The New Age of Football Analysis: A Three-Layer Structure of Data, Tactics and Scouting

In football, analysis is not the act of repeating the scoreline. The scoreline is already on every screen. A useful analysis explains three things: which structural causes produced the result, whether those causes are repeatable or accidental, and what can be changed the next time the same situation appears. Over the past decade this work has moved out of the commentator's chair and into an organised discipline, where data science, video analysis, scouting and coaching form a single chain. The cleanest way to understand the discipline is to separate the layers of information. The first layer is event data: passes, shots, tackles, duels, fouls. The second is positional data: who stood where, in which direction they were running, and how fast. The third is interpretation: what any of this means inside the game. Machines can gather the first two layers. The third requires a human who understands football. The most serious mistakes happen where clubs assume the third layer emerges automatically. A central example is expected goals. The logic is simple: not every shot is created equal. A calm attempt from inside the box and a rushed effort from thirty yards out are not the same event. xG converts shot quality — location, angle, body part, pressure, type of assist — into a single number, which tells you how many chances a team created rather than merely how many it scored. But xG measures quality, not sustainability, and it does not reveal intent. A low xG figure may mean a team was weak, or that it deliberately sat deep and waited to counter. The number looks identical; the meaning is inverted. Positional tracking data has changed the vocabulary of analysis altogether. It is now possible to measure how high a defensive line stood, how much space opened between two lines, which player covered the most ground, and how a team's shape contracted or stretched under pressure. A decade ago these observations were made by eye and were therefore arguable. Now they can be measured, and measuring changes the argument from what happened to what it means. Its most practical use is diagnostic: a team that wants to press but consistently leaves twelve to fifteen metres between its lines has the right intention and the wrong execution — a structural finding that is far more actionable than criticism of an individual. On tactics, one misconception must be cleared away. A formation such as 4-3-3 is only the starting arrangement. Modern teams can play the same formation in three completely different ways, defined by what shape they build in possession, how quickly they press after losing the ball, and what their first choice is after a recovery. That is why analysis now separates three phases: with the ball, without the ball, and transition. Transition is the most measurable of them, because the first three seconds after a recovery determine the entire shape of a match. Possession percentage alone cannot establish whether a team is strong or weak; controlling the ball and creating danger are different things, and conflating them is the most common error in football analysis. Set pieces have also moved from the margins to the centre. Corners, free kicks and throw-ins are static situations, which means they can be rehearsed under controlled conditions while open play cannot. Where two teams are evenly matched in open play, set pieces can separate them. Modern analysis asks how the opponent marks — man or zonal, who stands at the near post, how much freedom the goalkeeper has — and how prepared they are for the second ball. None of this is sufficient alone; it becomes useful only when connected to the design of the runs. Scouting is where data and human judgement meet most directly. An algorithm can tell you which player in a league matches a statistical profile. It cannot tell you how that player will adapt to the pace of a new league, an unfamiliar language and a new dressing room. The first is a filtering question, the second an investment question, and weighing them on the same scale is scouting's biggest mistake. Good scouting therefore runs in the opposite direction: first the club defines the role, the style, the budget and the responsibility the player will carry, then data produces a long list, then video is studied, and only then come personal conversations and background checks. Clubs that reverse this order often buy good players who are not useful to them. The age and development curve of a player is another frequently ignored dimension. Two players of the same age do not develop at the same rate, and physical precocity at twenty does not guarantee superiority at twenty-four. Scouting therefore increasingly looks at rate of improvement — how a player's decision speed, positioning and physical data have moved across two seasons. Youth development and load management have become intertwined. Injuries were once treated as fate; now muscle load, sprint counts, high-speed running volume and recovery intervals are tracked. The purpose is not to play a player less but to judge what load his body can absorb in a given week. There is an ethical dimension here too: if every metre is recorded, the limits on the use of that data must be defined just as clearly, or monitoring slides into surveillance. Opposition analysis has also been standardised. The usual sequence is to establish the opponent's stable identity — how they look at their strongest — then to identify when and why they departed from it. Weaknesses are sorted into three categories: structural, individual and situational. Structural weaknesses sit inside the system, such as a large gap between the lines. Individual weaknesses are the limits of a player. Situational weaknesses come from weather, scoreline or fatigue. The classification tells the coach what to trust: structural weaknesses can be planned for because they appear every week, individual weaknesses are riskier because a good player can hide them on the day, and situational weaknesses are the weakest link of all because they are not durable. Then there is the organisational side: communication. The best analysis in the world is worthless if players cannot recall it on the pitch, which is why good departments produce messages rather than reports. A full tactical review should compress into three sentences that can be repeated at half-time. Analysis that understands the game but cannot simplify itself is unfinished. This connects to the media's narrative instinct: the most comfortable stories have a hero, a villain and a clear moral, and they usually hide the real mechanisms, which are almost always structural. Analysis must also state its own limits. The first is sample size: four matches produce no robust conclusion, however clean the numbers look. The second is context: opponent quality, home and away, travel, rest days and weather are the background without which a number is incomplete. The third is overfitting: an analysis detailed enough to explain every detail of past matches loses the ability to predict the next one. The fourth is the unmeasurable: a player's trust in a teammate, composure under pressure, leadership in a dressing room — these have no satisfactory indicator, yet their effect on results is unmistakable. The fifth is resource inequality: elite clubs and smaller clubs play the same game but do not live in the same information world. For smaller clubs the smarter route is selection rather than breadth. The strength of modern football analysis therefore lies in the combination of three things: data that is objective, an eye that understands context, and communication that carries information onto the pitch. Remove any one link and the other two generate noise. The road ahead is not about inventing another metric, but about whether analysis can reach the point where it states its own uncertainty clearly. A large part of football is still chance, and good analysis is the kind that has the courage to call chance by its name.