Central Winger: Why possession stats are misleading
Possession is a statistic that is notoriously misleading. As suggested in last week's article, it has been found in MLS (and leagues abroad) that winning a game's possession battle can have little influence on actually winning that particular game. In light of some of these findings, this week's article focuses on ways to improve our understanding of this statistic.
Soccer games are supposed to be 90 minutes long. That's when the final whistle blows.
Depending upon how a particular game is being played, though, the amount of time that the ball is "active" can vary tremendously. For example, imagine two games – one relatively free-flowing game that has 80 minutes of active play, and the other a game plagued with ticky-tack fouls and anti-soccer tactics that sees the ball in play for just 60 minutes. These are not numbers pulled from a hat – many media outlets reported an average around 70 minutes of actual play during the 2010 World Cup.
Italians call it "The Right of the Weak" to destroy games against a superior opponent, and here's why: In these two extreme scenarios, teams that split possession 60-40 have incredibly different amounts of time on the ball. In the 80-minute game, possession is split 48 minutes to 32 minutes. In the 60-minute game, possession is split 36 minutes to 24 minutes.
Think a team like Barcelona would like an extra 12 minutes on the ball every game? At the very least, that's what they'll get if you don't slow the game down against them.
Another way to look at it: If Barça split posession 50-50 in a 70-minute game, they get 35 minutes on the ball. If they put in a dominant 60-40 performance in a 60-minute game, they get exactly one more minute of time controlling play.
|1||New York Red Bulls||6,624||50||194.8||132.5|
|2||Seattle Sounders FC||7,434||56||218.6||132.8|
|3||Sporting Kansas City||6,781||50||199.4||135.6|
|10||Real Salt Lake||6,920||44||203.5||157.3|
|15||New England Revolution||6,307||38||185.5||166.0|
|16||San Jose Earthquakes||6,911||40||203.3||172.8|
As you see in the table above, with info gleaned from the Opta Chalkboards, we'll then have to discuss possession a different way – by counting the amount of times a particular team possesses the ball.
When defining "possession" as a sequence of events by one team that included at least one attempted pass, the Sounders led the league with 7,434 (or 218.6 per game). Conversely, the New England Revolution landed on the bottom of the chart with 6,307 possessions (or 185.5 per game).
By adding goals scored, we can get a feel for the amount of possessions that it tends to take a particular team to score a goal. According to these statistics, New York Red Bulls were the most efficient, scoring a goal every 132.5 possessions on average. Vancouver, on the other hand, had the lowest possession efficiency, scoring at the rate of one goal every 190.9 possessions.
However, possession is not exclusively an offensive statistic. When winning, teams often use lengthy possessions to try and close out the game. The less possessions left, after all, the less likely your opponent is to score.
Thus, we have two different modes of possession: "attack-minded," when a team is actively searching for a goal, and "clock-killing," where a team is looking to grind an opponent down and protect a lead.
Leaving only "attack-minded" possessions, then, we can extract much more accurate offensive efficiency statistics.
|1||New York Red Bulls||4,885||37||132.0|
|4||Seattle Sounders FC||5,346||38||140.7|
|5||New England Revolution||4,763||32||148.8|
|9||Sporting Kansas City||4,919||31||158.7|
|15||Real Salt Lake||4,846||28||173.1|
|16||San Jose Earthquakes||5,363||30||178.8|
While these results are not stunningly different, there are some very interesting changes. For example, the New England Revolution jumped from 15th to 5th. I suppose this makes sense, as the Revs' main problem last year was defense. Their goals-against total was 32 percent higher than the league average while their goals-for total was only 16 percent lower.
Clearly, by digging into possession deeper than simple percentiles, we can begin to better understand this deeply flawed statistic.
Devin Pleuler is a computer science graduate from Wentworth Institute of Technology in Boston, where he played on the men's varsity team as a goalkeeper. He's certified as a coach through both the USSF and NSCAA, and writes the Central Winger analytics column for MLSsoccer.com.