What is the value of ball possession in football?
On the penultimate matchday of the 2022/23 season, Arsenal visited Nottingham Forest at the City Ground. The Gunners ranked second, theoretically still in with a chance of winning the title. While 16th ranked Forest were still in a relegation battle. It was Arsenal who dominated possession with 81.7 percent. A statistic that put them amongst the top 10 of teams with the highest possession during a Premier League match since such statistics were collected in 2003/04. And yet, with only 18.3 percent possession, 161 passes versus Arsenal’s 724, and just two shots on target, it was Nottingham Forest who won with 1-0.
The result was enough for Arsenal to officially lose the title race to Manchester City despite having led the league for a record 248 days. Forest, on the other hand, secured Premier League safety that day. They also set a Premier League record for having the least possession while winning. Taking over from Burnley, who beat Liverpool in 2016/17 (2-0) while recording a possession of 19.4 percent.
It is illustrative of how unpredictable and counterintuitive football can be at times. Teams can dominate matches with possession, shots, and any other statistic, and yet they can still lose.
The question is whether these are just outliers or not. How important is possession really when it comes to results? What is the value of having more of the ball? And has this changed over time?
Does more possession mean more success?
Monerals, CC BY-SA 4.0, via Wikimedia Commons
So far, the data on whether more possession results in more success seems inconclusive. With the type of competition (domestic vs European club football vs national team competitions) as well as the game changing over the years, making a difference.
During the 2008/09 La Liga season, for example, ball possession was concluded to be one of several key variables differentiating between winning and losing sides. A 2013 study concluded that in the Big Five leagues possession was a better predictor of victory. Yet, in the Champions League and during national team tournaments no (significant) impact was observed.
Euro 2020 data suggested possession did influence success at a national team tournament. With teams with more possession scoring on average more goals from inside the box. While teams at the 2022 World Cup who had more possession also ranked better for other statistics often associated with success, like shots, passes and offensive actions.
Possession has value above a certain threshold
A study of 19,949 domestic league matches across the 2016/17 and 2017/18 seasons, also showed a link between possession and success.1 However, importantly, only from a certain level of possession did it have a significant effect. A positive link was found between possession percentage and goal difference. With the real difference emerging from 57 percent onwards, at which time teams’ possession resulted in a significantly higher chance of outscoring opponents and thus winning.
Similar conclusions were drawn by using the difference in successful passes (threshold of 55 percent) and successful passes in the final third (threshold of 60 percent) as a measure of possession. With the latter being the best indicator for success (for both defined by points per match and goal difference) across a season.
What is the relation between possession and success? based on 35 leagues in 2016/17 and 2017/181
| Sort possession | Threshold for impacting success (i.e. goal difference) | Correlation | |
|---|---|---|---|
| Points per match | Goal difference | ||
| % of possession | >57% | 48.8% | 52.0% |
| % of successful passes | >55% | 46.7% | 49.8% |
| % of successful passes in final third | >60% | 53.2% | 57.8% |
During the two observed seasons, possession percentage in the Premier League had a 72 percent correlation with points per match at the end of the season. It ranked the English topflight third behind the Slovenian and Croatian leagues. With the Premier League champions having an average possession of 61 percent. While successful passes across the whole field and in the final third were respectively 68 percent and 65 percent.
Leicester recorded the lowest possession for a Premier League champion
Extending the period from two to 10 seasons (between 2014/15 and 2023/24) and the Premier League champions recorded a similar average possession: 62 percent. During that time, Manchester City recorded the highest possession with 71 percent (2017/18).
Leicester City recorded the lowest possession by a champion during their 5000-to-1 championship run in 2015/16. Their 42.6 percent ranked the Foxes 18th in terms of ball possession that season. Only Sunderland (41.0 percent) and West Brom (39.8 percent) saw less of the ball, and they finished the season in 17th and 14th place respectively.
Despite being dominated in possession and having the league’s second lowest percentage of completed passes, Leicester scored well on other metrics that are instrumental to results. Overall, they ranked third in both goals scored (68) and conceded (36). Their early pressing and high number of ball recoveries (league’s best) contributed to an effective counter attacking strategy. During their successful run they scored the most counter attacking goals (six) off all Premier League sides. Which accounted for nine percent of their total output. Only 2019/20 champions Liverpool had a higher percentage (13 percent) of their goals come from fast breaks.
The Foxes did not rank particularly high for number of shots on target and underperformed their expected goals. Yet, the percentage of their shots that were saved (65.3) was the league’s second lowest. While they had the second highest save percentage by a goalkeeper (73.5 percent). Implying effectiveness at both ends of the pitch.
Guardiola’s style impacting the Premier League
Leicester’s fairytale run was an outlier in terms of ball possession though. The Foxes have been the only Premier League champion during the indicated decade who saw less than 50 percent of the ball.
In six out of the ten seasons the champion also recorded the league’s highest possession. On all occasions this has been Manchester City. Which can be explained by the arrival of Pep Guardiola from Bayern Munich in 2016. Since the Spaniard’s arrival, City has won six Premier League titles. Furthermore, the Citizens have topped the possession table each season with an average of 66.4 percent. Which is eight percent points more than what City recorded pre-Guardiola between 2014/15 and 2015/16 (58.4 percent).
Before Guardiola’s spell at City, it was rivals United who recorded the league’s highest possession. However, with 61.2 percent and 58.5 percent in 2014/15 and 2015/16 respectively, they recorded lower percentages than City have since.
The highest possession recorded by a champion other than City, has been Liverpool in 2019/20. During their first Premier League title run, Liverpool recorded a possession of 62.9 percent. Only second to City’s 66.2 percent that season.
Low possession in the battle for relegation
Those finishing last between 2014/15 and 2023/24 always had possession under the 50 percent. Sunderland (2016/17) and Sheffield United’s (2023/24) possession even dropped below 40 percent. In half of the ten seasons, the team finishing 20th ranked amongst the bottom five for possession. Norwich City, however, finished last in 2019/20 despite having the ninth highest possession (49.3 percent). In 2015/16, Aston Villa had the 14th highest possession (46.4 percent), more than champions Leicester, and yet they finished at the bottom as well.
Having the lowest possession does not necessarily equal relegation either. On four occasions did the team with the lowest possession end up in the bottom three. During the six other seasons they secured Premier League safety despite, or maybe due to, their low possession. In 2014/15 and 2016/17, Crystal Palace and West Brom finished 10th despite having the lowest possession.
The season prior, West Brom saw the least of the ball (39.8 percent) as well while securing a 14th place finish. However, it was not sustainable. As a third consecutive season recording the league’s lowest possession (40.7 percent) saw them finish 20th in 2017/18.
Managers’ influence on possession
Football.ua, CC BY-SA 3.0 GFDL, via Wikimedia Commons
Guardiola’s move to Manchester City sparked changes in the Premier League. Highlighting the importance of a managers and their preferred style of play on how much possession teams have.
It is widely known that the Spaniard prefers his teams to dominate possession. Across his senior managerial career, Guardiola’s teams have had an average possession of over 66 percent in league games. Whilst at Bayern Munich, his team had even more possession than City have had under his management. In 2014/15 and 2015/16, the Bundesliga Meister had the ball over 70 percent of the time. Which was at least six percent points more than other teams.
Guardiola’s tiki-taka Barcelona side was known for their possession-based style with many short, fast passes. With amongst others Busquets, Iniesta, and Xavi, who were accurate passers and had great ball retention abilities even under high pressure, they had the players to execute Guardiola’s ideas. By dominating possession, they would tire out opponents and find space to attack. Often behind the last line through their full backs. Did they lose possession, they aimed for high pressing and winning the ball back quickly. With success as they won 14 trophies during Guardiola’s four years in charge.
Guardiola vs Simeone
Guardiola, and the teams he has managed, are slightly unique in their (possession) dominance. Yet, he is not the only manager who (prefers to) dominates possession when possible. Across the full seasons Jürgen Klopp has managed between 2014/15 and 2023/24, his teams had an average league possession of 61 percent. At Liverpool he never had a season where possession dropped below 60 percent. While new England manager Thomas Tuchel managed five full seasons at Dortmund, PSG, and Bayern, recording an average league possession of 64.6 percent.
Carlo Ancelotti’s sides at Real Madrid and Napoli were dominant in possession as well. However, in his full season at Everton his team finished tenth with 46.4 percent possession.
José Mourinho and Diego Simeone are two managers known for their more defensive style of play. Which is reflected in an average possession of 53.4 percent and 49.5 percent respectively between 2014/15 and 2023/24.
During the observed period Unai Emery managed full seasons at Sevilla, Villarreal, Aston Villa and PSG. And while his sides had an average possession of 56.9 percent, at PSG he recorded possession of over 65 percent. Indicating that squad quality and league affect possession as well, irrespective of the manager.
Possession as indicator of league strength or disparity
Despite Manchester City’s dominance in recent years, the Premier League is widely considered the best league in the world. UEFA’s (club) coefficients also indicate this. With England ranking highest since 2020/21 when they replaced Spain after their eight-season stint at the top. Opta’s league power ranking ranks the Premier League as strongest league as well, ahead of the Serie A.
Possession could be another metric to compare leagues’ competitiveness. With less variance in ball possession amongst clubs implying a more competitive league. In 2023/24, the Bundesliga recorded the smallest difference in possession between the team with the most possession (champion Leverkusen: 61.8 percent) and the least (eighth ranked Heidenheim: 42.5 percent). Amongst the Big Five leagues, Ligue 1 had the largest variance with 29.6 percent points. Slightly ahead of the Premier League in which champion Manchester City (65.2 percent) recorded 29.4 percent points more than relegated side Sheffield United (35.8 percent).
However, League 1’s clubs with the highest (PSG: 65.2 percent) and lowest (Metz: 35.6 percent) possession were outliers compared to the rest of the league. Without the highest and lowest values in all five major leagues, Ligue 1 had the least variance (12.3 percent points).
Furthermore, while the Bundesliga had the least variance, the median (46.8 percent) was the lowest amongst the five leagues. Indicating that across the season more teams (12) saw less than 50 percent of the ball than those who dominated possession (six teams).
Amongst the Big Five leagues, Manchester City and PSG (both 65.2 percent) recorded the highest possession. Ahead of Barcelona (64.3 percent). Despite having the highest possession in La Liga, Barcelona did not win the league. Just as Napoli did not win the Serie A. They finished second and tenth respectively.
The three clubs with the lowest possession across the Big Five leagues in 2023/24 were Metz (35.6 percent), Sheffield United (35.8 percent) and Udinese (39.3 percent). At the same time, they were the only three teams (out of 96) to record possession of less than 40 percent. Italian side Udinese was the only one avoiding relegation by finishing 15th.
What happens when clubs are more evenly matched?
League’s competitiveness is influenced by team quality which is influenced by how much clubs can and will spend on players. It is expected that these variances in money fall away between the highest revenue generating clubs, such as the Premier League’s so-called Top Six.
In Top Six ties between 2021/22 and 2023/24, the home team won 48 percent of the time. While 30 percent of the matches ended in a draw. The home team recorded a higher possession in 53 percent of the matches. With the biggest difference coming in 2021/22 when City beat Arsenal (5-0) with 80.4 percent possession.
The team dominating possession won 41 percent of the time. While 29 percent of the matches were won by the team having less possession. The other 30 percent were draws.
City dominated possession in 90 percent of their matches against the Top Six between 2021/22 and 2023/24. At home they even dominated all matches. It did not always translate into wins, but they won most of the Top Six matches (53 percent).
Manchester United dominated possession in only four out of 30 Top Six matches during that period (13 percent). The least amongst the Top Six. Making no difference whether they played at Old Trafford or away. Yet, the Red Devils won 30 percent of their 30 Top Six ties. More than Chelsea (20 percent) and Spurs (23 percent), who each saw more of the ball.
Arsenal is the only other club who won a higher percentage of matches (50 percent) than in which they dominated possession (33 percent). When dominating possession, the Gunners won 60 percent of their matches (losing 30 percent). The best amongst the Top Six. Ahead of City, who won 52 percent of the matches in which they saw more of the ball.
Spurs won only 23 percent of the matches they dominated possession. Implying they were the worst amongst the six in translating possession into wins. Liverpool lost the least number of matches in which they dominated possession (15 percent), while winning 40 percent of them.
To dominate possession or not in a Champions League final…
The Champions League’s later stages could also be considered a setting in which teams are more evenly matched. Of the 13 finals played between 2011/12 and 2023/24, eight were won by the team dominating possession.5
Two finals were won by the team seeing the least of the ball. Liverpool beat Tottenham in 2018/19 (2-0) despite recording 39 percent possession. Whereas Chelsea bested Manchester City (1-0) with 42 percent possession in 2020/21.
Another two finals required penalties to determine the winner. While the 2021/22 final between Liverpool and Real Madrid (0-1) saw the teams share possession.
Ball possession and winning in Champions League finals between 2011/12 and 2023/245
| Club dominating possession… | How often? |
|---|---|
| Won | 8 |
| Drew | 2 |
| Lost | 2 |
| Note: in the 2021/22 final possession was divided equally | |
Possession in Top Six ties strongest related to difference in number of shots
Beyond wins, draws and losses, performance can also be measured through goals, shots and shots on target for and against and their differences. Especially in the long run.
Based on the Top Six ties across the three seasons, the strongest relation (correlation not causation) was recorded between possession and difference in number of shots (67 percent). With possession having a stronger negative correlation with shots conceded (64 percent) than the positive relation with shots for (51 percent). The relationship between possession and shots on target was weaker. Possession and goals conceded had the smallest relation amongst these metrics. While possession and goal difference had a correlation of 24 percent.
City having less possession
In recent years, Manchester City has topped many statistics, most of all possession. In 2024/25, they once again have the highest possession with 61.05 percent through the first 25 matches. However, it is lower than what they have recorded ever since Guardiola took over.
This downward trend was also visible in their Top Six away ties. Only in three out of five such ties in 2023/24 did City still dominate. With average possession dropping to 52.02 percent. Being less dominant in possession, key to their style of play, is certainly related to, if not causing, their under-par performance this season. Only once did Guardiola’s City lose more matches during a season than the eight, they have already lost so far.
Based on the first 25 matchdays, there is a correlation of 62 percent between possession percentage and goals scored. While this increases to 68 percent for possession and expected goals. Number of successful passes (whole field and final third) has an 87 percent correlation with goal difference. With secure passing having a stronger relation with goals conceded than with goals scored.
Are Nottingham Forest defying the statistics?
Yet, possession and other metrics do not always tell the whole story or reflect the table. There are outliers and other (not yet measurable) factors that play a part.
A good example is Nottingham Forest in 2024/25. After winning promotion from the Championship in 2021/22, Forest recorded a 16th and 17th place league finish. This season they defy expectations, and statistics, by having been ranked in the top four ever since they beat United on matchday 15.
Throughout the first 25 matchdays, Forest has had the lowest possession (39.76 percent). Which is in line with their possession in 2022/23 (37.6 percent) and 2023/24 (40.6 percent).
They have also recorded the least number of passes (8,481), with their completion rate (78.56 percent) ranking them in the bottom three.
With 2,908 final third passes (13th) they have made over 34 percent of their passes in the final third though (second behind City). The percentage of their passes going forward (34.29 percent) is the highest in the league. Which could be explained by a low defensive block creating more space to pass forward. City, on the contrary, have the lowest percentage of passes going forward (24.21 percent).
Forest scored 41 goals through their first 25 matches. Which is in the middle of the pack. Yet, they significantly outscored their expected goals (+7.97). Only Arsenal (9.84) outscored expectation by a greater margin. The number of goals conceded by Forest have been in line with expectations (29 vs 29.34).
Counterintuitively, their low possession has not resulted in many goals coming from counterattacks. Despite recording 32 shots from fast breaks (tied third), they only scored four such goals.
In general, it is successful teams who tend to dominate possession. Explainable, as successful teams will likely have more quality players who are better capable of retaining the ball. Simultaneously, they will have the quality to create chances and goals, while reducing the number of chances and goals conceded. Yet, there will always be outliers who defy statistics like possession and number of chances. Something that makes football what it is.
