What is possession, and how is it measured?

That's the short version. Here's how the number is actually built, why two broadcasts can show different possession splits for the same match, and what FootyMetrics shows for it.

What is possession?

Possession is the share of a match a team spends in control of the ball, expressed as a percentage. Add the two teams' numbers together for any match and they come to 100, since the ball is always either with one side or the other while it's in play. A team on 62% possession had the ball for a clearly larger share of the match than an opponent on 38%.

That much is simple. The part worth being precise about is how "share of the match" actually gets worked out, because there's more than one way to do it, and the answer changes what the number is really measuring.

The two calculation methods

Two broad methods exist across the data industry, and they measure genuinely different things.

The time method tracks the actual clock. Every second the ball is in play gets assigned to whichever team has it, and the seconds are added up and divided by total playing time. It's the most literal reading of "possession": if a team had the ball for 33 of the 60 minutes the ball was actually in play, that's their possession share.

The pass-count method, sometimes called the touch-count or event-count method, doesn't use time at all. It counts the number of passes, touches or on-ball events each team records, then divides each team's count by the combined total for both sides. If two teams combine for 800 completed passes in a match and one side made 480 of them, that side is credited with 60% possession, regardless of how long the ball actually spent at each side's feet.

Opta, and by extension most of the football data industry, uses the pass-count family rather than a stopwatch. Opta's own methodology has moved on from a raw pass tally: since 2017 it has counted the number of discrete possessions each team has, where a possession runs from the moment a player takes a controlled touch until the opponent wins the ball back, the ball goes out of play, or the half ends. That's still fundamentally an event-count approach rather than a clock. It's a refinement of counting touches and passes, not a switch to timing the ball in seconds. Opta's own definitions describe a possession as one or more sequences in a row belonging to the same team, ended when the opposition gains control, which is an event-based framework rather than a timed one. FootyMetrics uses Opta-standard event data, so the possession percentage shown across the site sits in this pass-and-touch-count family.

Why the two methods can disagree

The gap between the methods shows up most with teams built around short, quick passing. A side stringing together twenty short five-yard passes in a build-up phase racks up twenty counted events under the pass-count method, even if the whole sequence only takes fifteen seconds of clock time. A side that instead holds the ball at a defender's feet for ten uneventful seconds before playing one longer pass registers as one event over a similar stretch, despite controlling the ball for a similar or longer spell of actual time.

That's why a team known for rapid, high-volume passing can show a noticeably higher possession percentage under the pass-count method than a stopwatch on the same match would suggest. It's not that the pass-count number is inflated or wrong. It's measuring how large a share of the ball-playing events in the match belonged to each side, not how many seconds each side spent with the ball. The two only match up neatly when both teams play at a similar tempo, which is often not the case.

Dead-ball time

The two methods also treat stoppages differently. A genuine time-based system stops the clock the moment the ball goes out of play, so a throw-in being retrieved, a goal kick being set up, or an injury being treated doesn't add seconds to either team's total. Opta's event-count approach doesn't run on a clock in the same way, so those gaps in play simply produce no possession events for either side while the ball is out, rather than being explicitly stopped and resumed like a stopwatch. Neither method hands one side extra credit purely for a stoppage.

FootyMetrics couldn't find a single published Opta rule spelling out exactly how the first touch after a restart is scored, so treat that specific detail as unconfirmed rather than assumed.

Why possession alone doesn't tell you who's winning

A high possession percentage isn't proof a team is playing well or creating danger. A team can hold the ball for long spells in its own half and in front of the opposition's defence without ever threatening the goal, while a team with far less of the ball camps in behind and scores from two or three direct attacks. Possession measures how much of the match a team had the ball. It says nothing about where on the pitch that happened or what the team did with it.

This is part of why modern analytics leans on more targeted ideas alongside the raw possession number. One example that's become common in tactical analysis is "field tilt", sometimes called attacking-third possession share, which looks specifically at how much of the possession in the final third or in dangerous areas belongs to each side, rather than treating every square yard of the pitch as equally meaningful. A team can be behind on overall possession and still be tilting the game their way in the areas that matter.

What FootyMetrics shows for possession

FootyMetrics tracks team possession percentage for every match, built from the same Opta-standard event data behind the rest of the site's stats. It shows on team pages alongside the wider set of match and season stats.