When the Analysis Room Falls Silent: The Empty-Data Trap in Modern Football

Chengdu, one March morning, the annual season just finding its rhythm. In the coaching staff's meeting room, the projector lit up a blank slide. The data analyst stood beside the screen, pen in hand, and said the sentence I have heard hundreds of times: "There are no red flags." The slide carried no expected-goals chart, no pressing-intensity metric, no heat map. It was empty. And in the language of the analysis room, that emptiness was translated into one word: safe.

I sat at the far end of the table. Outside the window, the team was warming up on the training pitch. What I knew, and what that slide did not know, was that across the last three matches, the midfield had lost the ball in dangerous areas fourteen times. I counted with my eyes, not with a machine. No line of data in the club's system recorded that, because the system was running in silent-failure mode. It did not raise an alarm. It simply stayed quiet.

In modern football, silence is becoming the most dangerous kind of data.

We begin in Chengdu, where barriers are not as towering as people assume. In 2026, while working as a tactical commentator for a local sports channel, I reconstructed fourteen passing sequences from a second-tier match in software, just to prove something the machine could not see. A male colleague mocked me then, saying a woman could not possibly understand high pressing. My video drew one hundred twenty thousand views, six times the official channel. The lesson that year was not the view count. It was this: the club's analysis system had missed that entire blind spot, and missed it quietly.

That quietness is the subject of this article.

Context: the promise of a decade of data

Within fifteen years, the analysis department has become indispensable at nearly every professional club. Data budgets, analyst hiring, contracts with motion-tracking providers — all have grown exponentially. The promise was clear: if you measure everything, you will decide better. If you quantify pressure, space and chance quality, you will see match outcomes in advance.

That promise is partly true. But it contains a gap few are willing to look at directly.

Data in football is not merely a pipeline carrying numbers from the pitch to a screen. It is a chain of layers: collection, cleaning, labelling, modelling, interpretation, presentation to the coaching staff. Every layer can break. And when a layer breaks, the result is often not a wrong number. The result is a blank.

That blank is what is truly frightening.

I once said in an internal presentation that wrong data is dangerous, but empty data is lethal. The room laughed. They thought I was exaggerating. But imagine a centre-back reading the pre-match report and finding the "opponent weaknesses" section empty. What is he to understand? He has no option but to read it as: the opponent has no weaknesses. Or worse, he reads it as: there is nothing to worry about. Both readings are wrong, and both lead to the same outcome on the pitch.

In an industry where every decision is justified by a spreadsheet, a blank spreadsheet becomes the strongest justification of all. The reason is simple: no one is accountable for an empty cell. The coach cannot argue with a chart that does not exist. The analyst cannot be questioned about a metric that was never produced.

In 2026, when the World Cup in Russia took place, I was invited to write a column for a major football outlet. In the quarter-final between France and Uruguay, I found Didier Deschamps deploying Antoine Griezmann deep to form a variant 4-4-2 that neutralised Edinson Cavani. I measured the distance between France's lines — nineteen point five metres — and wrote a two-thousand-word analysis overnight, ahead of even the European press. It was shared forty thousand times, and coaches in the Chinese Super League phoned me for advice on countering counter-attacks.

The Russia World Cup taught me that attack is a form of expression, while defence is the answer. But there was one thing I did not write in that piece, because I was not yet brave enough: most of the data I used to measure those distances I had rebuilt by hand from video, not drawn from the organiser's system. That system was not wrong. It simply lacked line-to-line distance data for every sequence. What I did, in the end, was fill a blank with the human eye.

And that is precisely what modern analytics is forgetting.

The mechanism of silent failure

To understand why blanks are dangerous, we need to look at how a football data pipeline operates.

The first layer is collection. Tracking cameras, sensors in the ball, devices on players' backs, event-labelling software. The second layer is cleaning: removing noise, synchronising time, fixing identification errors. The third is modelling: computing expected goals, pressing intensity, action value. The fourth is interpretation: turning numbers into language a coach understands. The fifth is presentation: turning that language into a decision.

An error at layer one produces wrong data. An error at layer three produces skewed data. But an error at layers four and five produces the worst of all: a blank presented as a conclusion.

Take a concrete example. A club hires a new data provider. The provider lacks sufficient historical data for the league the club plays in. As a result, metrics such as defensive expected goals, or successful pressing counts, return empty values for the entire upcoming opponent list. The analyst opens the table, sees the empty cells, and writes in the report: "Insufficient data to assess." The coach reads that line, nods, and walks into the match without any quantitative information about the opponent.

What stands out is not the lack of data. What stands out is that the lack of data has been legitimised into a valid decision.

I call this phenomenon silent failure. It has three features.

First, it raises no alarm. The system does not error, does not flash red, does not send an email. It returns a blank, and the blank sits quietly in the spreadsheet waiting to be read.

Second, it has no accountable subject. No one is responsible for an empty cell. The provider says data was insufficient. The analyst says he reported honestly. The coach says he asked. The loop of responsibility closes without anyone being held to account.

Third, it manufactures false confidence. In the mind of the report's reader, a slide with no red flags equals a slide with no risk. The absence of evidence is read as evidence of absence.

Those three features combine into a trap. And the trap does not exist only at small clubs short of budget. It exists at the richest clubs too, where the analysis department has twenty people and still contains data zones left blank because no one was tasked with filling them.

I once witnessed a textbook case in a league I have followed for years. A major club prepared for a derby. The pre-match report ran forty pages, full of charts. But one section was left blank: the opponent's ability to defend crosses on the left flank. The reason: the data provider could not label that type of sequence for the opponent's left-back, because he had just moved from another league and lacked a sufficient sample.

Match result: the big club lost by two goals, both from crosses to that left flank. After the match, the coach said at the press conference that no data had shown that weakness. He was right in the literal sense. But he lost a match because of an empty cell.

This is where I must state my position clearly. I am not against data. I use data every day. But I oppose the way this industry is turning data into a ceremonial performance, in which having a spreadsheet matters more than understanding the match. When data becomes ceremony, blanks become ceremonial gaps — and no one dares to point out that the emperor is wearing paper.

There is a deeper consequence I want to dissect next. When data analysis becomes detached from the actual rhythm of the dressing room, it does not merely miss information. It also creates a language that coaches must speak to be considered modern. And in that language, silence is a statement.

When blanks become statements

I remember a meeting in Chengdu a few seasons ago. The analyst presented a report on the upcoming opponent. The report had every necessary metric except one: the pressing structure when losing the ball in the opponent's half. He explained that the provider only tracked pressing in one third of the pitch, because the league had not purchased the extended data package. He went on: "But based on available data, the opponent does not press high. So we can build up from the back comfortably."

That sentence was the moment I recognised the problem. A data blank had been converted into a tactical conclusion simply by a connecting sentence. "No data" became "no press". And "no press" became "we can play short passes freely".

Three steps. There was no evidence for the second step.

I raised my hand and asked: "Do you have our fourteen losses in the middle third over the last three matches?" He said no, that data belonged to another package. I asked again: "Then how do you know the opponent does not press?" He went quiet.

That silence, in a room full of people, was the most valuable piece of data of the whole morning.

The problem was not the analyst. He followed procedure correctly. The problem was the design of the procedure: it allowed a blank to pass through the door unchallenged, and allowed the presenter to fill that blank with a guess without marking it as a guess.

In finance, people call a variable with missing data a missing value. Serious models must declare missing values and handle them by a dedicated method; they are not allowed to skip them. In football, most reports have no such convention. An empty cell is just an empty cell. The reader decides how to interpret it.

And under the pressure of a big match, people always tend to read an empty cell in the way that benefits them.

This brings me to a central question: if blanks are so dangerous, why does the industry not address them systematically?

The answer lies in the incentive structure. Data providers are paid to deliver data, not to admit blanks. Analysts are judged by the number of reports they produce, not by the number of blanks they detect. Coaches are judged by match results, not by the honesty of their decision process. No one is rewarded for saying they do not know.

And so silence becomes the optimal strategy for everyone in the chain. That is why it persists and spreads.

The execution blind spot: when empty data meets the pitch

This section is the one I consider most important, because it moves the problem from the meeting room to the pitch.

A data blank is not a harmless trap. It is a trap with concrete output on the pitch. And that output usually appears at moments when no one can correct it in time.

Imagine a situation I have encountered many times. Your team prepares for a match in which the opponent habitually changes formation mid-game. Data on this habit exists, but it lives in a source your system does not connect to. In the report, that section is empty. In the meeting, no one mentions it. On the sixtieth minute of the match, the opponent switches from four defenders to five. Your team takes fifteen minutes to adapt. Those fifteen minutes are enough to concede a goal.

That goal, on the scoreboard, is the fault of a player or the coach. But at a deeper layer, it is the fault of an empty cell left undisturbed throughout the preparation week.

I say this not to blame the analysis room. I say it to point out that empty data is not a technical problem. It is a human one.

Because when a blank is left undisturbed, the question is not "why is the data missing". The right question is: "who will fill this blank, with what, and based on what evidence?"

In thirty-four years of following this industry, I have not seen a single club answer that question systematically. I have seen clubs answer it better than others, but no one has turned it into a mandatory process.

My three-evidence principle originated precisely here. After the 2026 incident in Chengdu, I set myself a rule: every tactical claim I make must have at least three concrete in-match situations to illustrate it. No three situations, no claim. If I reach two situations but lack a third, I write "insufficient sample" instead of a conclusion.

That rule did not make me slower. It made me harder to challenge.

But that rule works only for an individual. The problem is at system level.

An honest analysis system must be designed to do three things that most current systems do not.

First: mark every blank clearly from the data layer, and keep that label through the entire chain until it reaches the coach. A blank must announce itself, with a symbol the reader recognises immediately.

Second: separate two kinds of blank. A blank because data was never collected is entirely different from a blank because the phenomenon genuinely did not occur. Both appear as empty cells, but they carry opposite meanings. Confusing the two is the source of most tactical errors I have witnessed.

Third: attach accountability to blanks. If a blank is left undisturbed and causes consequences, someone must be responsible. Not to punish, but to create the incentive to fill it before it is too late.

Those three tasks sound simple. But to carry them out, the industry must abandon a deeply ingrained habit: treating the report as a product for display rather than a tool for decision-making.

Evidence from the empty stadium

2026 was an unwanted but invaluable experiment. The empty stadium is the largest laboratory modern football has ever had. When the noise disappeared, everything the data normally ignores became clearer.

During those months, I watched hundreds of matches without crowds. And what I realised lay not in what I saw, but in the gap between what I saw and what the data recorded.

One example. In a league match, the team I was following pressed very aggressively in the first twenty minutes. My eyes saw it clearly. But the system's pressing-intensity metric sat at an average level, because the system's benchmark had been built from the previous season's data — a season with crowds, home pressure, a different tempo. The system was not technically wrong. But it had left one enormous variable blank: the no-crowd context.

The empty stadium strips away excuses. It does not make a team weaker. It makes weaknesses normally masked by noise become visible. And the data, calibrated for a world with noise, suddenly becomes skewed without raising any alarm.

2026 taught me: a team stands firm through its system, not its line-up. But it also taught me the reverse: a data system stands firm through its context, not its metrics.

When the context changes and the metrics are not recalibrated, blanks appear everywhere. They are not in any specific cell. They are in the entire table.

This is the most dangerous kind of blank, because it is invisible. There is no empty cell to see. The table is still full of numbers. Only the conclusion is wrong.

The counter-intuitive angle: absence is not evidence

Here I want to frame the problem in the opposite direction to the prevailing intuition.

The prevailing intuition in the industry says wrong data is the greatest enemy. By that logic, if we clean the data, fix identification errors, standardise formats, everything will be fine.

I believe that intuition is right but insufficient, and places the emphasis in the wrong place.

The greater enemy of tactical decision-making is not wrong data. The greater enemy is empty data presented as complete data. Wrong data can be detected, because it contradicts observable reality. Empty data cannot be detected, because it contradicts nothing. It simply does not appear.

A wrong number will make a coach ask a question. An empty cell will make a coach look away.

In my presentations, I am known for a style built on decision matrices. I list options down the rows, criteria across the columns, then strike them out until one option remains. But there is one rule I always follow that few notice: I never leave an empty cell in a matrix without bolding it. Every blank is bolded and annotated "no basis yet".

The reason is practical. Readers of a matrix tend to trust what they see. A white cell between numbered cells will be filled in by the eye with an assumption. Bolding it is the only way to stop that automatic filling.

This means the analyst must voluntarily make statements about his own ignorance. In an industry where credibility is built by appearing to know a lot, this is a counter-intuitive act. It demands courage.

But I believe it is precisely that act which creates real credibility.

Let me illustrate by comparing two schools of analysis. The first, common at big clubs, builds comprehensive models and tries to quantify everything on the pitch. The second, common at mid-sized and small clubs, focuses on a few key metrics and accepts leaving the rest blank.

Both have strengths. But the risk of the first is far greater, because a comprehensive model creates the illusion that there are no blanks. When a model returns a number for every variable, the reader tends to forget that many of those numbers are merely the result of interpolation. And interpolation, in many cases, is just a sophisticated way of filling a blank with a guess.

I once received a report from a leading analytics company in which the expected goals of a young striker were projected from three hundred minutes of play. Three hundred minutes. A third of a season for a substitute. The model did not return an empty cell, because it was designed to always return a number. But that number concealed a reality: we know almost nothing about this player at professional level.

That is empty data disguised as complete data.

And I argue this is far more serious than simply lacking data. Because when data is genuinely empty, at least we know we do not know. But when empty data is disguised, we believe we know, and we decide on the basis of that false belief.

When the right question is replaced by a convenient answer

There is a psychological dynamic in coaching meetings that I think needs naming.

Coaches, especially young ones building a career, want to appear decisive. Decisiveness, in the language of the meeting room, is usually shown by reaching a conclusion quickly. "I have seen enough, we go with Plan A." That sentence builds credibility. The opposite sentence — "I lack enough data to decide" — creates hesitation, and hesitation is read as weakness.

So blanks always have an incentive to be filled quickly. People do not want to leave an open question in the meeting room. They want to close it with a convenient answer, even if that answer rests on no evidence.

This is the biggest blind spot in football's decision-making process. It is not in the quality of the data. It is in the culture of the meeting room.

In a culture that treats fast decisiveness as a virtue, blanks will always be filled with guesses. And a guess, once presented in a beautiful spreadsheet, becomes a fact.

I learned this from my own career. In 2026, when I first started at a sports newspaper and worked as a correspondent in Madrid, I wrote with great confidence. I believed a good writer was one with answers. It took many years for me to understand that a good writer is one who asks the right questions.

The same applies to tactical analysis. A good analyst is not one with an answer to every question. He is one who knows which questions have no answer yet, and says so plainly.

Eight Olympic Games, eight World Cups, many seasons of the Giro and the Tour de France — I have moved through enough environments to see that this principle holds in every sport. In cycling, the analytics team can measure power, heart rate, pedal force. But they cannot measure the will of a rider on the third climb. There is no metric for that in any team's spreadsheet. And when such a cell is left empty, teams always tend to fill it with an assumption.

This is not a problem unique to football. It is a universal problem of every industry undergoing digitisation.

How a blank should be filled

I do not want this article to end at criticism. I want to show how a blank should be filled.

Principle one: a blank must be detected early, not after the match. This requires the analysis room to cross-check its data against at least one independent source — video, scout reports, or direct observation. A blank is truly filled only when at least three sources confirm the same conclusion. This is the basis of the three-evidence principle.

Principle two: the person filling the blank must be a direct observer, not a spreadsheet reader. In football, the direct observer is the scout, the assistant coach, or the head coach himself. A spreadsheet cannot fill a blank. It can only show that the blank exists.

Principle three: conclusions based on direct observation must be recorded in the same document as quantitative data, and must be marked as "observation", not "measurement". This distinction matters, because the credibility of the two types of information differs, and the reader needs to know which he is reading.

Principle four: there must be one person responsible for each important blank. Not to punish when consequences follow, but to ensure that the blank is not forgotten during the preparation week.

These four principles are not a technical solution. They are a cultural change. And like all cultural change, they are far harder to implement than writing a new model.

But I believe they are necessary. Because if this industry does not learn to treat blanks properly, then data — the very thing created to make decisions better — will keep being a shelter for bad decisions legitimised.

From Chengdu to every analysis room

Back to that March morning in Chengdu.

After the analyst finished his presentation, I raised my hand. I said the slide was blank in the midfield ball-loss section, and I wanted to know on what basis we concluded there was no problem. He replied that the system did not record that type of sequence.

I proposed spending thirty minutes reviewing video of the last three matches together. No software, no data. Just fifteen people in a room, one screen, and one person counting.

We counted fourteen losses in dangerous areas across three matches, and nine of them originated from the same positional error in midfield. Those nine sequences were not in the report. They were in the video.

That meeting ran nearly an hour longer. We adjusted one detail in our build-up structure. The detail was not large. But it came from a blank that had been filled properly.

What I want to convey through this story is: the problem is not technology. The problem is the attitude toward blanks.

A good analysis room is not the one with the most data. A good analysis room is the one that knows clearly what it does not know, and knows how to fill those unknown gaps with direct observation.

This is why I still believe in the pitch. Not because I distrust data, but because I know its limits. The pitch is where every blank must eventually be filled, whether one wants to or not. The match does not allow you to leave a cell empty. It gives you only two choices: either you prepare for that blank, or you pay for it.

I have chosen to prepare.

And that choice begins with a small habit: every time I open a report, I look for the empty cells before the numbered ones.

You could try doing the same thing this week.

When your team walks out for its next match, try counting how many blanks were skipped in the preparation report. Not to find fault. But to discover what the team's system is actually seeing, and what it is actually overlooking.

Because in modern football, a blank in a spreadsheet often tells you more than all the complete numbers in it.