Monaco on the Spreadsheet and Monaco on the Grass — When Football Data Deceives Itself

On the data board of an analysis room in Lyon I visited last October, the first result of a query tagged "Monaco" glowed red. Not because the Stade Louis II had just suffered another Ligue 1 defeat — but because the algorithm had attached that label to a document about a European filming schedule. The data analyst sitting next to me gave a wry smile: "The system got it wrong again." But having smiled, he left the row exactly where it was in the file sent up to the coaching staff, and nobody in the meeting checked it.

Three weeks later, in a meeting about the potential transfer target list, the name "Greece" surfaced again as a market to track closely. Then it was Gabriel — Arsenal's Brazilian centre-back — mixed up with an entirely different Gabriel, simply because the entity-recognition system could not tell the two contexts apart. And nobody asked a question.

I tell this story not to sneer at a piece of software. I tell it because I have seen the exact same thing happen on grass, every week, in analysis sessions where data is used as a substitute for the eye.

Modern European football lives on data. Every club from the Premier League down to Ligue 2 — and the clubs of Vietnam's V.League as well — has built its own analysis department. Recruitment driven by metrics, training driven by GPS tracking, tactics designed from chance-conversion probabilities. It sounds reasonable, but there is a nameless hole in it: data cannot distinguish "Monaco" the club from "Monaco" a filming location, "Greece" the national team from a base on a pre-season tour, and "Gabriel" the Arsenal centre-back from any one of several other Gabriels playing in Europe.

I have worked this trade since 2026, when the paper I started at was newly founded. Thirty-one years of observing is enough to draw one rule: every new technology in football is greeted as a revolution, and three years later people discover it is merely faster than the old method — not more accurate, unless the user understands it. The computer analyses data the way computers do best: counting, cross-referencing, probability. It cannot look at a player wiping sweat in front of goal and understand that his hands are shaking.

This is my point to every club staking its budget on data: without a verifying eye, data only makes mistakes faster and more confident. And the "Monaco" incident in that Lyon analysis room is not a software bug — it is a mirror of a problem sunk deep into the culture of contemporary football analysis.

Let us start with Monaco, because it is the example I know best, and the place where the crowd mocked me for three months.

In 2026-17, AS Monaco scored 107 goals in Ligue 1, won France and reached the Champions League semi-finals. That squad had an eighteen-year-old Kylian Mbappé, Radamel Falcao, Bernardo Silva, Thomas Lemar, Tiemoué Bakayoko, Fabinho, Benjamin Mendy, Djibril Sidibé. Leonardo Jardim's attacking machine made all of Europe tremble, and on the spreadsheet every number was beautiful: high goals, high xG, superior chance-conversion rate.

At the end of 2026, aged thirty-eight, I wrote my first piece on a newly created personal page, headlined: "Monaco will collapse after selling Mbappé." I cited the previous season's numbers, but what I actually relied on was not on the spreadsheet — it was on tape. I re-watched more than twenty Monaco matches and spotted a detail the metrics overlooked: their ability to build from the back depended heavily on Mbappé stretching opposing defences with raw speed, opening space for the satellites around him. When Mbappé left, that structure would collapse.

My contrarian call was mocked for three months. But when Mbappé officially moved to Paris Saint-Germain for a fee of 180 million euros — then the second most expensive transfer in the world, behind Neymar — the piece was shared more than fifty thousand times. Monaco went into the next season second in Ligue 1, but one season later slid to mid-table, then fought relegation, and kept selling off what remained of its pillars.

The lesson was not "data is right" or "the eye is right" — it was that data has value only when set against a concrete pitch-side reality, and the eye has value only when checked by numbers. Monaco then was a beautiful technical machine, but the data about them was "contaminated" by the system's own strength: the numbers were all clean, all smooth, so that when a specific human being departed, they became a self-fulfilling prophecy. No different from an algorithm labelling a filming article "Monaco" — the name was right, the context was entirely wrong.

And this is the point people miss: Monaco did not lose the audience, they lost the shield that covered their weakness. With Mbappé, you could hide a slow defence. Without him, every gap opened at once. Data never saw the shield — it only saw the goal tally staying high until, suddenly, it was not. Do not look at the number on the price tag; look at the team after the player leaves.

Now consider Greece, the Greek national team. At EURO 2026 they won the title with an iron defensive game under the German coach Otto Rehhagel — one of the great shocks in European football history. They beat Portugal twice, knocked out Zinédine Zidane's France in the quarter-finals, defeated the Czech Republic in the semi-finals, then beat Portugal in the Lisbon final. That was a disciplined, quick-transition side built on organisation and collective spirit rather than stars.

Twenty years later, data models still routinely file "Greek defence" as a stylistic label — while the current national team is entirely transformed, with a young generation playing possession football and fast transitions. They are no longer Rehhagel's team. But the "Greek defence" label survives, and it makes anyone who reads the label without watching the game predict wrongly.

This is the most dangerous kind of analytical error: a stylistic label outliving the very team that produced it. People see the "Greek defence" tag and misjudge the match, because the algorithm does not know that today's line-up is not the line-up of twenty years ago. Just as an algorithm labels a filming article "Monaco" — the ambiguity is not in the name, but in the failure to check context.

I remember a Greece qualifier where the co-commentator beside me kept talking about the "Greek defensive tradition" while the team was playing a back three and pushing high. By the twentieth minute Greece had taken the lead and kept attacking. The old label was in the commentator's head, not in the eleven on the pitch. And the viewer at home, listening without watching, carries the wrong label into the next match.

Then Gabriel — the most interesting case, because it exposes the mechanism of "entity collision" in its bare form.

At Arsenal there is Gabriel Magalhães, the Brazilian centre-back, rated among the best defenders in the Premier League in recent seasons. At Brazil and formerly at Arsenal there is Gabriel Jesus, also Brazilian, a forward. And elsewhere there is Gabriel Martinelli — yet another Gabriel, also Brazilian, also at Arsenal. Three men with the same given name, the same nationality, the same club. It is the nightmare of every entity-recognition system.

Nor does the problem stop there. Across Europe, in the Premier League alone this season there are at least three Brazilian players named Gabriel. Add the other Gabriels in Portugal, Spain and Germany, and the number runs into the dozens. Entity-recognition systems built on names will conflate these men into one unless the context window is long enough. In football, the context is position, club and tactical role. When context is truncated — when people read only headlines, only one social-media line — machines and readers commit the same error: confusing a centre-back with a forward purely because of a shared name.

I witnessed this in a live television analysis in Marseille, when a guest pundit insisted Arsenal were "short of strikers" while the starting line-up plainly had Gabriel Jesus up front. What that guest got wrong was not football knowledge — it was the habit of fast lookup, skim-reading, name-assignment without context verification. A shock take has value only when it stands on a detail others overlooked — and a shock take built on a false detail earns one share and is forgotten.

And here is the larger point: matches are decided where the audience is not looking. In all three cases — Monaco, Greece, Gabriel — the common thread is that the data system saw part of the truth and discarded the rest. Monaco has a name and a club. Greece has a place and a national team. Gabriel has a name and several different men. The good analyst is not the one who reads more data, but the one who knows to ask: where did this data come from, what is its context, and what has been left outside the frame?

I have applied this principle throughout recent years. In 2026, at the Moscow World Cup as a commentator for an online sports channel, before the Croatia-England semi-final I wrote that Croatia would reach the final because Luka Modric and Ivan Rakitic controlled midfield. I watched them circulate the ball, stretch England's shape, and predicted a 2-1 win after extra time — exactly as it finished. The piece reached two million views and I was invited onto national television. But what I remember most is not the two million — it is the feeling in the stands when I realised: Croatia's midfield was not strong because they completed many passes, but because they moved the ball at the right rhythm once England's midfield lost its structure. The data gave me the pass count; the eye gave me the rhythm. Numbers walk me to the stadium gate, the eyes lead me into the dressing room.

Since then I have shifted fully to situational tactical analysis — using concrete passages with minutes, names and positions to illustrate a point, rather than speaking in labels. Not because I disdain data, but because I learned that data is a map, not the territory. The map tells you where Monaco is; it does not tell you whether the principality's club is pressing high or sitting deep in the seventieth minute.

Nor is this confined to Europe. Looking at Vietnam's V.League, I see the same signals emerging. Clubs are hiring data analysts, buying foreign software, importing models from Europe — but many lack enough people who understand data to verify it with their eyes. A striker scoring heavily in a lower division can be promoted to the first team on a pretty metric, then fail because the metric did not account for a completely different system. A coach can be judged by win percentage on a dashboard while nobody re-watches how his side operates in the last fifteen minutes.

I say this as a man of thirty-one years in the trade who has won the Football Writers' Association Commentator of the Year award around five times, most recently in 2026. Those awards did not come from reading more data than others — they came from going to the stadium, sitting in the stand, and seeing what the spreadsheet never records.

Where could I be wrong? One point I must be honest about: I have fallen into the trap I am criticising. Not the data trap, but the opposite — trusting my own eye too much.

Last March I watched a match involving a club I follow closely and became certain of an emotional conclusion: that their defence had lost its structure. I had watched the first fifteen minutes, seen two misplaced passes, and immediately wrote a paragraph on my personal page. When I checked the numbers, the back line had conceded only two clear openings all match, both in stoppage time while protecting a lead by dropping deep. My eye was right about what it saw, wrong about what it concluded. Had I not checked the numbers, I would have published a wrong verdict.

That is why I always say: do not make an idol of the eye, and do not make a dogma of the number. Truth lives where the two meet, and the good analyst is patient enough to stand at that intersection longer than anyone else. What I oppose is not data, but the habit of using data as a shield to avoid watching football.

And one more honesty: my arrogance. I am known for shock headlines, and I know my reputation was largely built on the times I was mocked and later proved right. But that reputation has a price: it makes me stubborn at times, holding a prediction even as reality refutes it. People laughed at me for three months, but laughter never scores — and that cuts both ways. Laughter does not score when I am right, and it does not score when I am wrong. The only thing that scores is the reality on the pitch.

At Monaco this season there is a young midfielder I have watched for three consecutive rounds. He has no standout metrics: not many goals, not many assists, a pass-completion rate merely decent. But sitting in the Stade Louis II I saw him constantly directing the midfield, adjusting the team's distances, making off-ball runs to open space for teammates. The data calls him an average player. The eye calls him the orchestrator. If that club sells him in January because his metrics are not pretty, that will be a data mistake — and I will write about it, as I wrote about Monaco in 2026.

But I also do not forget that not long ago another club kept a player because I trusted my eye, and that player failed completely over the following six months. The data said otherwise then, and the data was right. I have no hesitation in saying it: I was wrong. For a man known for shock verdicts, admitting error is not a loss of credibility — it is the only way to keep the eye honest.

If you ask me for a prediction this season, I will answer with a question back: which data are you reading, and have you watched the match? Modern football is paying for cognitive laziness — we have too many numbers and too few stories told from the ground. The club that understands this first will win, not because it has better data, but because it knows data is the beginning of a question, not the answer.

As for Monaco, Greece and the name Gabriel — I will keep watching football with both eyes, and keep a finger on the keyboard to double-check whenever I want to rage at a number. Because matches are decided where the audience is not looking, and a sportswriter has value only when standing in exactly that place.