Why possession stats lie
Possession percentage is the most-cited team statistic in soccer broadcasts. It is also one of the least useful, in the sense that the relationship between possession and winning is much weaker, much more conditional, and much more team-dependent than the broadcast framing suggests. A team with 65% of the ball is sometimes a dominant team. Just as often, that team is being walked into a counter-attacking trap by an opponent that wanted the ball to stay on their feet because their plan worked better without it.
The stat survives because it's intuitive — having the ball feels like an advantage, and for most amateur soccer it is. At the elite professional level the relationship inverts in a meaningful fraction of matches, and the broadcasts have not caught up with what every modern manager has been doing for fifteen years.
What possession actually counts
Possession percentage is the share of total match time during which each team had the ball, summed across the ninety minutes and rendered as a fraction of one hundred. The metric counts every kind of possession the same: a back- line cycle in your own third, a probing build-up across the halfway line, a sustained attack in the final third, and a nervous late-match clearance all contribute identically. The result is a single number that summarizes how often the team had the ball without saying anything about what they did with it or where they did it.
Two teams with the same 55% possession can produce wildly different match shapes. The first might have spent the whole match patiently moving the ball in the opponent's half, producing fifteen shots and three big chances. The second might have been pinned back, recycling possession across the back four because the press wouldn't let them advance cleanly. The column doesn't know which version it is.
The Leicester case
Leicester's 2015-16 Premier League title remains the cleanest empirical refutation of possession-as-dominance. Leicester averaged 42.4% possession across the entire title-winning season. The teams they finished ahead of averaged considerably more. Across the modern history of the league, Leicester's possession share would have been considered the profile of a mid-table team at best, a relegation candidate at worst. They won the league.
They didn't win by accident. Their tactical setup was explicitly built to concede possession in non-dangerous areas and counter-attack into the spaces that the opponent's higher line and committed defenders had vacated. Possessing the ball in your own third while waiting for the opponent to commit was the plan. The opponent's high possession share was the symptom of their plan working too — they were doing what they were supposed to. Leicester was also doing what they were supposed to. One of the two plans was better matched to the league's actual game state, and it wasn't the possession-heavy one.
Leicester was the most visible example, but it wasn't an outlier. Atletico Madrid won La Liga in 2013-14 and 2020-21 with possession profiles meaningfully below their opponents'. Several Champions League finals in the modern era have been won by the team with the lower possession share. Counter-pressing systems regularly produce winning teams whose possession looks pedestrian on the broadcast. The relationship between having the ball and winning is approximately the same as the relationship between completion percentage and quarterback value — real but loose, and meaningfully dependent on what kind of plays the team was running.
What possession is correlated with
Possession does correlate with something. It correlates, loosely, with team quality at the top of the league — Manchester City, Bayern Munich, Real Madrid in their dominant years, and Barcelona at their peak have all had high possession shares. The correlation comes from the fact that better teams keep the ball better, not the other way around. Possession is an output of high-skill technical play, not an input to winning.
The mistake the broadcast framing makes is to treat the output as a cause. A team with 65% possession isn't a good team because they have 65% possession. They have 65% possession because they're a good team and their players retain the ball when they get it. A worse team trying to copy the possession profile — by forcing back-passes and sideways play in their own half — won't win more. They'll lose differently.
The xG version of the same data
The cleaner way to measure attacking dominance is xG differential, not possession. A team's xG total counts the quality of chances they generated; the opponent's xG total counts the quality of chances they conceded; the difference tells you who had the better night attacking. The correlation between xG differential and match outcome is much tighter than the correlation between possession and match outcome, because xG measures what teams actually did with the ball rather than just whether they had it.
A team with 60% possession but a -0.6 xG differential lost on chances even though they kept the ball. A team with 40% possession but a +0.9 xG differential won the chances battle even though they didn't have the ball most of the time. These are common patterns in elite soccer. The possession column tells you almost nothing about which version of the match you watched. The xG column tells you a great deal.
The PPDA companion stat
For teams whose system is built around pressing, the most informative companion stat is PPDA — passes per defensive action — which measures how many opponent passes the team allows before making a defensive intervention. Low PPDA means the team presses aggressively and disrupts the opponent before they can build. High PPDA means the team sits deep and lets the opponent move the ball until they cross into the final third.
Reading a possession stat alongside a PPDA stat tells you much more than either alone. A team with high possession and low PPDA is dominating both ends. A team with high possession and high PPDA — letting the opponent have the ball when they get it — is more passive. A team with low possession and low PPDA is the modern Leicester archetype: pressing hard, conceding the ball back, and winning the transitions when they get the ball back in dangerous places.
How to read possession the next time you see it
Three rules. First, ignore single-game possession unless it's combined with the xG totals. The two numbers together tell you whether the team that had the ball did anything with it. The possession number alone is roughly noise. Second, ignore the standings-relevance of possession across teams of very different styles. A counter-pressing team's 42% possession is not worse than a passing team's 62%; they're different plans being executed. Third, remember the Leicester case. Possession is not a leaderboard. It's a measurement of how a team chose to play, and the choice depends on the opponent and the plan.
The broadcast convention of treating high possession as synonymous with dominance is a hold-over from an earlier era of soccer analytics, when xG didn't exist and possession was the only on-screen number anyone had to characterize attacking play. That era is over. The stats that capture what actually matters — chance quality, set- piece efficiency, defensive structure under pressure — are all publicly available now. The graphics will get there. In the meantime, when you see a 60-40 possession split, check the xG before drawing conclusions about who played better.