The Empty Envelope: When Tennis Data Is Sold Without Anyone Verifying It

On the night of June 26, 2026, in a bar about seven hundred meters from Luzhniki Stadium in Moscow, I saw a familiar face. His name was Hung, a university friend of the vice president of a football club in Binh Duong. He sat at the last table, his phone glowing constantly, his hand scribbling into a small notebook. When he left, I photographed two pages left behind on the table. They were numbers. But today's story does not begin in Moscow, nor does it begin with football. It begins with an empty envelope.

Seven years later, in the press room of a tennis event in Ho Chi Minh City, I received an analysis file from a data provider. The file weighed 4.2 megabytes and arrived with a service invoice. I opened it in spreadsheet software. Player name: blank. First-serve percentage over seventy-two hours: blank. Last twelve match results: blank. Break points saved: blank. Every cell contained two capital letters and a slash.

What chilled me was not the emptiness. It was that the market had already reacted to it before I could open the file. The odds at a regional bookmaker had shifted 0.08 points within forty minutes. Someone had read the empty envelope, and someone had made a decision based on it. A file containing no information was still capable of moving money. That is the question I carried through this entire season, and it led me into a room the tennis industry rarely opens to outsiders.

Context: a market built on numbers nobody rechecks

To understand why an empty envelope still has value, you must understand who buys tennis data and why. There are three main customer groups. The first is media: broadcasters, news sites, and live feeds need numbers to fill the time between games. The second is academies and coaching teams, who need data to plan training. The third, and by far the most generous payer, is betting companies and the intermediaries who broker odds.

Since 2026, money flowing into sports data has grown exponentially. Sportradar, one of the largest providers, is publicly listed and valued in the billions largely on the strength of its live-data division. Tennis Data Innovations, a joint venture between the ATP and ATP Media, was created to commercially exploit tournament data exclusively. Every point, every serve speed, every line judge's position becomes a product that can be resold repeatedly to different buyers.

The problem is this: data products have no independent quality regulator like medicines or food. No body checks whether the 68% first-serve figure one provider sells to a broadcaster matches the figure that same provider sells to a bookmaker. When I asked a mid-level manager at a data company at a 2026 Singapore conference, he answered bluntly: "We sell data, not truth. Truth is the buyer's job."

That answer sounded like a confession. And when I traced the money from a Challenger event in Vietnam back up to international providers, I realized it was more than a confession. It was a business model.

Core analysis: three data tiers and the hole in the second

Tennis data is not a unified block. It splits into three tiers, each with its own revenue model.

The first tier is raw data from on-court sensor systems. Since Hawk-Eye entered the US Open in 2026, every major court has had cameras and sensors recording ball position, speed, and spin. This is the most reliable tier, because machines produce it, with timestamps and video cross-references. But installation and operating costs are high, so only Grand Slams, Masters 1000s, and some ATP 500s have full coverage. A Challenger in Da Nang or an ITF event in Binh Duong does not.

The second tier is manually entered data. This is the tier that generates money and also the tier that generates holes. At most events below ATP 500 level, one or two staff sit in the stands with tablets, pressing a button every time the ball lands in, lands out, or a player faults a serve. Every press is a number. Every number is a product. If the operator is tired, distracted, or paid by an untraceable third party, the number that goes to market no longer resembles the match.

And here is the point I want to make explicit: there is no mandatory cross-check mechanism between tier one and tier two. A match at an ATP 250 may have sensors on center court but not on court three. Data from both courts is merged, sold together, labeled identically. Buyers cannot tell which parts were machine-captured and which were human-pressed. They only see a smooth number in the report.

The third tier is derived data: prediction models, odds, form indices. This tier is built on the two below it. If tier two drifts by half a cent, tier three drifts by a dollar. And because tier three is what gets sold to fans and bookmakers, the smallest error at the bottom amplifies into large decisions at the top.

I spent three months cross-checking data from a Challenger event in Vietnam against replay footage from an amateur camera in the spectator stands. The result: in a match lasting two hours and forty minutes, 14 points were entered incorrectly, roughly 4.7% of total points. Three of them were decisive in the second-set tie-break. Nobody in the stands knew. Nobody in the betting market knew. The 4.7% went into the machine, was multiplied, was sold.

People call that operational error; I call it the first lesson on my home court.

Since Moscow 2026, I no longer watch a sporting event as a match, but as a money-flow balance sheet. Every number, accidental or intentional, has a price. And when the price of a number exceeds the price of the truth, someone will always choose to edit the number.

A money trail worth retelling

Back to the empty envelope in the Ho Chi Minh City press room. That data file came from a brokerage based in a neighboring Southeast Asian country. The service contract stated clearly: supply pre-match analysis reports for twelve players at an ITF Men's World Tennis Tour event in Vietnam. Contract value: 2,800 dollars a month, paid by international transfer.

But the report contained no data. Only empty fields. So what did that money buy?

I called a former employee of that company. He said something I recorded verbatim: "The report doesn't need to be correct, it just needs an identifier. The client sees the code and knows we have system access. Content comes later." In other words, what was sold was not data but access. And access can be granted before real data exists.

This is the model I call "the stamped empty envelope": the buyer pays for a legitimate shell, so that when real information arrives, they are first to know. If real information never arrives, they still hold the shell, and the shell can be resold to others with the story "we have exclusive data sources."

When the bookmaker knows in advance and the button-presser knows it, the match is just a script in the stands.

This explains why odds shifted 0.08 points after the empty file was distributed. Not because anyone read the content, but because someone knew an envelope had been sent, and that envelope would soon carry a number. In a market where information is priced by speed, simply having information — regardless of content — is already an advantage.

The counterintuitive angle: emptiness is more honest than filling

Here I want to break from the crowd. On reading an empty analysis report, the standard reaction is to immediately conclude fraud. I do not go that way. I argue that the most dangerous thing in the tennis data industry today is not empty reports, but filled reports nobody checks.

A file with all fields marked N/A is, at minimum, honest: it admits it has nothing. A careful reader stops. A responsible reader picks up the phone. The evil lies in the shell, not in the content.

Conversely, a file full of beautiful numbers — 68% first serves in, 42% return points won, 6/8 tie-break win rate — will not make anyone stop. It is too smooth to suspect. And precisely because it is too smooth, it may contain numbers generated to match a betting script rather than a real match.

I have seen this before. In 2026, when news broke that world number one Jannik Sinner had tested positive for clostebol in a March sample, the data market shook not because of the case itself, but because a wave of analysis reports was pushed out within twenty-four hours, all richly populated with his form data from the preceding six months. Not one report mentioned that those figures were collected while he was in the process of handling the case. The numbers were correct; the context was dropped.

The blind spot of the tennis data industry is not missing numbers. It is having too many numbers and too little ability to read them. A number with no origin and no date is just a carefully packaged illusion.

In the phantom season of 2026, I sat in an empty stand watching money flow into the pockets of the powerful. Tournaments had stopped, but data reports were still being sold. No match, but a contract. No score, but a prediction model. The only thing suspended was the court. The market was not.

Why this matters for Vietnamese fans

Vietnamese fans consume more tennis data every year, but most of it is imported from foreign providers without passing through any domestic verification step. When a Vietnamese news site cites a young player's serve percentage, it does not know how many hands that number passed through before reaching its own.

I checked three domestic sports outlets covering a Challenger event held in Vietnam in 2026. All three gave three different figures for the same player's first-serve percentage in the same match. The maximum discrepancy was six percentage points. No outlet named the original data source. No outlet cross-checked. And no reader demanded it.

This is where the role of Vietnamese sports journalism must be reset. In nineteen years observing this industry, I have never seen a single Vietnamese newsroom with internal procedures for data entry, verification, and archiving. Every number is copied from an unidentified source, pasted into an article, and survives for years without anyone rechecking.

I do not believe in intuition; I believe in the half-cent discrepancy in a transfer ledger. And the half-cent discrepancy in Vietnamese tennis data is far larger than half a cent.

The common thread in every case

After gathering evidence for this piece, I sat down and made a table. One column for the case. One for the data source. One for who is ultimately accountable if that number is wrong.

I built this table seven times with seven different cases across the sports industry, not just tennis. Every time, column three was empty.

Every scandal has one thing in common: the powerful stand outside the sideline but sign their names on the scoreboard. The data seller is never held responsible for the number they sold. The data enterer is never named. The data buyer is never asked how they read it.

This accountability structure breaks precisely at the joint between tier one and tier two. At tier one, machines are responsible and machines cannot be sued. At tier two, humans are responsible but humans carry no legal obligation. So the gap sits there, and money flows through that gap every day.

I record every footprint on the court so that when they wipe their hands, I can identify each hand.

The unknown: three questions I leave open

I do not publish everything in this piece, because my professional principle is to go all the way only when three independent sources agree. At this point, I have two sources confirming the "stamped empty envelope" model and a third still under verification. I choose to wait for the mature moment rather than chase speed.

Three questions I leave open for those working in Vietnamese tennis. One: who domestically has direct access to raw data from ITF and Challenger events held on Vietnamese soil? Two: with whom are those events' data-supply contracts signed, and what do the quality-check clauses say? Three: in the past decade, has any event in Vietnam ever cross-checked manually entered data against replay footage?

A thought to carry forward

Global sport is entering an era where data is valued higher than the person who generates it. A player can retire, but the dataset about him keeps earning. A match can end, but the numbers about it keep selling to people who never watched the match.

In this context, demanding a verification mechanism for sports data is no longer a technical request. It is a professional ethics request. Who has the right to turn one button press in the stands into a number that affects odds, articles, and the career of a twenty-year-old player?

The answer lies in the fact that no one answers. And that silence, in turn, is itself a piece of data.

From now on, whenever I receive an analysis file, I will not just read the content. I will ask when it was sent, who sent it, what the identifier is, and which system confirms it was entered before the match began. Without those four answers, I write one line in my notebook: not enough sources.

And I keep waiting.