How it works
What the figures on this site are, where they come from, and what they cannot tell you.
By Lamin Diallo. Code, data and the register of tests: github.com/lamindiallo/competitive-mobility-project.
The question
Can a club that is not already at the top get there — and how often does it happen? The project looks at every club in the divisions where clubs move between the top and the rest: the top two divisions in Spain, Italy, Germany and France, and the top four in England, from the early 1990s to the end of 2025/26.
"How high can my club realistically go?"
We do not predict anything. We look back. For every season since 2004/05, we note where each club finished and then follow it forward: did it reach the top four within ten seasons? Win the league? Stay up for the next five? Go up?
Clubs are grouped by where they finished — for example, "finished 5th to 7th in Serie A", or "was promoted from the Championship". The figure on a club page is the share of clubs in the same group that reached each goal. "About 1 in 6" means that, of all the times a club finished there, roughly one in six went on to do it.
Ten-year goals use finishes from 2004/05 to 2015/16; five-year goals use finishes from 2004/05 to 2020/21. That way every case has had its full time to play out. We chose this window because it is recent enough to reflect today's football and long enough to give stable figures.
It describes the position, not your club. A club with far more money or history than is usual for its position — a giant in a bad season — has better chances than the figure suggests. That is why each club page also shows the club's own last ten seasons, and says so plainly when a club is much bigger than its position.
"Against the money"
In England and France, clubs' wage bills are public: English clubs file accounts at Companies House, and the French league's financial watchdog (the DNCG) publishes every club's accounts. We use the wage bill because, across football, it is the single best guide to where a club finishes.
For every season we compare each club's wage bill with the average in the top two divisions and work out the finish that wage bill usually brings. The gap between that and where the club actually finished is "places above (or below) its money". An overall verdict needs at least eight seasons, and a club is said to beat (or fall short of) its money only when the gap is clearly bigger than chance would produce.
Most of a big over-performance fades, but not all of it. Clubs that finish well above what their wage bill usually buys are, on average, still above it the next season, though by about a third as much; if it were pure luck they would be back to zero. A few clubs beat their money almost every season. We found no sign that over-performance turns into a bigger wage bill later. The measure mixes good management, luck and resources we cannot see, so it is not a score for how well a club is run.
In England and France, club pages also show a "realistic ceiling" that takes the wage bill into account. We checked it on clubs the model had not seen. Adding the wage bill clearly helps for reaching the top four in England and for staying up in France; it helps a little for staying up in England and for getting back to the top flight in France; and it adds nothing for getting back to the top flight in England. On reaching the top four in France, one of our two accuracy scores improves clearly and the other barely moves. Where the wage bill adds little, the club page says so. The full check is in money_ceiling_models.csv.
What the project found
Each of these was written down, with the test and what would count as a "no", before any result was calculated. A "no" is reported as prominently as a "yes". The register of tests is published with the code. It is the project's own dated document, not an entry in an outside registry, so its dates are not independently timestamped.
| Question | Answer | In plain words |
|---|---|---|
| Has it got harder to reach the top across European football? | No, as registered | On the data the test was registered on, only England (since the early 1990s) and Italy (since the mid-1990s) declined, with Spain borderline. Re-run on the corrected tables (official positions, 2025/26 added), Spain also declines, so a wider decline cannot be ruled out. Germany and France show no change on either version. No league has become more open. Details of the re-run. |
| Does a club's history matter, beyond where it finished last season? | Yes | But the pull is back towards each club's usual level: newcomers to the top four rarely stay, and fallen giants often return. "Success breeds success" is not what the data show. In England and France, most of this pull disappears once clubs' wage bills are taken into account: it mainly reflects lasting differences in money. |
| Is the top harder to break into than the bottom is to escape? | Yes | Top positions stick more than chance would allow, and more than bottom positions, in every league except Germany. |
| Does promotion and relegation tell us whether the top is open? | No | How many clubs move up and down at the bottom says nothing about whether the top four can be reached. |
| Has the growth in league revenue made the top more even? | No | Across the five leagues, seasons with far more money in the league have not had a less concentrated table. If anything the reverse. Based on Deloitte's published league revenue figures for 1996/97 and 2008/09 to 2024/25 (one season missing); this describes what happened and does not show why. |
| Does finishing above your money lead to a bigger wage bill later? | No | On English and French figures, clubs that beat their money are not reliably followed by bigger wage bills. |
| Can a club climb for good by being cleverer than its wage bill? | No sign of it | Clubs that beat their money three seasons running were no more likely to move up for good than their position and wage bill already suggested. Among clubs whose wage bill did not grow, almost none climbed. This does not mean good decisions never matter; it means that, without money following, the gains do not last. |
| Does the extra money from doing well arrive too late to help? | No | It arrives quickly: each place a club finishes above its money brings about 4% more revenue, compared with the rest of the league, by the next season, and nothing more after that. So rules that tie spending to revenue do not hold clubs back by making them wait for their money. |
| At the same wage bill, is the top four harder to reach in England than in France? | Not proven | English clubs did break in somewhat less often, but not clearly enough to pass the test set in advance. Having more well-paid rivals nearby explained none of it. |
The mobility measure
Every club is placed on one ladder each season: the champions first, and the top club of the second division straight after the bottom club of the first. The ladder covers England's four divisions and the top two divisions elsewhere, so clubs that drop out of the top flight are still followed. From the way clubs actually move on that ladder we work out how long, on average, it takes a club in the bottom part of the top flight (11th to last) to reach the top four, and compare it with a league in which each season's positions were reshuffled at random. A value of 1 would be as open as pure chance.
| League | Seasons | Times as long as by chance | Likely range |
|---|---|---|---|
| England | 1992/93–2025/26 | 4.3 | 3.7–5.4 |
| Spain | 1993/94–2025/26 | 3.2 | 2.8–3.8 |
| Italy | 1994/95–2025/26 | 3.2 | 2.7–3.7 |
| Germany | 1992/93–2025/26 | 2.5 | 2.2–2.9 |
| France | 1993/94–2025/26 | 2.4 | 2.1–2.6 |
The likely ranges come from drawing whole seasons at random, with replacement, and recalculating. Drawing them in blocks of four neighbouring seasons, to allow for seasons that resemble each other, widens the ranges a little (England 3.7–5.8, Spain 2.7–4.4). Either way, England's figure is higher than every other league's in at least 97% of the redraws. Spain's and Italy's ranges overlap heavily, as do Germany's and France's, so we do not rank within those pairs.
The registered test, run before the site existed on data to 2024/25, gave England 4.0, Spain 3.4, Italy 3.3, Germany 2.6, France 2.4. The figures above repeat the calculation on this site's own tables, which add 2025/26 and the official corrections, so that anyone can reproduce them from the code on the data page. The numbers move a little; the order does not.
Spanish reserve teams (Barcelona B, Real Madrid Castilla and others) are on the ladder, because they played in the second division and the other clubs' positions depend on them. Taking them out changes Spain's figure by less than 0.1. They are left out of every "realistic ceiling" figure, because they cannot be promoted into their first team's division.
How this relates to other work
- Buzzacchi, Szymanski and Valletti (2003), "Equality of Opportunity and Equality of Outcome: Open Leagues, Closed Leagues and Competitive Balance", Journal of Industry, Competition and Trade 3(3), 167–186, compare how often new teams reach the top ranks with what equal chances would produce. Our mobility measure follows the same idea, applied to the whole pyramid.
- Ramchandani, Plumley, Boyes and Wilson (2018), "A longitudinal and comparative analysis of competitive balance in five European football leagues", Team Performance Management 24(5/6), 265–282, also find Ligue 1 the most balanced of the five. Unlike us, they report a decline in four of the five leagues (all but Serie A); they measure concentration and dominance within the top flight, while we measure movement across the whole ladder.
- Hall, Szymanski and Zimbalist (2002), "Testing Causality Between Team Performance and Payroll", Journal of Sports Economics 3(2), 149–168, find that in English football higher payrolls lead to better performance, which is why we judge clubs "against the money".
- Noll (2002), "The Economics of Promotion and Relegation in Sports Leagues", Journal of Sports Economics 3(2), 169–203, discusses why promoted clubs often struggle to stay up.
- Sanderson and Siegfried (2003), "Thinking about Competitive Balance", Journal of Sports Economics 4(4), 255–279, review how competitive balance is defined and measured.
- Kuper and Szymanski, Soccernomics (first published 2009), is the best-known popular account of football economics, including wages and results.
What these figures cannot tell you
- Why. We can see where and when the top of each league closed. We cannot say which rule or decision caused it: every league changed at once and nothing was tried at random.
- Your club's future. Every figure describes what happened to clubs in a position, not what will happen to one club.
- How well a club is run. Beating the money mixes skill, luck and things we cannot measure.
What we never publish
- Anyone's individual pay, or anything from which it could be worked out.
- Transfer-market valuations, or anything derived from them.
- Club badges: they belong to the clubs.
How the data are checked
- League positions to 2024/25 are worked out from match results and checked against an independent source (in England every club and position in 88 division-seasons agreed). Positions then follow the official tables: head-to-head tie-breaks where the league uses them (Spain; Italy from 2005/06), point deductions, titles taken away, and in England the two divisions decided on points per game in 2019/20. The second divisions of Spain, Italy and France come from published final tables. Every change is listed in the downloads, and a sample of corrected seasons was checked against the published tables.
- The 2025/26 final tables are the official tables, checked against a separate results file for the five top divisions and the Championship. Where the two differed, the official table had applied a head-to-head tie-break, a points deduction or an awarded match.
- Every wage figure must add up from its own parts as printed in the accounts, and most are confirmed again by the next year's accounts. No figure was read by an AI model; all were read by fixed rules that can be re-run.
- This season's table is provisional. It comes from football-data.org and is refreshed every 30 minutes; if that service is unavailable, a volunteer-run public-domain results file is used instead, which can run behind.
Spotted something wrong? Tell us, with a source. Version 2.5, 2 November 2026.