The Blank Data Sheet and Football's Discipline of Admitting It Does Not Know

At minute 63, the screen in front of me had four columns. The first counted passes. The second measured distance covered. The third logged pressures. The fourth was expected goals. All four were blank. Not zero. Blank. The data feed from the provider had dropped at minute 61, and the editing software in the control room handled the incident the way it always does: it erased the cells.

In my earpiece, one voice suggested filling the gap with last season's averages. Another spoke faster: just estimate it, nobody checks. I stayed quiet. Thirty-nine years in this trade taught me that the most dangerous moment for a commentator is not when there is nothing to say. It is when there is a gap, and everyone around you wants to fill it with something that sounds reasonable.

That day I chose differently. I read the gap out loud: we currently have no positional data for either team from minute 61. The host looked at me as if I had refused to work. I had not refused. I was doing the hardest part of the job, which is stating your own limits before somebody else discovers them.

Data does not lie. The people who clean the data do.

Who counts, who cleans, who tells

Professional football runs on a data supply chain almost no viewer ever sees. People see the statistics strip in the corner of the screen, the expected-goals figure flashing after every shot, the heat map of a winger. They do not see who pressed the button that created those things, who corrected them at two in the morning, and who decided that one through ball counts as a chance and another does not.

The feed I use daily has a fairly traceable history. Opta launched in England in 2026 and quickly became the largest event-data collection network in the world. In 2026, STATS and Perform merged into Stats Perform, creating an entity that holds much of the event data broadcasters across Asia buy. Alongside sit Sportradar of Switzerland, Hawk-Eye, whose goal-line technology debuted at the 2026 World Cup in Brazil, and Second Spectrum, which tracks player positions optically. The 2026 World Cup in Russia was the first with VAR. The 2026 World Cup in Qatar was the first with semi-automated offside technology. Every one of those steps came with the same promise: we will be more accurate.

More accurate than what, and accurate by whose definition, is the question that never makes the bulletin.

Drawing on my experience watching matches in V.League stadiums and across Chinese competitions, I split the chain into four links. The first is the collector. In Europe's top leagues these are taggers sitting in the stands, usually part-time, paid per match, with high turnover. In many Asian leagues they are sports students hired by the session, trained for three days, and handed a codebook of twenty action types to distinguish in a fraction of a second. The second link is the cleaner. After the match another group reconciles the raw data and removes impossible values, such as a player covering 22 kilometres in 90 minutes, or a team holding 140 percent of possession. The third link is the definer. They decide what counts as a successful tackle, what counts as a key pass, whether a shot from inside the box after a rebound counts as a big chance. The fourth link is the narrator, which is me.

Four links, four opportunities for the truth to bend, and only the last one has a name the audience knows.

When money flows onto the pitch and data flows off it

To see why this chain matters so much, look at rights valuations. In 2026, the Chinese platform PPTV, part of the Suning group, was reported by international media to have signed a three-season English Premier League rights deal said to be worth around 700 million US dollars, the largest ever for the Chinese market. Less than a year later the deal was terminated amid the pandemic, when matches were postponed and payment flows froze. Contracts like that are not priced on passes or goals. They are priced on a belief that viewers will pay for a product that can be packaged, sliced and resold.

What makes that product sellable is data. A modern rights package includes the broadcast signal, a second-screen signal for handheld devices, augmented graphics, real-time data for the broadcaster to build its own tables, and exploitation rights on betting platforms. Every one of those layers depends on a supply chain that, when it breaks at minute 61, nobody compensates for.

That is why in May 2026 I told the leadership of my broadcaster something they did not want to hear. Global sport had frozen, broadcast contracts faced the risk of default because there were no matches to air. The meeting circled around deferring payments. I left the room and noticed a different gap: audiences were desperate to talk about football, not just to listen one way.

I used the community I had already built, and produced my own online show analysing the 2026 Istanbul final between Liverpool and AC Milan, inviting viewers to interact minute by minute and propose hypothetical tactical changes. Management rejected the idea, arguing that audiences only want live action. I did it on my personal channel. It reached 250,000 views, fifteen times a second-division commentary in the same slot.

In a stadium with no singing, I heard the future of broadcasting.

And the lesson attached matters more than the view count. Fans do not leave the stands when they can bring the whole stadium into their living room. With no terraces, value shifts from monopolising the signal to hosting a conversation. But that conversation only lasts if the person leading it is honest about what they know and what they do not.

The blank cell and the zero

Here is where I believe sports media is making a systematic mistake, and making it at the deepest layer.

On a spreadsheet, a blank cell and a zero look identical, and they are opposite truths.

A blank means we never looked. A zero means we looked and found nothing. In research, those two states sit at opposite ends of every conclusion. In football statistics they are blended into one. A young player at a provincial academy with no data at all is not essentially the same as a young player with data equal to zero. The first is invisible to the model. The second was assessed by the model and rejected. Both vanish from the list, but only one was actually seen and turned down.

The consequences of this confusion are far larger than a wrong table. When a scouting network in Asia, Africa or South America treats the silence of data as evidence of low quality, it does not merely miss talent. It creates a distorted market where a child's value depends on whether somebody happened to record them. I have watched families stake everything on an academy because of a fifteen-second clip that went viral, and I have watched better players than that go unmeasured. Scouting networks in developing countries both find genius and produce football lottery tickets and broken families. The confusion between a blank cell and a zero is the pedal that drives that wheel.

At the same time, this industry has developed a concept I want to translate into football. In systems engineering, a pipeline that returns valid structure with empty content is called a silent failure, and it is more dangerous than a loud one. A crashing programme raises an alarm. A programme that returns a correctly formatted but hollow table quietly flows into every downstream report, and at each later stage fabricated content is added to fill the hole.

Football has its own version of the silent failure. It is the scouting report written in professional language that says nothing at all. It is the match review full of strong adjectives with not one verifiable detail. It is analysis that reads beautifully and leaves no understanding behind in the reader's head.

Where the four links break

The collector breaks when speed is placed above accuracy. A tagger has to classify an action in roughly two hundred milliseconds. Six taggers on the same match can produce six different key-pass counts, with deviations commonly in the ten to fifteen percent range. In leagues without tracking technology, that error can be wider than the gap between the league leaders and fifth place.

The cleaner breaks when cleaning rules are not published. Removing an impossible value is technical. Deciding whether a shot after a rebound is a big chance is a judgement. Those two jobs are usually folded into one step, performed by a group nobody can name, following a document nobody reads.

The definer breaks when different models produce different numbers for the same match and nobody tells the audience that they differ. Expected goals is not a constant of nature. It is the output of a modelling choice: how many seasons of data, how to treat a shot with the weaker foot, whether goalkeeper position is included. Three reputable providers can publish three expected-goals values for the same shot, and no court arbitrates.

The narrator breaks when asked to fill the gap before going on air. This is the only link with a name, a face and the capacity to be criticised. Which also makes it the link most exposed to temptation.

I know that temptation well, because I have beaten it and I have lost to it.

The match that taught me to trust data

In July 2026, in the AFC Champions League quarter-final between Guangzhou Evergrande and Shanghai SIPG, I used positional data from twelve on-pitch sensors to show that SIPG's 4-2-3-1 effectively became a 3-4-3 in possession, and that this transformation stretched Evergrande's back line horizontally to the point where both centre-backs had to keep leaving their positions. A male colleague sneered: women can only read numbers, they do not understand football. Three days later, head coach Andre Villas-Boas confirmed exactly that in a press conference. My analysis was shared 8,400 times. My under-25 audience grew by 210 percent.

That success made me complacent. I believed I could beat any prejudice with data, and I forgot that data is only right when the input is right.

The match that taught me to fear data

In June 2026, at the Nizhny Novgorod stadium, during Croatia's 2-0 win over Nigeria, I mispronounced Ante Rebic's name three times in the first half. Social media reacted instantly. The confidence of the previous year had made me careless about checking identities, and I paid for it.

That night I did not delete the clip. I rewatched the whole match, took notes on Croatian pronunciation, and spent the thirty days after the tournament building a Vietnamese transcription table for 736 players, published free on my blog. The post was shared 12,000 times, became a reference for several broadcasters, and built me a loyal following I still have today.

A table of 736 names is not discipline. It is an apology, systematised.

While building it, I learned something no training course teaches. A player's name, even mispronounced, is how we open our arms to a culture. The error does not lie in getting a syllable wrong. It lies in refusing to spend thirty days understanding that behind every name there is a person, a family, a town.

The match that taught me a person can become a number

In June 2026, in Bucharest, France lost to Switzerland in the Euro round of sixteen on penalties. Kylian Mbappe missed the decisive kick. Amid the storm of criticism, I received information from a contact in the transfer world, a relationship built through the pandemic-era live shows. Real Madrid had just had a bid formally rejected by PSG: 180 million euros for Mbappe. And the young player had been psychologically shattered before the match had even started.

I wrote a 3,000-word analysis. I did not defend Mbappe. I explained the psychology of a human being turned into a transfer figure, of having your value publicly listed before the whole world during exactly the period when you must perform as a footballer. Le Parisien cited the piece.

Since then I never write about a single match in isolation. I always place it inside the economics and the transfer market, because the match is only the surface of much deeper water.

The contrarian angle: clean data is the frightening kind

My industry holds an almost religious belief that the enemy is dirty data. I think the real enemy is clean data that does not say who cleaned it.

A dataset described as clean is one where every original ambiguity was decided on your behalf by somebody. Those decisions are erased from the surface, leaving a flat table that carries the authority of a table. Dirty data incriminates itself. You look at it and see gaps, outliers, warnings. Clean data accuses nobody. It simply sits there, tidy, waiting to be cited.

In football this means the more beautifully a dataset is presented, the harder it is to trace. The prettiest tables are usually the ones that passed through the most hands. And every hand has a reason to want the number to look slightly more reasonable.

The second contrarian point concerns the wave of automated content. Language-model systems can now write previews and post-match texts from data, and they write fluently. Their structural flaw lies elsewhere: they are optimised to produce coherent text, and the sentence there is not enough data to conclude is not coherent. So they always have an answer. A machine that never says I do not know is not an analyst. It is a fountain.

And here is the uncomfortable part. The industry's incentive structure rewards the fountain. Bookmakers need a number before kick-off. Broadcasters need a graphic before going on air. Sponsors need an index to attach a brand to. None of them pays for a blank cell. The market prices certainty, including fake certainty, above honesty. That is the entire economics of this problem in one sentence.

Data only becomes rebellion when somebody is brave enough to believe it. And bravery here usually means accepting that you will be seen as someone who cannot do the job.

Two laboratories, one mistake

I live in Guangzhou and work for the Chinese market, while still following Vietnamese football every week. Looking at the two markets side by side, I see them making the same mistake from opposite directions.

The Chinese market bought rights faster than it built the infrastructure to exploit them. Major platforms invested in augmented broadcasts, second screens, real-time data layers overlaid on the picture. The collapse of the Premier League rights deal in 2026 showed that the valuation rested on assumptions about monetisation rather than on a tested model. When the assumption broke, the contract broke with it, while the data stayed in the hands of foreign providers.

The Vietnamese market moved differently. Domestic league rights values rose, but data infrastructure inside stadiums remains some distance from regional standards. Much of the event data is still recorded by hand, dependent on a person in the stands. That means Vietnam does not only import pictures. It imports the narratives built on a dataset it does not produce and therefore cannot verify.

One side buys data it cannot yet monetise. The other monetises a match it has not yet measured. Both are paying for the same gap: the absence of a layer of people responsible for explaining data in the language of fans rather than the language of the spreadsheet.

With the 2026 World Cup expanding to 48 teams, data volume will grow exponentially, while the number of people who genuinely understand where data comes from will not. The distance between those two figures is where fabricated stories will breed, because there will always be a gap to fill and always somebody willing to fill it.

The most valuable thing I learned in thirty-nine years

I am not writing this to convict anyone. The badly paid tagger is part of the system, not its cause. The cleaner follows the rules they were given. The narrator works under broadcast deadlines. If any individual deserves interrogation, it is the structure that designed a profession in which saying I do not know is treated as failure.

The most valuable error of my career has 736 versions, and every one of them was worth making again, because they forced me to build a process instead of a promise. Every time I am wrong, I correct it publicly within twenty-four hours, explaining precisely which hypothesis failed and why. The trust I hold today does not come from being wrong less often than others. It comes from audiences knowing exactly what I will do when I am wrong.

So the next time you see a perfect statistical table for a match you have just watched, ask one simple question. Who pressed the button, and who cleaned up what that person pressed.

Football will not become more transparent through more data. It will become more transparent when the number of people willing to name the gap before going on air grows. The next platform for sports media is not camera technology. It is somebody saying into a microphone that we lost the feed at minute 61, and the audience staying anyway. If audiences stay after a sentence like that, then this industry has just found the thing I have been searching for across thirty-nine years: a business model built on honesty rather than on the silence of the table.