The Breath That No Spreadsheet Holds: When Transfer Data Misprices the Dressing Room

In March 2026, at PSG's Camp des Loges training centre, I stood behind the training fence and clicked my stopwatch. Neymar warmed up with a dribbling drill through six cones in an unchanging order, never off by a single beat from the day before. Over six consecutive months following that club, I recorded him repeating the drill 127 times in a single month, always finishing with a left-footed shot into the top right corner. When I asked why he never changed it, he just smiled: "It's a habit." In an office less than two kilometres from the training pitch, the club's data analysis department was running its player valuation models. They had xG, PPDA, successful dribble rates, distance covered per match. No column recorded "127 times", and no cell measured the order of six cones. That absence is not a minor technical flaw. It is a mirror reflecting how modern football values human beings. Over the past fifteen years, European football scouting has changed its skin. The notebooks of old scouts, filled by eye and memory, have gradually given way to vast databases. Big clubs hire dozens of data scientists, buy subscriptions from analytics firms, and build their own algorithms to rank players across hundreds of metrics. A midfielder can be judged by progressive passes, escapes from pressing, expected assist value. A centre-back is measured by interceptions, aerial duels won, long-pass accuracy. That revolution brought good things. It helps uncover gems in minor leagues where the human eye struggles to reach. It makes the market more transparent, reducing deals driven by one person's gut feeling. But it also creates a quiet belief: that if a player does not look good enough in the spreadsheet, he is not worth the money. And that belief is making football misprice two things at once — young potential, and dressing-room chemistry. To see the scale of the change, just look at the numbers. If the world transfer record in 2026 — the year I began this profession — was roughly £46 million for Zinedine Zidane, then by 2026 PSG paid €222 million to bring Neymar from Barcelona, breaking the record by nearly five times. A decade after that deal, the market has grown used to transfers above €100 million for players not yet 22. That enormous flow of money runs through valuation models, and the more money there is, the more people need numbers to justify their decisions. Let me start with the example I followed most closely. At the 2026 World Cup, I was given a special assignment: to record the entire behaviour of France's substitute players across seven matches. Olivier Giroud scored only one goal, and in the press he was labelled "the striker who doesn't score". But my notebook recorded a different figure: 214 pressing actions and 38 ball recoveries in the final third — the highest in the team, even though he was not the player who ran the most. Those presses did not directly produce goals, so they were almost invisible in every attacking statistic. They only appear when you sit long enough to see who freed up space for Kylian Mbappé and Antoine Griezmann. That is the first blind spot of the data model. Attacking metrics capture the final outcome — goals, assists, xG — but usually ignore the chain of actions leading to that outcome. A run that drags an opposing centre-back out of position goes unrecorded. A screen that lets a teammate break away appears in no column. Football is a chain of cause and effect, but data only snapshots the endpoints. And when you value a player only by those endpoints, you are buying a fragment of a person. The second blind spot lies at the opposite end: how the model overvalues youth. In the transfer world, there is an almost default rule — the younger the player, the higher the price, because of "resale potential". A 19-year-old can be valued above a 28-year-old at the peak of his powers, simply because the value curve is assumed to be steeper. This sounds financially sound, but it ignores a variable that cannot be measured: the maturity of the dressing room. In seven years following teams in France, I learned that a champion team is not the sum of its best players, but a system of rituals and trust. Someone must know when to speak, and when to stay silent. Someone must take responsibility when the team loses three games in a row. At PSG in the 2026-18 season, I once saw a talented young player become isolated simply because he had not yet learned to stay quiet at the right moment in the dressing room. The club's spreadsheet gave him very high marks. But the team did not play better when he came on. No statistic recorded that, and no one on the coaching staff wanted to admit it in black and white. The third blind spot concerns time. Data measures a moment, but the dressing room operates across seasons. Whether a signing succeeds or fails usually only becomes clear after two or three years — once the player has grown used to the language, the teammates, the climate, and even the way a city treats him. Neymar's ritual, which I recorded in 2026, is a small example. Those 127 repetitions were not superstition. They were his way of anchoring himself to something unchanging, so that when everything around him shifted — coach, tactics, pressure — he still had a fixed point. No algorithm can price that fixed point. But it decides whether that player stands firm in a big match. My experience of following the game taught me a simple habit: come back to a player three years later, and compare what was promised with what was delivered. I once heard an expert say of a Ligue 1 youngster that he would be "the next Neymar". Many seasons later, he was still drifting in lower divisions. Not because he was poor — but because every model judged him by what he might become, not by what he actually did on a cold Saturday afternoon. That is the difference between pricing a promise and pricing a person. My method is nothing sophisticated. I keep a notebook for each team, recording every training session, every trip, every communal meal. I count how many times a player shakes his teammates' hands after being substituted, how many times he is the first to comfort someone after a conceded goal. Those numbers appear in no scouting report. But after seven years, they give me a picture of the team that no algorithm can draw. When I asked an old Ligue 1 scout why he still travels to watch games despite having all the data, he answered: "Because I need to see who stays behind after the last training session." I do not reject data. I use PPDA myself to understand why a pressing team fails. I check xG myself to verify a match my eyes got wrong. But I use them the way people use a map, not the way people use GPS coordinates to replace walking. A map tells you the road. It does not tell you how rough the surface is, or how the locals treat a stranger. The most counter-intuitive thing I have drawn from 27 years watching this industry is this: transfer data models are not wrong because they are bad. They are wrong because they are too good at measuring what is easy to measure. A club can calculate the resale value of a 20-year-old to the nearest percentage point, but cannot calculate whether he will break dressing-room discipline in the third week. That precision creates an illusion of safety: that everything important has been accounted for. And when that illusion seeps into the boardroom, people begin to believe that buying players is like buying stocks — all you need is the right model. But football is not a financial market, because the asset here is human, and humans operate on emotion, belief, and rituals that cannot be digitised. The paradox is this: the more data there is, the more clubs tend to buy young players for "potential", and the less patience they have for older players who bring stability. They optimise for resale value instead of optimising for winning. And then they are surprised when a squad full of expensive young stars collapses against an older, lesser-known team that understands each other down to every breath. There is a second-order consequence few mention. When every club uses the same set of metrics, the market becomes dangerously uniform. Players who look alike in the numbers get priced alike, and the qualities that cannot be measured — stubbornness, the ability to ignite others, the way a player makes his teammates better — become cheaply forgotten. That is why the smartest clubs in Europe are going back to hiring "naked-eye" scouts, not to replace data, but to recover what data left out. There is a counterweight I must acknowledge. Not every ritual is good, and not every older player brings stability. There are toxic dressing rooms where experience turns into conservatism, and data is a necessary liberating tool. There are spectacularly successful young signings, and "stability" deals that are just money thrown at a mediocre season. So the issue is not data versus people — it is using one to cover for the other. The real enemy is intellectual laziness: believing a number can replace a conversation. What I learned at 43, after following eight World Cups and eight Olympic Games, is that no model replaces time. Data can tell you how many kilometres a player runs per match. It cannot tell you what he will do in the 89th minute of a final, when his legs are heavy and his mind is stretched. Only a person can answer that question, and only time gives you enough evidence to trust the answer. The signal I am watching next season is not on the transfer ticker. It is in the clubs brave enough to keep a 29-year-old with unimpressive numbers, simply because the dressing room needs him. If more and more teams pay for chemistry rather than only for potential, we will know the industry is maturing. If not, we will keep watching Europe's most expensive squads relearn an old lesson: that what decides a season never fits neatly into a spreadsheet.