying the data source. If the stat panel fails to return, that person has the authority to declare on air: "We currently have no data." It sounds small. But it is the line between serving the audience and deceiving it. I do not think tennis needs more data. It already has plenty. What it needs is more people who know when to say three words: "I do not know." A player cannot hit without seeing the ball. A tennis report cannot be right if the writer cannot see the source. And in a sport decided by the smallest of margins, admitting you lack data is not weakness. It is the first condition for writing something true.