For years, football analytics has revolved around a question of distribution. Is it better to concentrate the budget on one star or spread it across eleven solid players? The most popular version of the answer holds that a team depends above all on its weakest player, so plugging gaps yields more than signing stars. It is an appealing idea that has guided many decisions, but when someone bothered to test it against data, it did not hold up.
Stefan Szymanski and Guy Wilkinson designed a study in the Premier League precisely for that purpose, using the price paid for each player as a proxy for quality. They found the opposite of what they were looking for: the expensive players were the ones who mattered most to the result, while the cheap ones barely moved the needle. The most direct test of the weak-link thesis left it in worse shape than before.
It is worth adding that this result has a known weakness. Measuring talent by what a transfer cost is circular, because an expensive player receives more minutes, plays in more central positions and touches more balls partly because of what he cost in the first place. Finding that he carries more weight also measures the manager's decision. That does not invalidate the study, but it prevents it from being fully settled.
The work that best orders the matter is that of Egon Franck and Stephan Nüesch on nearly two thousand Bundesliga matches, because it unsettles both sides of the argument. In the eleven that takes the field, a large inequality of talent among the players reduced the probability of winning, just as the weak-link theory predicted. But the average level of talent weighed far more than that inequality in almost every measure they used. And when looking at the full squad across a season, inequality even appeared associated with better league positions, because substitutes and internal hierarchies serve a purpose.
Recent studies point in the same direction. One from 2025 covering three major leagues found that exceptional individual performances have a clear relationship with goal difference. Another from 2026, covering nearly six thousand matches across five competitions, found no consistent relationship between relying on extraordinary performances or on distributed contributions and points accumulated over the long term. There are teams that win by concentrating talent and teams that win by spreading it.
Put all that together and a fairly clean conclusion emerges, even if it is a disappointing one for anyone looking for a manual. What determines performance is the total amount of talent accumulated. How it is distributed matters little: in the specific eleven, inequality penalises slightly; across the full squad it even helps; and neither effect comes close to the weight of the average level. You can win with one star and ten ordinary players or with eleven good ones. What you cannot do is win with eleven bad ones.
If the answer is that simple, the interesting question shifts. It stops being about how to distribute talent and becomes about how to recognise it. And that is where football runs out of instruments.
At the very top there is no need to debate it. Nobody doubts that Mbappé or Lamine Yamal have talent; it is visible without any model. But consider what happens even there, where the judgement is unanimous. In July 2025, Football Benchmark valued Yamal at 279.7 million euros. Six months later, CIES placed him at 343 million. Two professional organisations, two credible models, sixty million apart on a player whose quality nobody disputes. If that is the precision available in the easiest case on the market, expectations for everyone else should be set very low indeed.
And everyone else is where almost all transfers are decided. Here appears the most revealing limitation in all the available research, and it is an absence, not a finding. There is no standard statistical definition of an average squad player. It is not that data is lacking: it is that the category has not been built. A substitute at Osasuna, a starter at Brighton and the fourteenth player at Real Madrid are not comparable to one another, and any figure that averages them is mixing different things. Football economics can tell you that you need talent and has no operational definition of the ordinary player against whom it would need to be evaluated.
The measures currently in use all carry the same bias. Individual ratings from providers such as WhoScored, which form the basis of the most recent studies, capture what happens on the ball better than positioning, covering runs and movement off the ball. Transfer prices incorporate goals, age, contract length and name recognition. What cannot be seen is not measured, and what is not measured is not paid for. The old debate between technique and physicality is a symptom of the same thing: there is disagreement about what talent is because nobody has managed to define it in a way that can be counted.
This goes a long way towards explaining why market prices take the shape they do. Franck and Nüesch found that one additional goal translated into a few tenths of a million in valuation around the market average and into four times as much near the ninety-fifth percentile. It is the same goal. What changes is not its contribution but the certainty you have about the person scoring it. At the top end you know what you are buying; in the middle range you are placing a bet. The elite premium does not pay only for performance; it pays for the elimination of doubt.
On top of that premium sits something else that has been measured. A study covering nine years of attendance in the Bundesliga found that superstars fill stadiums even when their team plays away, something local heroes do not achieve. And valuations incorporate a popularity component that is not explained by sporting performance. For the ten richest clubs in Europe, close to half of revenues already come from commercial activity. Signing someone whose talent is beyond question produces two goods at once, and only one of them is settled on the pitch.
There remains the final paradox to explain, which is the part of the matter that most contradicts common sense. Szymanski and Wilkinson did not only find that expensive players carry more weight: their model indicated that the optimal distribution of spending would be even more unequal than the one observed, and that if clubs do not concentrate resources further at the elite level it is probably because there are not enough players of that calibre available. The narrative of the irrational director infatuated with the star is reversed. The brake is not prudence; it is scarcity.
This is where the only known way of circumventing the problem fits in. If unquestionable talent is scarce and its price largely reflects certainty, developing it in-house is not an identity virtue but a first-order economic advantage. A club that produces a player of that tier internally obtains what the market prices at three hundred million without paying the certainty premium, because it has watched the player develop and does not need to buy information about him. The point is not to forgo stars, which is the usual sentimental reading, but to acquire them before the market knows their names.
What the research ultimately leaves behind is not a method for building squads but a fairly uncomfortable description of the terrain. We know that talent is necessary and that how it is distributed matters little. We do not know how to define it, we do not know how to measure it outside the summit, and we do not even have a statistical category for the ordinary player. The market is not solving that problem: it is dodging it, paying for the signals it can read, which are goals, name recognition and what the previous club paid. The extraordinary prices of the elite do not measure extraordinary performance. They measure how little football knows about everyone else.
