Football Analysis in the Data Era: A Complete Framework Can Still Be Empty

Football Analysis in the Data Era: A Complete Framework Can Still Be Empty

November in Incheon. Steam clung to the newsroom window, and on my screen sat a nine-page document. It had a title, a subtitle, a table of contents, tables, a methodology section, a conclusion, and even a carefully written paragraph headed "limitations of the study." It was missing exactly one thing: content.

I sat in front of it longer than necessary. It resembled a stadium already built — freshly painted seats, floodlights installed, the tunnel to the pitch numbered step by step — but no match had ever been played on that grass. Not one shot. Not one shout. Nobody could remember the name of a single team.

Forty-seven years in this trade have taught me many things, but the most expensive lesson is this: a document that looks complete can be more dangerous than an empty one, because nobody bothers to check it. When a blank page lies on the desk, you know you have to write. When a skeleton lies on the desk, you assume the writing is done.

I am Jack Lee, born in Brazil, living and working in Incheon. I came into commentary from esports, then moved to football reporting for the Korean market. The two worlds look different, but they share an identical disease: the belief that a correct analytical framework will automatically produce a correct conclusion.

When the data table becomes ritual

Over the past fifteen years, football analysis has seen an explosion of metrics that once existed only on data companies' servers. xG — expected goals — measures the quality of a chance rather than its outcome. xA — expected assists — assigns value to the decisive pass based on the probability it creates. PPDA — passes allowed per defensive action — turns pressing intensity from a feeling into a quantity comparable across teams.

That is real progress. I have seen it rescue hundreds of articles from laziness. But I have also seen it breed a newer, subtler laziness.

Based on my experience watching matches, a modern analysis tends to open by listing metrics and close by turning those metrics into an emotional story. The middle — the hardest part, the part that demands the writer answer why the metric looks that way — is usually skipped. The data table becomes an opening ritual, like a player touching the grass before kickoff: everyone does it, nobody remembers why.

The real worry lies elsewhere. Metrics do not create meaning by themselves. A team with a PPDA of 7.5 is performing defensive actions at a very high rate. But that number does not tell you whether they press so hard because they are proactive and good, or because they are behind and desperately chasing the ball. The same quantity, two entirely different stories. A reader given only the quantity will never tell them apart.

That is the first crack.

Club finance: where data became law

If there is one field where data stopped being decoration, it is club finance. UEFA's Financial Fair Play was introduced in 2026 and took effect from the 2026-2026 season, obliging clubs in European competition to move toward break-even. In England, the Profit and Sustainability Rules gradually replaced older regulations, capping permissible losses over a three-year cycle.

Those quantities carry real weight. On 17 November 2026, Everton were deducted 10 points for breaching the loss threshold. On 26 February 2026, after appeal, the deduction was reduced to 6 points. On 18 March 2026, Nottingham Forest received a 4-point deduction. In Italy, Juventus were deducted 15 points in January 2026 in a case concerning the valuation of players in transfer transactions; the sanction was overturned and replaced with a 10-point deduction in May 2026. And since 6 February 2026, Manchester City have faced 115 charges of breaching the Premier League's financial rules, a file stretching back more than a decade.

This is where data discipline stops being a matter of style. When sanctions are counted in table points, an analytical error does not merely make an article wrong. It makes supporters misunderstand the future of the club they love, and skews the arguments on the terraces for years.

I understood this on a winter night in 2026, sitting in Beijing and writing about a final that lasted only 78 minutes. I wrote three thousand two hundred words, used the word "legend" fourteen times, and likened the tears on stage to rain falling on an old tower. My editor — a former statistician — underlined twelve passages. He said the winning side averaged 98 wards per game, while my article did not contain a single quantity.

The piece still spread. But the analysts laughed. And I understood that emotion is not at fault. Emotion standing alone is at fault.

Transfers: where money meets bleeding dreams

The transfer window is not a village market; it is the collective testament of dreams.

I wrote that line in 2026 and still believe it. But a testament needs a notary. In modern football, the notary is data.

The Transfermarkt database, founded in 2026, has become the neutral yardstick almost every newsroom uses to anchor player value. The problem is that it is an estimated yardstick, updated periodically by community users and an editorial team, not an actual transaction price. When an article says Club A paid "double the Transfermarkt valuation," readers easily take it as proof of irrationality. In reality it is a comparison between two quantities of different natures: an estimated market value and a signed contract value, which depends on remaining contract length, player age, the seller's need, and the buyer's desperation.

The consequence is what I call the panic premium. With seventy-two hours left in the window, a club in a squad crisis will pay a price it would have rejected outright six months earlier. That gap is not data error. It is data — data about how frightened the board is.

A decent transfer analysis must separate three layers: the nominal fee, the payment structure across years, and performance-related add-ons. Many reports carry only the first layer, because it is tidy and shocking. The other two decide whether the deal strangles the wage bill three years later.

And the wage bill is what kills clubs. Not the transfer fee.

Results cycles and the small-sample trap

One of the most common errors in football commentary is concluding from three matches. Three straight wins generate a story.

The same three matches, if they happen in October, produce talk of a title contender. If they happen in May, they produce talk of a team with nothing left to play for, playing freely. The content of the matches does not change. The meaning changes completely, because timing is part of the data.

This is why I force myself to state absolute dates for every form-based claim. No phrases like "this week" or "lately." A claim without a date is a claim that cannot be verified, and a claim that cannot be verified has no cumulative value.

The phenomenon commonly called the "new-manager bounce" is a fine example of where data and intuition agree on the outcome but disagree on the cause. The outcome is real: teams usually perform better in the first few games after a managerial change. The cause is contested. Some argue it is a psychological release effect. Some argue it is regression to the mean — the team was never that bad, and the preceding losing run was simply an unusual small sample. Some argue the first few fixtures after a change are often easier.

Three hypotheses, three entirely different implications for the future. If it is psychology, form persists. If it is regression, the team returns to its old level and you bought the top. If it is fixtures, the good run ends the moment they meet a strong side.

An article that gives only the outcome without distinguishing these three hypotheses is an unfinished article. It has numbers. It has structure. It is still empty.

The dressing room, the age curve, and things that cannot be measured

There is a paradox I have met throughout forty-seven years: the factors that decide a team's success are often the hardest to measure.

Dressing-room leadership structure is one example. Who speaks last when the team is two goals down? Who stays calm when a young defender makes a mistake? No metric captures that, which is why teams with beautiful data sheets can still collapse within two weeks.

The age curve is measurable but frequently misused. A player at thirty does not automatically decline. Decline depends on position, on accumulated career minutes, on injury history, and on whether the club adjusts its style to compensate. A playmaker at thirty-two dropping deeper can still be the most important player on the pitch, while a pace-dependent winger can lose value after a single hamstring injury.

I learned this from an old mentor in Madrid, who always said age is not in the passport but in the footage. He told me to watch the first thirty minutes and the last thirty minutes of the same player across ten different matches, and draw my own conclusion. I did. My conclusion was this: many players labelled "finished" are simply being played in the wrong position.

In 2026, when stadiums stood empty because of the pandemic, I analysed forty matches of a championship-winning side and found their win rate rose by roughly twenty-three percent in games where the support left the bottom lane before the eighth minute. I opened with this line: "An empire does not rise from thunder, but from half a second of a jungler's reaction." For the first time in my career, I joined poetry and statistics without having to choose a side.

But I also know the limits of that finding. Forty matches is a sample large enough to suggest, not to declare. Anyone who turns it into a fixed formula for every team has turned an observation into a dogma.

Industry transmission: from academy to broadcast deal

Football is a supply chain, and most analyses look at a single link.

At the head of the chain sit academies and scouting networks. In the middle sit clubs and competitions. At the tail sit broadcast rights, commercial revenue, and derivative markets.

A decision at the head can shake the tail, with a delay of several years. When a club decides to sell a young player to balance the books within the current accounting period, the consequence does not stop at losing a player. It affects squad value three years later, sponsorship appeal, negotiating leverage on broadcast contracts, and even the national team squad list.

In Korea, where I live, this chain is unusually easy to observe. Son Heung-min won the Premier League Golden Boot in 2026-2026 with 23 goals, sharing the award with Mohamed Salah. Kim Min-jae won Serie A in 2026-2026 with Napoli. Those two stories are not merely personal. They are the output of a development system and a player-export strategy run for more than two decades.

People often tell Son Heung-min's story as an individual legend. I prefer to tell it as an indicator of the quality of an entire football nation.

When a complete framework becomes a shield

Back to that nine-page document.

After checking, I found the cause. It was not writer laziness. It was a failure in input extraction: the original content could not be read, and the system retained only a single label — "football" — from which it generated the entire analytical template. The template was so complete that every cell contained text, except every cell read "insufficient information."

So an empty document was packaged as a finished one.

This is the greatest trap of the data era in football. Not the trap of missing data, but the trap of having enough form that nobody questions the content.

A report with nine sections, nine headings, complete tables, and a "risks and warnings" section creates a false sense of security. The reader skims, sees tight structure, sees technical terminology, and believes. The editor skims, sees every field filled, and passes it.

In football, the same thing happens daily. A scouting report has all twelve criteria, but all twelve were scored by one person who watched the player in two video clips. A metrics comparison table is complete, but drawn from two leagues with different opponent quality. A transfer forecast has every link in the deal, but the root link — the source — is an anonymous account.

The real risk is not that people say something wrong. The real risk is that they say the right form and the wrong substance, and nobody notices.

I have sat long enough in analysis rooms to know that the most frightening thing is not an obviously wrong conclusion. It is a conclusion presented too neatly to be challenged.

What must be done, and what must stop

On the writer's side, there are three principles I imposed on myself after the shock of 2026.

First, every claim must be anchored to at least one verifiable fact: a specific date, a specific score, a specific fee, a specific published source. If there is none, I write a different sentence.

Second, every quantity must carry the context that produced it. A team holding the ball a lot does not mean the team controls the match; it may only mean the opponent is deliberately conceding possession to counter-attack. Context is the hardest part and cannot be omitted.

Third, I must state my level of certainty. Distinguish between what happened, what can be inferred, and what is merely a guess. Those three levels must never be written in the same voice.

On the reader's side, one question should be asked far more often of any analysis: what does this tell me that I did not know, and on what basis?

If neither part has an answer, the piece is unfinished. Even if it runs nine pages. Even if the headings are all there. Even if it looks entirely professional.

A counter-intuitive angle: discipline can become a shield

There is a widespread belief in modern football analysis that I consider wrong.

It holds that if you have enough data, you are immune to bias. I have watched the opposite happen too many times.

Data is not immune to bias. Data gives bias a new outfit, more polite, harder to catch. Someone convinced that a player is finished will find three metrics to prove it. Someone convinced that a manager is a genius will find three matches to exonerate him. The process does not produce the conclusion. People produce the conclusion, then use the process to defend it.

This is why I no longer trust analyses presented too smoothly. The rough patches in a piece — where the writer admits uncertainty, where two hypotheses stand side by side, where the data is insufficient and the writer accepts it — are a more reliable signal than any handsome table.

Snow falling on the summit of glory resembles the truth: light, silent, and it whites out every legend.

I have seen gold in the snow, and I know the most precious metal does not sit on the podium. It is not in any metrics table. It sits in the moment a writer is brave enough to say the evidence he has does not support the conclusion he wants.

That is the rarest moment in this trade. And the most valuable.

What remains when the lights go out

An era does not die from a defeat; it dies when people stop telling its story.

The same holds for analysis. A method does not die because it is wrong. It dies when it is used in the right form but with empty substance, day after day, until nobody believes in it any more.

At sixty-three, I do not count trophies. I count the stories that remain when the lights go out.

An empty stadium taught me that the loudest applause is the applause inside the heart of someone who still believes.

And supporters keep believing, as long as the person holding the pen is still willing to retrace the road he walked.

For forty-seven years, verifiable evidence has been the hardest thing in any football analysis. No algorithm has replaced it, and I doubt one ever will. Our job, as writers, is to keep the next analysis — and the one after that — from ending as a complete framework hiding an empty space. Supporters deserve to know the difference.