Football Analysis Built on Empty Data: Why the Dressing Room Remains the Final Court

Football Analysis Built on Empty Data: Why the Dressing Room Remains the Final Court

6:40 a.m. in Shenzhen

I opened my laptop before the city had switched on all its lights, and a file arrived. Eleven pages. The cover was handsome: bold headings, neatly drawn tables, a one-to-five-star rating scale, a "Risk Matrix" spread across six rows, an "Industry Transmission Diagram" sketched in ASCII characters, a glossary at the back explaining xG, PPDA, PSR, FFP, transfer amortisation, illegal approaches to players. Whoever wrote it had spent hours formatting it. Every cell was aligned. Not a single line was crooked.

The document contained nine sections. All nine answered the same way: "N/A — insufficient information."

No team names. No player names. No scores, no dates, not one figure. Only a line in the input integrity check noting that Step One's list of information points was empty, and therefore any conclusion drawn in Step Two would be fabrication. The writer had chosen not to fabricate. In nine years in this trade, I had never held a document so honest that it was useless.

I read it three times. By the third reading I realised I was holding an artefact far more interesting than its surface suggested.

What football is written with

Every day, the global sports content industry produces an enormous volume of text, most of it generated to fill a quota. How many articles a site needs, how many words an article needs, how many times a word must appear in a search engine — all of it is arithmetic. That arithmetic does not care whether there is anything real to say.

In professional content operations, the workflow usually runs in two stages. Stage One reads sources, extracts events, identifies entities, records information points. Stage Two builds the analytical frame: tactics, finances, results cycles, league landscape, regulatory compliance, dressing room, risk, media narrative, industry transmission. Stage Two is only meaningful when Stage One has material. When Stage One is empty, Stage Two becomes a formal machine: it still runs, still outputs every section, still prints every table — but every cell is blank.

I first recognised this paradox at seventeen.

On 10 July 2026, in Saint Petersburg, the World Cup semi-final between France and Belgium. I sat in front of a camera in a Shenzhen apartment and told a few thousand viewers that Didier Deschamps would have France press high, that Belgium's back line would be smothered in their own half. I said it with enormous confidence. I had diagrams. I had lineups. I had arrows showing movement.

France held under forty per cent of the ball, deliberately ceded territory, and won 1-0 through Samuel Umtiti's header in the 51st minute from a corner. Deschamps did not press. He built a mid-block, invited Belgium forward, and released Kylian Mbappé into the space behind the defence.

I was mocked. I did not delete the video. I rewatched all ninety minutes for seven consecutive days, noting every touch, every change of direction, every metre of movement from every player. That was the day I learned that a beautifully drawn diagram without verified data is just a painting. And a beautiful painting is the hardest kind of lie to detect.

Three years later, on 3 July 2026, in Rome, the Euro quarter-final between Ukraine and England. I was a data contributor for a football site, updating live. During half-time I suffered appendicitis and was taken to hospital. I sat on a hospital bed, a drip in the back of my hand, using a laptop and a phone to cover the remaining forty-five minutes. I split the work between two remote colleagues: one handled data, one checked events, while I set the frame and edited. England won 4-0, with a Harry Kane double and goals from Harry Maguire and Jordan Henderson. The piece was finished twelve minutes after the final whistle.

Writing from a hospital bed, I understood that the pulse of a match never waits for anyone. And from that moment I also understood that in a crisis, the only thing that preserves professional dignity is a clear process: identify the core information, sort the data, delegate, cross-check.

But there is a variant of crisis that process cannot solve: when there is nothing to identify.

When xG has no context

Expected goals, xG, estimates the probability that a shot becomes a goal based on location, angle, the type of pass that preceded it, defensive pressure and a handful of other variables. It is a good tool. It is also the most abused tool of the past decade.

Based on my experience covering matches in both the Super League and European competitions, the most common error in reading xG is treating it as a verdict independent of people. A team with higher xG is usually described as "deserving to win", when in fact the metric only says that team created chances from better positions. It does not say who shot, with which foot, at what minute, whether they were leading or trailing, and above all it does not say how much confidence the shooter still had.

Brentford is the example I return to. On 29 May 2026, the small west London club beat Swansea 2-0 in the play-off final to reach the Premier League for the first time in seventy-four years. In 2026-22 they finished thirteenth. There is nothing miraculous in that if you look at the xG table: Brentford sat mid-table for attacking output that season. The difference lay elsewhere — set-piece organisation, timing of accelerations, and a recruitment philosophy that only buys players who fit a model defined in advance.

This is the blind spot of metrics: they measure what happened, not what was coached to happen. A team that practises corners three hundred times a week will have better corner numbers, but those numbers do not tell you who coached them three hundred times, when, and after how many sessions the players began to believe in it.

Conversely, some teams have beautiful xG and poor results for months. The standard explanation is "bad luck". Sometimes that is right. More often it signals a problem that does not appear in data: a squad losing faith in the idea, or a group of players performing to protect their own positions rather than to win. No model scores that. You have to be in the room.

PPDA and the pressing trap

PPDA — passes allowed per defensive action — measures pressing intensity. The lower the PPDA, the more aggressively a team presses.

Marcelo Bielsa popularised the concept at Leeds United. His team pressed in ways few dared, and for long spells they posted the lowest PPDA in the division. The media called it beautiful football. In 2026-22, Leeds conceded seventy-nine goals and survived only on the final day, 22 May 2026, beating Brentford 2-1.

A low figure here carries two opposite meanings. First: the team controls the game, pressures constantly, forces long balls and turnovers. Second: the team chases the ball in vain, the midfield is dragged out of shape, and every intercepted pass becomes a counter-attack into the space behind.

PPDA does not distinguish between the two. It gives one number. The reader needs context: where on the pitch does this team press, who triggers it, and most importantly, who runs back first when the press fails.

No metric explains itself. Every metric needs a person standing beside it to translate.

I learned this in a season without crowds. In 2026, when the pandemic forced matches in China to be played in empty stadiums, I was embedded with a club and had access to both the training ground and the dressing room. A match without cheering still tells you more than an entire noisy season. In a league without fans, I could hear boots striking grass more clearly than the referee's whistle. And I could hear things that, in a packed stadium, no one could hear.

Five winless matches at Shandong

In 2026, when I was twenty-one, I followed Shandong Taishan closely through a compressed fixture period. The club went five matches without a win and dropped from third to seventh.

The default media reaction is to find a name. The defence is playing badly, the full-backs are out of position, the centre-backs are slow. Social media filled with those judgements, and they all sounded plausible because they matched the goals everyone could see.

I did not believe it. I requested GPS data on distance covered and high-speed bursts for the whole squad across those five matches, compared with the previous five. The answer was not in the defence.

Total distance covered fell only slightly, but high-speed bursts in central areas dropped sharply, while movement density within thirty metres of our own box rose. In other words, the midfield had stopped running to cut passing lanes from the second line, so the back four had to drop deeper and face more one-on-one situations. What the eye saw was defenders out of position. What actually happened was the midfield stopping work three seconds earlier.

Collapse does not come from a single conceded goal, but from hundreds of small details ignored. Three seconds of lost focus, multiplied by fifteen per match, multiplied by five matches, produces a conclusion nobody wants to hear: the problem was not defensive technique but accountability in midfield.

Watching training sessions made the picture clearer. A young midfielder lost focus after an internal disciplinary sanction, and that loss of focus did not show up as an obvious error but as half a step late in every duel. In a stadium with fans, that half-step is drowned out by noise. In an empty stadium, it shows like a scratch on glass.

At the same time, goalkeeper Wang Dalei showed signs of a shoulder problem but kept it hidden. He did not tell the club doctor, did not tell the assistant coach. He trained normally, played normally, and differed only in a detail visible to someone sitting close: long distributions toward the right flank became noticeably rarer than usual, shifted instead to the left. That data exists in no tracking system. It exists in a notebook and in a head trained to observe.

In my report to the coaching staff I drew firm conclusions and proposed solutions. Colleagues later called me rigid. Perhaps rightly. But the dressing room is where truth outlives any contract, and had I softened the conclusion to keep everyone comfortable, I would have betrayed the very reason I was allowed in.

Four places where fabrication lives

Back to the empty file. It is an extreme artefact, but the principle behind it is common. There are four areas where football hides its gaps in the finest packaging.

The first is match data. A statistics table can hold thirty metrics and still say nothing meaningful if the presenter cannot pick the two that answer the question being asked. I once received a four-page match analysis containing possession, passes, pass accuracy, tackles — and not one line on how the trailing team changed its build-up. That is data without an argument. Data without an argument is decoration.

The second is the transfer market. This is where misinformation breeds fastest, because every participant has a motive to circulate the version that suits them. Agents want leverage for a pay rise. Clubs want leverage to sell high. Intermediaries want to prove they are active. And news outlets want traffic.

In the winter of 2026, Chelsea signed Enzo Fernández from Benfica for a fee recorded at around 121 million euros, making him the club's most expensive signing at the time. In the same window, Mykhailo Mudryk arrived from Shakhtar Donetsk for roughly 70 million euros plus add-ons, on an eight-and-a-half-year contract. In the summer of 2026, Moisés Caicedo moved from Brighton for 115 million pounds, an intra-Premier League record at that moment.

Those three numbers travelled everywhere. Almost nobody explained the contract structures behind them. Eight- and nine-year deals allowed clubs to amortise transfer fees over a longer period, reducing the annual book cost. It was a legal accounting advantage at the time, letting clubs spend more while keeping compliance ratios within limits.

In 2026, European football's governing body closed the loophole, capping amortisation at five years regardless of contract length. A technical change, three lines long, upended the squad-building strategies of several major clubs for years afterwards.

This is why I always verify at least three sources before writing anything about transfers. One headline, one social media quote and one fan comment do not make three sources. Three sources are three independent parties with different interests, and if they align on the facts while differing on interpretation, you have something to stand on.

The third is financial compliance. On 17 November 2026, Everton were docked ten points for breaching the Premier League's profit and sustainability rules. On 26 February 2026, on appeal, the deduction was reduced to six. On 18 March 2026, Nottingham Forest were docked four points. On 6 February 2026, Manchester City were charged with more than one hundred and fifteen breaches of financial regulations, and the process remains unresolved.

Those milestones are public data. But most content produced around them falls into one of two extremes: convicting before a verdict, or defending with arguments grounded in no document at all. Both are forms of the blank page: the writer has no material, so they fill it with emotion.

The fourth is post-match tactical analysis. This is where I see the most rubbish. After every major round, hundreds of diagrams get redrawn, with arrows indicating player movement in situations where the player had no intention of moving that way. The writer has not rewatched the full tape, only the highlights, and then constructs a story that sounds very reasonable.

The data analyst walks into the dressing room

Over the past decade, a new profession has entered professional football: the data analyst. At leading clubs they attend every session, every meeting and, increasingly, the dressing room itself.

This is progress. It also carries a specific risk.

Models are built to remove emotion from decisions. Football is played by people, and people run on emotion. A model says player X should start because his metrics beat player Y. A coach sitting in the dressing room knows player X has just been through a family crisis, and that his first touch in this morning's session was four metres off.

No model scores those four metres.

The data analyst sees the map. The coach sees the terrain. And football is only ever played on terrain.

During the Shandong season I witnessed a forty-minute argument in a meeting room between one side presenting passing-efficiency metrics and an assistant coach who kept asking a single question: "But will he take the ball in that area in the eightieth minute?" That is not a technical question. It is a question about human nature. And it has no answer in a data table.

I am not against data. My entire career rests on it. A master's in exercise science taught me the body moves according to measurable laws, and ignoring them is arrogance. But I am against using data as a shield to avoid responsibility for one's own judgement.

When you say "the numbers show", you are lying in a very sophisticated way. Numbers show nothing. You show. The numbers are simply there.

The miracle trap

There is a bias in sports media I call the underdog bias.

Stories of weak teams beating strong ones always spread faster than stories of strong teams doing their job properly. A ninetieth-minute winner by a bottom club generates more shares than a perfectly controlled 3-0. Algorithms amplify this, and writers unconsciously adjust.

The result is a distorted picture of football. Weak teams appear in media mainly at their moment of glory, and vanish across the other thirty-eight rounds. Nobody writes about a nineteenth-placed club's twenty-second weekend, about losing their first-choice centre-back to suspension, about a chairman three months behind on wages, about a coach fielding four eighteen-year-old academy players just to fill the bench.

Only by following a weak team all year do you understand the price of a miracle. Miracles are not free. They are paid for in weeks nobody watches, in winter sessions with no heating money, in twenty-year-olds forced to grow up ten years early.

Underdog bias also seeps into how we analyse. When a big club wins, we look for reasons they got lucky. When a small club wins, we call it character. Two different yardsticks for the same event. That is not analysis. That is storytelling, and storytelling can be beautiful, but it carries no obligation to truth.

The honesty of a blank page

Here I have to return to that eleven-page file.

On the third reading, I realised it had done one thing very few pieces of football content manage: it refused to draw conclusions without a basis. Every "N/A — insufficient information" cell was, in effect, a confession. Facing pressure to produce nine analytical sections, the writer left them blank rather than filling them with speculation, appending a warning that anything going beyond the integrity check would be fabrication.

I have spent years criticising hollow analyses. But I realised I had often criticised the wrong thing. The problem is not the existence of empty sections. The problem is empty sections presented as though they were full. A document that admits it is empty, like this one, is more useful than a document pretending to be full — by a ratio of roughly one to a thousand.

In the content industry, the most honest moment is usually the only moment the writer dares to say: I do not know.

This has practical value for football readers. When you read an analysis, look for whether the author states their sources. Look for whether they acknowledge the limits of their conclusion. Look for whether any sentence begins with an admission rather than an assertion. If there is nothing like that, you are probably reading a product filled in by form.

What outsiders cannot see

The counter-intuitive angle is this.

Fans often believe the quality of football analysis depends on the volume of data. More metrics, more tables, more diagrams means more credibility. That belief is right at the collection stage and completely wrong at the conclusion stage. A table with fifty metrics does not help you understand a match more than a table with five, if the presenter cannot pick the ones that describe what is happening.

In practice, an excess of data is usually a sign of insufficient understanding. Someone who understands needs two figures. Someone who does not needs all of them, because they do not know which matter.

Another outsider misconception: the public believes the dressing room is where emotion rules, where everything is settled by heated argument. My experience embedded with a club showed the opposite. The dressing room is the strictest information-discipline environment I have ever entered. What is said there usually never leaves. And what truly matters is often never said — it is expressed through a player arriving twenty minutes early, or through another staying on the training pitch half an hour after everyone has gone in.

This is also why data cannot replace observation. A tracking system records that a player stayed out thirty-two extra minutes. It does not record that he stayed because he was substituted in the sixtieth minute yesterday and needed to prove something to himself.

And this is why I keep one rule across my career: never write a tactical claim before verifying at least three sources and rewatching the full tape. That rule was born from a failure at seventeen, in Saint Petersburg, when I spoke very loudly about something I did not know.

The pulse waits for no one

There is a temptation I understand very well: the temptation to fill gaps with process.

When a crisis hits at work — a power cut, missing staff, corrupted data, a deadline arriving before the piece has taken shape — my instinct is to turn everything into a task list. Identify the core information. Sort the data. Delegate. Cross-check. That process has saved me many times, and I still publicly champion it.

But I must distinguish process from duty. Process tells me what to do next. It does not tell me whether I have the standing to say anything at all. That duty belongs to another level, which cannot be automated: I must answer for every sentence I set down, even when every step of the process was completed correctly on time.

I was once taken to hospital during half-time of a Euro quarter-final, and the piece was still finished twelve minutes after the final whistle. The pulse of a match waits for no one. But precisely because that pulse does not wait, every word must be written faster, not more carelessly.

The hospital could not slow the match down; it only taught me to run faster with every sentence.

The beat keeper

There is one role in this industry that I believe is the most underrated: the person who follows the team.

In English they call it a beat keeper. They do not write about matches from the stands. They travel with the team, fly with the team, eat with the team, sleep in the same hotels, stand on the training pitch at seven in the morning when it is still dark. They know what coffee the coach drinks, who sits next to whom on the bus, who is first out of the dressing room after a defeat.

Their information appears in no statistics table. But it is verifiable in a different way: it is verified by the consistency of small details, recorded over weeks and months, which, assembled, form a pattern nobody intended to create.

Truth forms in the dressing room, where people do not have time to hide what they are. Media can lie. Contracts can lie. Statistics can be cut and pasted to say the opposite of what they originally meant. Only what forms in that space, when nobody has time to prepare, is allowed to stand as a final conclusion.

And that is why I still read the empty cells in that file the way I read a reminder. A blank page is not a failure. It is the initial state. Failure is filling it with what we want to believe.

The internal signals to watch

If you want to know whether the club you follow is rising or falling over the next six months, do not start with the league table. The table is the result, not the cause.

Start with the three smallest signals. One: how many players stay out after official training ends. Two: whether substitutes celebrate the team's goals or do not. Three: who speaks after a defeat — when the speaker is the oldest player, the club is fine; when it is the coach and nobody else steps forward, the club has a problem.

No model scores those three signals. They never appear in media, never in statistics tables, and never in a nine-section analytical document. But they are the truest data I have collected in nine years in this trade.

As for the question I carry every morning, opening my laptop while the city has not yet switched on all its lights: if I have nothing to say today, do I have the courage to say that I have nothing to say?