Runs That Lead Nowhere: A Week in Lyon and How We Read Football Data

Twenty minutes after the final whistle at Groupama Stadium, row seven of the press tribune had emptied out. I stayed in my seat, staring at the laptop screen. The match data packet had just been uploaded to the internal server. File name: L1_J_export_final. Size: four kilobytes. Contents: nothing. Not a single line of coordinates. No PPDA. No xG. No heat map. There was only a row of column headers and a lone comma on the second line, as if someone had opened the door to the room and walked out halfway through. The data-centre technician, the young man who still pours me coffee every Saturday afternoon, shrugged: "Server sync error, sir. It will be there tomorrow morning." I nodded and packed up my laptop. But that shrug, on a late-autumn evening in Décines, suggested far more to me than a technical fault. It reminded me that for twenty years I have trained myself to read a match through numbers — and that some nights the numbers vanish while the match remains exactly where it was, full of things that cannot be measured. The pulse never lies in the ball; it lies in people. I have written that sentence for years. Only on that evening, when the file came back empty, did I understand how true it is. European football entered this regular season with a paradox that has become structural. The volume of data generated per matchday now exceeds the total volume of the entire 1990s, yet our ability to read it correctly has not grown at the same pace. Each Ligue 1 match produces millions of positional data points, hundreds of thousands of ball events, and a metric list so long that no coach reads it all. The analytics departments of French clubs — even mid-budget ones — now employ at least three to five full-time staff, a figure reserved for major corporations fifteen years ago. Olympique Lyonnais is a clear example. This club was once the emblem of a self-developing football culture, with seven consecutive league titles between 2026 and 2026 and an academy referenced across Europe. When I began following the team in 2026, the coaching office held a tactics board, videotapes and a fax machine. Today, along the same corridor, there is an air-conditioned room, six large screens, and a data-sync process where a single wrong click sends everything downstream. Alongside that explosion, financial pressure has tightened. French clubs must balance volatile broadcasting revenue, steadily rising wage costs, and the league regulator's financial control requirements. In that environment, data becomes a pricing tool: a player who runs twelve kilometres per match is worth something different from one who runs nine. A defender with 92 percent pass accuracy is valued above one with 84 percent. Precisely because of that, the quality of data infrastructure is no longer a technical matter. It is a sporting one. When the data file is empty, the question is not what went wrong with the server, but what happens to the club's decision-making over the next forty-eight hours. One rule change has profoundly reshaped how matches unfold, and I believe it is the reason data has become dangerously important. The five-substitution rule, widely adopted after the pandemic disruption, was initially praised as a player-welfare measure. In practice, it turned the final twenty minutes into an organised war of attrition. Previously, with three substitutions, the closing twenty minutes were a period of shared fatigue. Both teams tired together and accepted a slower rhythm. With five substitutions, the squad with greater depth can change almost half the team at minute 60, sending on faster, fresher players and turning the last twenty minutes into an entirely different physical contest. Based on my experience covering Ligue 1 matches over the past four seasons, I have noticed a striking repeating pattern. Big-budget clubs often deliberately extend the game to minute 65, holding a draw or a fragile lead, before releasing their high-quality bench options. They do not need to win the first seventy minutes. They only need not to lose, and to let squad depth do the rest. This is something broadcast data rarely mentions, because it produces no attractive headline. But it explains why the winning team's distance-covered figure in the final twenty minutes is often no higher than the losing team's. They do not run more. They run at the right time. And here the story grows complicated. Distance covered and sprint counts have long been packaged as effort metrics. Post-match graphics print the name of the player who ran most, as an implicit compliment. But I have sat in enough club analytics rooms to hear the data specialists themselves admit the opposite: ineffective running also produces very beautiful numbers. A midfielder repeatedly dragged out of position will run a great deal to compensate for his own mistakes. A defender who cannot read a pass must sprint more often than one who reads it and intercepts before it becomes a problem. On the spreadsheet, these two players can look nearly identical. On the pitch, they are two different tiers. That is why I never accept reading a single metric in isolation. PPDA, the measure of a team's pressing intensity, means something only when you know whether the team is leading or trailing, home or away, chasing points or protecting a man. xG means something only when you know how the shooter struck the ball, with which foot, after how many seconds of running. A complete data file without context is as dangerous as an empty one. Both lead to the same outcome: people draw conclusions about a match they never actually watched. There is a memory I carry with me every time I sit down in a press room. It is from the summer of 2026, in Russia, when a substitute of the France national team sat in front of me for two full hours in a hotel near the training centre. He did not cry. He simply talked. About waking each morning, checking the projected line-up, and knowing his name would be on the twelfth line. About sitting on the bench in a World Cup quarter-final, hearing the stands break open in waves, and feeling as though he were watching a film he had once been invited to act in. No metric captures that. Not xG, not PPDA, not heat maps, not recovery indices. Yet the substitute's loneliness is a structural part of modern football, and it only grows as substitution quotas expand. Five allowed changes mean seven players left sitting. Seven human beings, seven psychological states, seven stories nobody interviews. I did not write that piece to provoke tears. I wrote it to reframe the question of what we are measuring. A football culture that measures everything on the pitch but nothing about the person off it has a systematic blind spot. And that blind spot is not in the technology. It is in what we choose to measure. The same logic applies to the transfer market. Every transfer contract is a turning point of a life, not of a spreadsheet. I have watched deals valued by amortisation algorithms, by age, by minutes played, by value-added-per-minute indices, collapse within four months for reasons nobody modelled: a player's wife could not find a suitable school for their children. No analytics department in Europe includes that as a variable. Not because they lack competence, but because it falls outside the model. And what falls outside the model exists in a different, more persistent, harder-to-process way. I came to Lyon to find a media wave, and I stayed to learn how to listen. I still use that line when students ask why I remained in a city I had intended to visit for a few months. The reason is specific. In 2026, when tension between the club and its most loyal supporters peaked, I organised a face-to-face dialogue at Groupama Stadium involving around 120 supporters' group members and the coaching staff. I did one thing: I kept people talking to each other as human beings, not as two sides of a newspaper page. I mention this not to claim credit, but to show that those periods taught me to read a club in a way no metric can touch. A communications crisis does not appear in a PPDA table. Trust eroded over months does not show up in xG. Three years later, when the pandemic emptied the stands, I fell into a serious psychological crisis. An empty stadium, yet I could still hear applause from distant balconies. That applause was real, and it came from people who were not present in any measured space. I began calling fifteen Lyon players' families. I listened to players' wives describe the fear of a husband losing form, of a contract nearing expiry, of a household income depending on a run of ten matches. I listened to a mother whose son plays in the youth team weep when supporters called her child by name over a video call. No data exists on those calls. Nobody aggregates them. Nobody puts them in the end-of-season report. Yet they are football, in the fullest sense of the word. At this point I must say something many colleagues dislike hearing. The football analytics industry has a weakness that does not lie in the quality of its algorithms. It lies in the assumption that what cannot be measured does not matter. That assumption becomes a belief, the belief becomes a habit, and the habit becomes the way an entire industry sees the game. The consequence is that we grow ever better at describing what happened and ever worse at understanding why it happened. We know how many kilometres more a team ran than its opponent, but not who was speaking and who was silent in the dressing room. We know how many shots a player took, but not how many hours he slept the night before. This trend is not merely academic. It shapes decisions directly. A club buying players on a model can overlook signals any beat reporter spots after two training sessions. A coach judged on a dashboard can be sacked at precisely the moment his team needed continuity. I write from the inside, where outside noise cannot reach the pulse. That is the privilege of the long-serving beat reporter, and also the responsibility. Because when the data file comes back empty, the only person left who can recount what happened in the match is the one who sat there, watched, listened, and recorded it with his own ears. So which signals should we track in the coming matchdays? First, watch how teams manage the final twenty minutes. This is the earliest indicator of genuine squad depth, as opposed to depth printed on a transfer ledger. A team that sustains intensity after minute 70 without replacing its entire midfield has a sturdier structure than appearances suggest. Second, note teams that begin the season with unusually low first-half pressing and a sharp second-half spike. This pattern often signals a side deliberately saving energy for later phases — or one still searching for an identity and compensating with effort. The two look identical in the data but lead to very different outcomes by March. Third, pay attention to players with high distance covered but a low number of touches in dangerous zones. This group is the most misjudged in modern football. Tactics will age, but human stories never expire. At a time when every stride is recorded and every pass is encoded, there is a question I think readers should ask themselves each matchday: in the game just played, who was mentioned, who was forgotten, and what was truly happening behind the scoreline? I will answer that question by staying in Décines, waiting for tomorrow's data file, and continuing to write down the things the spreadsheet refuses to keep.