Manchester City ran further than every other Premier League team in the available 2025/26 club-distance data.
Across their first 27 league matches, City averaged 115.47 kilometres per match. Leeds were barely behind them on 114.75km, followed by Arsenal on 114.28km and Newcastle on 113.98km. At the opposite end, Chelsea averaged just 105.70km. Premier League physical data
It looks like the beginning of a useful performance table.
Until you add the football.
Manchester City finished with 78 points. Leeds, despite running only 0.72km less per match, finished 31 points behind them. Burnley had the sixth-highest running average but finished with 22 points. Chelsea covered the least ground yet produced a positive expected-goal difference across the completed season. Premier League physical data StatMuse
FIVDA compared the available 27/28-match distance table with all 20 clubs’ eventual results.
There was a slight tendency for the teams covering more ground to collect more points, but it was nowhere near consistent enough to treat running distance as a reliable measure of performance.
The same was true when we replaced points with expected-goal difference.
Even possession barely explained which teams accumulated the most kilometres. Premier League physical data StatMuse
The lesson is not that running does not matter.
It is that total distance tells us how much movement happened, not whether that movement was useful.
To understand physical data properly, we need to move beyond volume.
We need intensity.
Then we need purpose.
Who ran furthest? That is the easy question
The complete club-distance table available to FIVDA covers 27 matches for 18 clubs and 28 for Arsenal and Wolves, rather than the full 38-match season. It therefore gives us a substantial snapshot of each team’s running profile, but not a complete season of tracking data. Premier League physical data
Manchester City led it at 115.47km per match.
Leeds averaged 114.75km.
Arsenal were on 114.28km.
Newcastle averaged 113.98km.
Brighton were fifth on 113.59km, with Burnley sixth at 112.51km. Premier League physical data
Already, the list should make us suspicious of any simple interpretation.
City and Arsenal were near the top.
So were Leeds and Burnley.
Meanwhile, Chelsea were bottom for distance at 105.70km, while Liverpool and Everton both averaged 108.57km. Premier League physical data
These clubs were not playing football of equivalent quality simply because their running totals happened to resemble one another.
So FIVDA tested the obvious question.
Did the teams covering more ground actually collect more points?
Running more did not reliably mean winning more
Not really.
There was a slight pattern towards the higher-running clubs collecting more points, but it was weak. Knowing how far a Premier League team ran would have told you very little about where that team would eventually finish. Premier League physical data StatMuse
Figure 1 — Running more did not reliably mean winning more.
The individual clubs show why.
Manchester City averaged 115.47km and finished with 78 points.
Leeds averaged 114.75km — almost exactly the same running volume — and finished with 47.
Arsenal averaged 114.28km and won the league with 85 points.
Burnley averaged 112.51km, the sixth-highest figure in the distance table, and finished with 22. Premier League physical data StatMuse FBref
If total running were a useful shortcut for team quality, clubs with similar workloads should have produced much more similar results.
They did not.
The Premier League’s running table contained title challengers, mid-table teams and relegation sides alongside one another.
Similar workloads accompanied radically different outcomes.
That does not mean those kilometres were irrelevant to how each team played.
It means the total itself cannot tell us whether they were productive.
Maybe results are too noisy. What about xG?
Points are an imperfect way to investigate physical performance.
A team can play well and lose. Another can create little, defend its box effectively and win.
Expected goals give us another way to ask whether heavy-running teams were consistently producing stronger underlying performances.
Again, there was only a slight relationship.
FIVDA calculated each club’s final expected-goal difference — xG minus xGA — using the same StatMuse model for all 20 teams. Teams that ran further tended to have marginally better underlying numbers, but the pattern was weak and inconsistent. Running distance alone still told us very little about which teams consistently created better chances than they conceded. Premier League physical data StatMuse
Figure 2 — Running more did not reliably produce a better xG balance either.
Once again, the club comparisons are more revealing than the overall pattern.
Manchester City paired their league-leading 115.47km running average with a +26.42 xG difference.
Arsenal combined 114.28km with +36.15 xGD.
But Leeds, running 114.75km, finished at -0.76 xGD.
Burnley’s 112.51km accompanied a -41.56 xGD.
At the other extreme, Chelsea had the lowest running average in the dataset at 105.70km but finished with a +14.25 xGD. Premier League physical data StatMuse
The Premier League therefore contained elite teams that ran heavily, struggling teams that ran heavily and productive teams that covered considerably less ground.
Total kilometres could describe their workload.
They could not reliably rank their football.
Are low-possession teams doing all the chasing?
There is an obvious tactical explanation for a high running total.
Perhaps teams without the ball simply spend more time chasing it.
If that were the dominant league-wide explanation, we would expect low-possession clubs to accumulate substantially more distance and high-possession clubs to accumulate less.
The 2025/26 numbers do not show that pattern.
Across all 20 teams, possession explained almost none of the differences in total running distance. Knowing which teams had more of the ball would have been very little help in predicting which teams covered the most ground. Premier League physical data StatMuse
Figure 3 — Low possession did not simply mean more running.
Manchester City provide the clearest contradiction to the chasing theory. They had 60.6% possession while also recording the highest running average in the league-distance dataset.
Burnley reached a similarly high position in the running table with only 42.3% possession.
Chelsea went the other way: 57.7% possession, but the lowest total-distance average.
Liverpool averaged 59.3% possession while covering 108.57km per match. Leeds had 45.7% possession while covering 114.75km. Premier League physical data StatMuse
High possession did not automatically mean less running.
Low possession did not automatically mean more.
That does not mean possession is irrelevant to physical demand.
It means total distance is too blunt to show the relationship properly.
The kilometres hide what the players were actually doing
A kilometre is a kilometre on the tracking system.
Tactically, it can mean something completely different.
Players can accumulate distance while pressing, recovering after losing possession, tracking opponents, supporting possession, making attacking runs or shifting across the pitch to maintain the team’s structure.
Adding all of those movements together produces one clean number.
It also removes much of the context that makes the movement interesting.
Research across 1,675 Premier League matches from 2019/20 to 2023/24 illustrates the problem. The study used optical tracking data and separated physical output according to whether teams were in or out of possession. Teams that controlled more possession generally collected more points, while the importance of their running changed depending on whether that movement happened with or without the ball. five-season Premier League tracking study
That is a much richer question than asking who ran furthest.
Another Premier League study found that formation influenced total distance and high-speed running, while the combination of formation and possession affected total distance, high-speed running, sprinting and other high-intensity efforts. Premier League running-performance study
In other words, running load is partly a product of the football being played.
Change the formation, role or possession phase and the physical demands can change with it.
That helps explain why two teams can accumulate similar kilometres for completely different reasons.
The same principle appeared in FIVDA’s analysis of the Premier League’s shift towards more direct football: similar outcomes can emerge from different tactical routes, and aggregate volume alone rarely explains the method.
Possession changes the type of running
An earlier study of 810 Premier League players makes the distinction particularly clear.
Overall total distance was similar between players from high- and low-possession teams.
But the composition of that running was different.
Players in high-possession teams performed 31% more high-intensity running with the ball, while players in low-possession teams performed 22% more high-intensity running without it. Premier League possession study
That is almost the perfect warning against reading too much into a total-distance leaderboard.
Two teams can cover comparable ground while using that physical work in different phases of the game.
One total does not reveal that distinction.
This also helps make sense of FIVDA’s 2025/26 findings.
Possession and total kilometres showed almost no clear relationship across the 20 clubs.
The finding does not mean possession has no effect on running.
It means possession may change how the running is accumulated more clearly than how much running appears in the final total.
Sprinting tells us more — but not everything
If total distance is too broad, the obvious next step is intensity.
That improves the picture.
But it still does not produce a simple performance formula.
Across the five-season Premier League study, sprint distance was the only overall running-load measure that showed a clear positive relationship with points per game.
Even then, the relationship was modest. Teams sprinting more were somewhat more likely to collect more points, but sprint distance on its own was still nowhere near a reliable measure of team quality. five-season Premier League tracking study
So “more sprinting equals better football” would simply replace one oversimplification with another.
Opta’s analysis of Premier League physical data provides another useful warning. Manchester City led total distance at its January checkpoint, while Bournemouth led sprint volume. Fulham, West Ham and Everton were among the lower-sprinting teams, and Opta found no apparent relationship between success and occupying either extreme of the sprint table. Opta Analyst
The metric becomes more specific as we move from total kilometres to sprinting.
But it still needs context.
A sprint to press.
A sprint to recover.
A sprint beyond the defence.
A sprint caused by being caught out of position.
That distinction becomes especially important in a league where teams were also changing how they confronted and bypassed opposition pressing.
They all count as high-intensity movement.
They do not necessarily have the same tactical value.
Arsenal and Burnley show why context matters
Arsenal are particularly useful because they undermine another easy theory: that strong teams dominate possession, control matches and therefore conserve their legs.
They did not.
Arsenal covered more ground than their opponents in 35 of their first 37 league matches on their way to the Premier League title. Opta Analyst
They were also third in FIVDA’s available team-distance table at 114.28km per match. Premier League physical data
Burnley were sixth at 112.51km. Premier League physical data
Yet their final outcomes could hardly have been more different.
Arsenal finished with 85 points and a +36.15 expected-goal difference. Burnley finished with 22 points and a -41.56 xGD. StatMuse FBref
It would be wrong to conclude that Arsenal’s running caused their success.
It would be equally wrong to conclude that Burnley’s running caused their struggles.
The useful observation is simpler.
Both teams could accumulate heavy physical workloads while producing vastly different football outcomes.
The missing information is what those kilometres represented.
Player leaderboards contain the same problem
The issue becomes even clearer when running statistics are applied to individual players.
James Garner finished 2025/26 with 415.46km, the highest total distance of any Premier League player. Elliot Anderson was second on 411.06km, followed by Adrien Truffert on 402.59km and Morgan Rogers on 400.29km. Premier League
But a season total rewards two things simultaneously:
running rate and playing time.
Garner accumulated his 415.46km across 3,414 minutes.
FIVDA calculates that as approximately 10.95km per 90. Premier League
Among the same ten players who finished highest for total distance, Garner was only sixth for kilometres per 90.
Ethan Ampadu reached approximately 11.47km per 90, Jarrod Bowen 11.23 and Elliot Anderson 11.10, despite all finishing below Garner for total kilometres. Premier League
Figure 4 — The player distance leaderboard also rewards availability.
That does not diminish Garner’s achievement.
It changes what the statistic means.
His league-leading total was evidence of a high workload sustained across exceptional availability, not evidence that he ran further per 90 than every other Premier League player.
There is another layer too.
At the February checkpoint, seven of the ten players with the highest total distances were central midfielders or No 10s. Premier League physical data
Position affects opportunity to accumulate distance, just as minutes do.
Before comparing two players’ physical totals, we therefore need to ask how long they played and what their roles required.
What running data should actually tell us
Football’s physical data are seductive because kilometres look objective.
Manchester City: 115.47
Chelsea: 105.70
James Garner: 415.46
The numbers are precise.
The interpretation is not.
FIVDA found a similar distinction when testing Premier League travel: a large physical difference between clubs does not automatically produce an equally large performance effect.
FIVDA’s 2025/26 analysis found only a slight tendency for higher-running teams to collect more points or produce a better expected-goal difference. Neither relationship was strong enough for total distance to tell us much about which teams were actually better.
Possession was even less helpful in explaining the running table. High-possession and low-possession teams appeared at both ends of it. Premier League physical data StatMuse
Those results come with important limitations. There are only 20 clubs, and the complete running table covers 27/28 matches rather than the entire season. The analysis also cannot separate running by speed, possession phase, game state, opponent or tactical role. Premier League physical data
But the broader research points in the same direction.
Possession phase changes the composition of physical work. Formation changes running demands. Sprinting contains information that gross distance misses, but even sprint volume does not provide a simple route to winning. five-season Premier League tracking study Premier League running-performance study Premier League possession study
That gives us a better way to read running statistics.
Start with volume: how far did the team or player move?
Then ask about intensity: how much of that movement happened at high speed?
Finally ask about purpose: was the running happening in possession, out of possession, during pressing, recovery or attacking movement?
The first number is easy to publish.
The third question is where the football begins.
The Premier League running table can tell us who covered the most ground.
It cannot, by itself, tell us who used that ground best.
Stay tuned for future posts and please send us a message if there is a specific topic you would like to see covered.
As always, thanks for taking the time to read these posts!
JC
Sources for the data presented are referenced within the article.
The main image is from Manuel Pajon on Unsplash with the article title overlay.