Data Journal: Four Summers That Shaped How I Read Football

Hook

July 2026. World Cup final night on Russian soil. I was fifteen, sitting in a small apartment in Da Nang, watching an old television display the score France 4-2 Croatia. The match ended close to three in the morning. My father had dozed off on the sofa since the first half. I stayed awake — but not because of the thrill of a final.

It was because of a number the commentator read only twice all match: Luka Modric ran 12.7 km in 120 minutes. Harry Kane ran 11.9 km. But Kane touched the ball fewer than thirty times. A striker running nearly as much as a creative midfielder, yet almost nonexistent in the biggest match of his career. Something did not match between the number and the feeling the stands were cheering.

That night, I reread English-language analyses of expected goals — a concept then unfamiliar to most Vietnamese audiences. I found something absent from the post-match reports: Croatia won only three of six knockout matches, yet their xG exceeded their opponents' in all six. The press called Croatia "deserving." The data showed Croatia created more chances. Two different stories, same tournament. In one night, I learned to hear a number whisper amid the roar. Amid the cheers of Russia, I heard a number whisper — and it was more accurate than the crowd.

Context

That was the beginning. Seven years later, I still sit at the same desk, the same laptop, the same habit: before every major tournament or transfer window, I rebuild an analytical framework of my own. But the journey from that Russian final to today is not a straight line. It passes through four summers, each leaving a number that forced me to rewrite how I read football.

The current context is the transfer window. Names swapping shirts, transfer fees shouted in every headline, fan pages racing to post rumors. But amid that noise, I keep returning to those four summers — because each taught me one principle: how to read a contract, a squad, a match is really just different versions of the same equation. Separate signal from noise.

I was not born into a football family. No father who coached, no brother training at an academy. I came to sports data via a detour: from personal curiosity to profession. From a tenth-grader in Da Nang unable to sleep over xG, to a sports betting analyst at twenty-three, living in a coastal city where I can follow both the Premier League and V-League in the same week. Da Nang gave me a strange advantage: geographical distance from Europe — where the data originates — forced me to read rather than believe. I could not watch football in the stands, could not be swayed by the crowd. I was forced to hear the number first.

But hearing the number first does not mean trusting only the number. That is a distinction I had to learn over seven years — and this article is part of that process. I retell four summers not to boast about results, but to show that every correct conclusion carries a blind spot with it, and that the best data practitioner is the one who knows where their own blind spot is.

Core

First summer, 2026 — Russia, and how xG broke a prejudice.

The final at Luzhniki taught me the first lesson: feeling and statistics do not negate each other, they simply answer two different questions. The stands roar when Croatia touches the ball. The data says Croatia shoots from better positions. Both are true. But if you need to predict the next match — not recount the last one — your question must be the question the data answers.

Across all six of Croatia's knockout matches at World Cup 2026, their xG exceeded their opponents' in every single one. In the semifinal against England, Croatia generated 1.7 xG against 0.6 for their opponents, while the 90 minutes ended level at 1-1. In the quarterfinal against hosts Russia, Croatia dominated in clear chances but had to go to penalties to advance. These are matches where, if you look only at the score, you think Croatia was lucky. Look at xG and you see they deserved it — and more importantly, you see a repeatable trend.

My first principle was born that summer: do not judge a team by the result, judge it by the chances it creates and the chances it allows its opponents to create. A result is one instance. A chance is a trend. I began rewriting my football-watching journal as a data journal. Each match, I recorded three lines: xG for both teams, passes into the box, and shots conceded inside the box. No model yet, just a recording habit. But that habit was the seed of everything that followed.

I do not watch football to enjoy it. I watch it to test a long-term hypothesis. I wrote that line first at fifteen, and it remains true today. But I must also admit: at fifteen, I did not know what I was doing. I only knew that when a match ended, the story commentators told did not match what I saw in the stats sheet. That mismatch became an obsession. And that obsession became a profession.

Second summer, 2026 — empty stadiums, and the most perfect laboratory.

In 2026, the pandemic forced European stadiums to close. When the Bundesliga returned in May with empty stands, the media called it a temporary fix. To me, it was the first opportunity in my life to observe football under controlled variables. Roughly speaking: if the crowd is a variable, removing it from the equation tells you how much it matters.

I was seventeen. I spent nearly a month gathering data from 312 matches across six top European leagues — matches with crowds before the pandemic and matches without crowds after football's return. The results forced me to rewrite everything I had believed about "home advantage."

Home win rate dropped from 46% to 38%. That is a steeper fall than any tactical change I had ever seen in my data. More striking: home teams' PPDA (passes allowed per defensive action) rose by an average of 1.8. Meaning without fans, home teams pressed less — not more. They lost what the stands gave them: a form of social pressure forcing them to surge.

That raised a question my data could not yet answer. Where does home advantage come from? The pitch, the referee, the travel fatigue of the away team — or the stands themselves? If it is that final factor, then the story of "home advantage" every fan page tells is missing its most important variable. And if I could isolate that variable, I could predict a match better than anyone.

An empty stadium is the most perfect laboratory I have ever entered. I wrote a three-thousand-word analysis of this effect and posted it on a small data forum. Days later, I received a message from a manager of a First Division club — asking if I had more detailed data. That was the first time I understood my data could step beyond the screen and influence a real decision.

I still keep that article. Not because it was perfect. But because it was the first time a question of mine had someone waiting for the answer. Since then, I began recording more systematically. One table per week. One trend per month. One hypothesis per season. But it took until the summer of 2026 for that habit to return an answer large enough for me to put two million dong on the table.

Third summer, 2026 — PPDA, Morocco, and my first big bet.

In 2026, I was nineteen, a first-year university student. Before the World Cup in Qatar, I built a ranking model for all thirty-two teams based on three years of defensive data: PPDA, distance covered, and shots conceded inside the box. This was not a complex model. It was how I forced myself to stay faithful to a framework rather than let emotion lead.

PPDA is the lens — through it, I saw Morocco in the semifinal two months early. The model placed Morocco in the top eight. Nobody believed it. I shared the prediction with my classmates. They laughed. I shared it with an older brother working in the betting industry. He called me "naive." I kept the prediction unchanged and recorded my reasoning in the journal — because I wanted to hold myself accountable for every prediction I made.

Morocco's first group match was against Croatia — the 2026 World Cup runner-up. Their second was against Belgium, then ranked second in the world. Belgium had Kevin De Bruyne, Romelu Lukaku, Eden Hazard. Morocco had a defense nobody in the international media's prediction lists named. Croatia drew Morocco 0-0. That was the first sign I needed.

I placed two million dong on Morocco to beat Belgium, at odds of 5.80. It was the largest amount I had ever staked at that point — and the first time I bet not on a feeling about a team, but on a number I had calculated myself. My first big bet did not come from bravery. It came from the crowd's mistake.

Morocco won 2-0. Belgium took nine shots, but their xG was only 0.7. Morocco defended with a low block, their PPDA among the lowest in the tournament, yet their shots conceded inside the box were very low. That is the paradox only defensive data sees: a team pressing less does not mean defending badly. They chose to cede space in midfield to protect the box. That was a deliberate trade-off, not passivity.

Morocco reached the semifinal. They eliminated Spain in the round of 16 and Portugal in the quarterfinal. They became the first African team to reach a World Cup semifinal. I won the bet, but the money was the smaller part. The larger part was conviction: defensive data can predict results with higher accuracy than the intuition of an entire media industry. From there, I began building my own betting analysis system — and logging every stake with the reason for win or loss. Not to show off. But to force myself to follow the framework rather than emotion.

Russia taught me that the crowd and the data always tell two different stories. Qatar taught me that the data's story can be written before the crowd's — if you are patient enough to wait for it to come true. But both summers taught me something more dangerous: after a streak of being right, people begin to believe they can see the future. And that belief is the beginning of every mistake.

Fourth summer, 2026 — Euro, Yamal, and an offer from Malta.

In June 2026, I was twenty-one, interning at a small sports data company in Ho Chi Minh City. Spain unleashed a teenage wing pair: Lamine Yamal, sixteen, and Nico Williams, twenty-one. The media called it a "young, exciting wing partnership." I wanted to know exactly what they produced.

I spent three weeks gathering data on receiving positions, movement directions, and chances generated. The result: the Yamal – Nico Williams pair generated 4.2 xG per match from central runs — higher than any other midfield pair at Euro 2026. Yamal received the ball an average of 11.3 times per match in the zone where opponents pushed high, and those receptions not only created chances for him, but opened space for full-back Dani Carvajal to overlap.

This matters not because it is new. But because it breaks an old prejudice: that a young team must play counter-attacking football to be effective. Spain did not play that way. They pressed high, kept the ball, and used both wings like two scalpels into the opponent's central space. Their style raises a question about the future of tactics: does the model of "wings as two independent pipelines" still hold in modern football?

I wrote a twelve-page report on Spain's left- and right-wing ecosystem — how they do not use the wings as two independent pipes, but as two branches of a central network. My boss sent the report to three European betting firms. A week later, a company in Malta sent a part-time job offer.

I accepted. But I kept my studies. Because I believe a system built slowly lasts longer than one built to beat a deadline. One report correct at one tournament is not enough to prove a career. But it is enough to prove a direction.

Spain won Euro 2026. In the final against England, Nico Williams scored the opening goal after a move originating on the left wing — exactly as my model had drawn. But what I remember is not the goal. It is the moment in the seventeenth minute, when Yamal received the ball at the edge of the box, and I knew — from three weeks of data — that he would not shoot. He would lay it back to the second line. And he did exactly that.

That was the first time I felt the strange sensation of an analyst: watching a match I had already seen in the data. Not prophecy. Just a trend clear enough, repeated often enough, that you recognize it when it happens again. And my job, ultimately, is to make that recognition slightly faster than the crowd's.

Contrarian

But confidence is a kind of poison. And I must say this before anyone reads this article and believes data is always right.

Over the past four summers, I have been right more often than wrong against the crowd. But being right often does not mean being right for the reason you think. My PPDA model placed Morocco in the top eight — but if Kevin De Bruyne had not missed two clear chances against Morocco, the story would be different. Belgium had a higher xG in that match if blocked situations are counted. The final result rewrites everything, including what the data had not yet reflected.

The drop from 46% to 38% in home win rate in 2026 may have been affected by congested schedules, injuries, and fitness — variables my data could not isolate. Some matches were played in sweltering summer conditions or on a three-day schedule — factors that could strip home advantage even with fans present. I did not control those variables. I only know I observed a real phenomenon but may have attributed it to the wrong cause.

At Euro 2026, Spain won, but if Rodri had been injured in the semifinal, their wing pair might not have had enough ball to flourish. If an opposing defender had marked better in the final, Nico Williams might not have had the space to score the opener. Every correct model depends on a chain of "ifs" the analyst does not control.

That is the core problem of every sports model: correlation is not causation, and data cannot predict random events. A model that is right three times does not prove the model is right. It only proves the model has not failed three times. That is a small distinction in logic, but a huge one in consequence. Many confident analysts — and many fans — fall at exactly this point.

The worst thing a data analyst can do is confuse their streak of being right with a complete theory. I nearly made that mistake. After Qatar 2026, I began to believe defensive data was king. I nearly ignored attacking data through the first half of 2026. Then I was wrong — repeatedly — in the Euro 2026 qualifiers, when teams with strong attacks but average defenses still won through individual brilliance. Football is never one-dimensional. Every one-dimensional model collapses when it meets reality.

The true discipline of a Data Monk is not believing in your model. It is actively seeking out data that refutes it. In every analysis I write now, I dedicate at least one paragraph to what could make my model wrong. If I cannot think of anything — it means I have not read enough. It is a rule I imposed on myself after the 2026 lesson, and it has saved me many times from arrogance.

In football, the only thing trustworthy is what the crowd has not yet seen. But to see it, you must accept that you too might be looking wrong. That acceptance does not make me weaker. It makes me more careful, and in this line of work, carefulness is a form of strength.

When the stadium was empty, I realized I had been betting on a myth for four years. Not a myth about a team. But a myth about myself — that I could always be right, as long as I had enough data. The truth is no amount of data is enough to erase randomness. And the good data practitioner is the one who lives with that, not the one who tries to prove otherwise.

Takeaway

The current transfer window is in its noisiest phase. Rumors about fees, inflated numbers, fan pages racing to post "insider sources." Amid that noise, four summers have taught me one thing: do not read rumors — read structure. The structure of a contract (release clauses, duration, wage bill) tells the real story better than any headline. The structure of a squad (positional gaps, average age, fixture density) reveals what a team needs better than any rumor.

If you ask me which signal is worth tracking next month, I will not answer with a name. I will answer with a question: which team is changing its structure, not just its personnel? Which team is filling a tactical gap, not just buying a star? Which team is building a system — rather than chasing a headline?

Four summers, four numbers, four lessons. But the journey is not over. It has only begun in a new transfer window — where the numbers are whispering again, and I am listening again. If there is one thing I want to carry from Modric, from the empty stadiums, from Morocco, and from Yamal — it is patience. Not patience to wait for data to answer. But patience to accept that some questions data will never answer. And my job is to keep asking, even when the answer has not arrived.