Possession Percentages and Game Outcomes

Discussion in 'Statistics and Analysis' started by EvanJ, Feb 21, 2010.

  1. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    www.flashscore.com has possession percentages for some leagues including Engalnd, Spain, Italy, Germany, France, Greece, and Switzerland. I'm interested in the following:

    1. Are games more likely to be draws when each club's possession % is closer to 50%?
    2. In games with winners, what percent of the time does the winning club have a majority of the possession?
    3. How often do games with large margins of victory (3 or more goals) go with a big possession % advantge for the winning club?
    4. When a club has a big possession % advantage how often does that club have a large margin of victory?
    5. What is the mean possession % for the club with the higher possession % overall?
    6. What is the mean possession % for the winning club if there is one?
    7. What is the mean possession % for the losing club if there is one?
    8. What is the mean possession % for the club with the higher possession % in draws?
    9. Do the answers to these questions vary significantly from league to league?

    I might be interested in other things also. I will post some findings later.

    If anybody knows a site with possession percentages for other good European leagues and/or big leagues in countries in other confederations like Mexico, Brazil, Argentina, and Japan please post.
     
  2. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    ESPN Soccernet and Flash Score have different possession percentages for the same games, so I'm abandoning my project.
     
  3. revelationx

    revelationx Member+

    Jun 5, 2006
    London
    Skysport display game stats including possession for many games
    http://live.skysports.com/ScoreCentre/football.html

    Possession is a largely direct consequence of successful passing. If you pass it well you will boss possession and if you will pass it badly your possession will go down. The Big 4 Prem clubs all boss possession and have by far the most completed passes in a season. Possession usually equates to scoring goals. at least from open play.
     
  4. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    You used England for an example. From the limited data I did before I deleted the file, the correlation between the home club's possession % and the home club's goal differential for the game was positive and strong (near 0.76) for the Premier League, but positive and moderate for Serie A (near 0.5) and positive but weak for La Liga (between 0.2 and 0.3). For today's Manchester United-West Ham United game, Manchester United had 63% of the possession according to Sky Sports, 60% according to FlashScore, and 54% according to ESPN Soccernet.
     
  5. RichardL

    RichardL BigSoccer Supporter

    May 2, 2001
    Berkshire
    Club:
    Reading FC
    Nat'l Team:
    England
    It also depends a lot on the style of play. A team playing a patient possession game will invariably have a better possession stat than a team attacking with pace, possibly prefering to play a percentage ball.

    My own club played a very deliberate possession game earlier in the season, and changed to a game based on breaking quickly from defence in the new year. I'd be surprised if we didn't pass the ball more earlier in the season, but we've certainly got better results since then.

    Possession stats do also vary from source to source. Some just look at passes made. Some look at successful passes. Some judge it by territorial possession - i.e. how much time the ball spends in each team's half.
     
  6. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    I could see there being separate statistics for possession percentage (what team's player had the ball) and ball location percentage (the percent of the game the ball spends in each half of the field) for both soccer and American Football.
     
  7. revelationx

    revelationx Member+

    Jun 5, 2006
    London
    All of the Big 4 play passing football and have more possession and more completed passes than the opposition in almost every match. If you look at the stats for the completed passes for the Prem season then these 4 teams will have most completed passes.

    A team in the Prem this season that plays a non-possession passing game is Aston Villa. Villa have played a game based on quick counter-attacks and set-pieces and have not scored or conceded many (until the battering at Chelsea). Wenger was condemned when he commented on this. It is true though. Portsmouth, the worst team in the Prem have completed more passes than Villa this season. Rafa's comments about Blackburn's style of football is backed up by the low number of completed passes by Blackburn. Stoke are well known for playing through set-pieces and this is backed up by them having the lowest number of completed passes.

    It is possible to get results by playing a more direct game, see Villa and Stoke but if you keep possession the other team cannot score.

    All this data can be seen here
    http://www.skysports.com/story/0,19528,11096_2705370,00.html

    PS. My Dad knows MacAnuff! :D
     
  8. ZeekLTK

    ZeekLTK Member

    Mar 5, 2004
    Michigan
    Nat'l Team:
    Norway
    Lots of possession is not important; it only matters what you do when you have it. If you can score right away there is no point in playing "keep away" just to have more possession. Last year Norway beat Germany 1-0 with like 22% possession. :D
     
  9. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    that's really all to be said. the only reasonable use of possession statistics is to quantify how the game is being played, not who's playing better. commentators who point out one team's 54% to 46% possession "advantage" deserve nothing but scorn. 60% shows how the game's being negotiated by both teams but can't be called domination per se. it's only when you get near 70% that a proper verdict can be reached, but at that point there's no need to count beans to gather one side's being passed off the pitch!

    territorial "advantage" is best shown not as a percentage, but as a line, the way canal+ does to show hauteur d'interception during french matches. it's a derivative indicator, but a valid and interesting one. still, it doesn't tell you anything you can't cotton on to by just watching the match.
     
  10. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    well, i imagine 22% possession... and at least one shining moment of blind luck!78/22 is more than just sterile keepaway!
     
  11. Elninho

    Elninho Member+

    Sacramento Republic FC
    United States
    Oct 30, 2000
    Sacramento, CA
    Club:
    Los Angeles Galaxy
    Nat'l Team:
    United States
    And I'll add to this: even if two sources both look at time in possession, there can be questions about when possession changes hands.

    Consider the extremely common scenario where a pass is too far ahead of a striker, and goes out for a goal kick. What do you do with the time between the ball going out of play, and the goalkeeper kicking it?

    Some sources might say possession changes hands when the ball goes over the endline. Others might say possession changes hands when the goal kick is taken. Still others might simply not count the time in between. Same events, three different ways it can be counted toward possession time.
     
  12. Kuppeb

    Kuppeb New Member

    May 24, 2010
    Princeton, NJ
    I think everyone is kind of looking past the point here. Regardless of what actually constitutes a change of possession or possession time, the affect that the stat has on the game is completely relative to the game. Someone simply put it earlier that it doesn't matter if you play keep away for 80 minutes, if the other team scores and you don't its over. This is a very rudementary way of putting it, but in a sense pretty accurate. The system and style of a soccer team has just as much influence on possession time as does their skill. A team might favor a contracted shape that falls back with an offensive advance in order to force a turnover and attack the depleted defensive third of the other team. That team could win a game 3-4 to nothing on no more than 10 shots, and in addition have a far less possession time. But you also need to look at the teams that take possession as a chance to setup their system, then you aren't looking at a game of keep away, but rather a building up of their attack. This is witnessed in a matchup like the US and Spain. Spain uses possession to open up gaps in the defensive, then attacks in a series of one touch slot passing. On the other hand the US sucks them in, then tries to force a quick turnover that allows them to cause havoc in transition with their speedy forwards. If Spain wins or loses their possession time will be around the same, if it is higher though, it there is a better chance that they were able to establish their style of play, while a lower number might indicate the US was able to disrupt it, which is their style. Basically, possession does matter, but you have to look at the numbers from team to team rather than establishing a standard to judge everyone by.
     
  13. NHFootyFan

    NHFootyFan Member

    May 28, 2007
    Outside Concord NH
    Club:
    New England Revolution
    I thought about how I'd like to study this, I just have NO idea how I'd go about doing it:

    Break possession down by:

    Location: Defensive / Central / Attacking third
    Time: 1st Half / 46-60 / 61-75 /75-90+
    Score: Lead >2 / Lead / Tied / Chase / Lose >2

    Then I'd like to start looking at things. Now if anyone has any actually practical ideas where to find the data or develop the data, it might be fun to try to see what emerges from that kind of analysis. But I don't know if it even exists.
     
  14. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    i can guarantee you no such stats exist or ever will unless you do it yourself.

    for your first criterion, location, a first substantial hindrance exists in that there is no marked division of the field in thirds, only in two halves. a second is how to count a pass that starts in one sector and arrives in another (or even crosses all three). a related and even finer indicator is sometimes illustrated by 2 lines graphically superimposed on the field showing hauteur d'interception (don't know a good english term, i've only ever seen it on french telecasts): the average point at which the team won possession of the ball. but this is only used to indicate the physiognomy of play at a certain moment, not at regular intervals and isn't collected as a statistic.

    there is no technical difficulty in the second: goals scored 15min/15min is a statistic commonly used (and very easy to collate using even the most rudimentary box scores). and often you see a possession stat onscreen for "last 15 min"... but never laid out by procrustean quarter-hours. again, the intention is only to give momentary insight, not generate categorical statistics.

    as to the third of your criteria, you are often informed what a team's record is after being led, or down by 2, or ahead at halftime, but as for possession in all that... even if data could be collected, it couldn't really give any insight. primo, at each end of your gauss curve you're going to get two populations: one of teams that logically bunker down when up by more than two and another of teams that are so superior that they are simply passing their opponents off the field. secundo, a tied score can mean different things to each opponent, depending on many different factors a draw can be a result ranging from great to disastrous.

    as i posted earlier, possession really only indicates style of play or the texture of a match in progress, it's of little use in judging the quality of a team.

    more generally, soccer is simply not a sport that lends itself to stat crunching the way baseball or american football does. but here's something for you if that's your cup of tea:

    http://www.cahiersdufootball.net/classement-en-relief.php?id_pays=FR&division=1

    you can find all kinds of french league standings sliced up just about anyway you like: who would have been champion if goal difference was the reference, if only home matches counted and they lasted only 75 minutes? three clicks and bob's your uncle! (for what that's worth!)
    much more interesting is an occasional feature they run called le statistique inutile (the useless stat). these are generally very derivative, concern often obscure criteria, but far from "useless" are usually mindbogglingly insightful. I'll try to dig some of them up for you when i have a bit more time.
     
  15. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    here you'll find one comparing where PSG and OL are in the standings minute by minute in the last 20'.

    http://www.cahiersdufootball.net/article.php?id=3487

    pretty telling stuff. sometimes you don't know what numbers are useful until you see the graphs they make!

    sorry all this is in french. one no longer needs la langue de molière for diplomacy, philosophy or travel, but you do need it to read les cahiers du football, and that's reason enough. by far the most astute, subtle and stimulating source for football analysis and discussion in any language or medium (their paper version just went under). everything else is for adolescents, retarded or otherwise, in comparison.
     
  16. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    here's one showing how loïc rémy, considered nice's star player is actually poison to his own team:
    http://www.cahiersdufootball.net/article.php?id=3304
    the guy who crunched the calculations doesn't understand it either, because the player is undeniably quality, but the numbers are as clear as the nose on your face, no matter how you look at them!
     
  17. guignol

    guignol Moderator
    Staff Member

    Apr 28, 2005
    mermoz-les-boss
    Club:
    Olympique Lyonnais
    Nat'l Team:
    France
    after going off on a serious tangent, here's something that should interest you (and not just me). since it does indeed concern possession, and since i'm a super nice guy (none nicer, ask anybody) and since it’s a slow day (is that the machines humming or just a fly buzzing?) i'll even translate it!

    http://www.cahiersdufootball.net/article.php?id=3243#sta

    today we will analyze the usefulness of ball possession in Ligue 1 season 2008-2009.

    The home team plays...

    1% of their matches with between 30% et 34% possession
    3% of their matches with between 35% et 39% possession
    13% of their matches with between 40% et 44% possession
    25% of their matches with between 45% et 49% possession
    28% of their matches with between 50% et 54% possession
    20% of their matches with between 55% et 59% possession
    8% of their matches with between 60% et 64% possession
    2% of their matches with between 65% et 69% possession
    Giving an average possession for the home team of 51%. Nothing noteworthy.

    The matches at either end of the spectrum were :

    Nantes who earned a 1-1 draw against OM on matchday 11 despite having only 33% possession, lowest figure for a home team in the season.

    Saint-Etienne, who on matchday 17 dominated Le Havre both in score (2-0) and in possession (70%), highest rate of possession in 2008-2009.

    On average one finds :
    50.7% possession for teams who win at home .
    51.2% possession for teams who draw.
    51.3% possession for teams who lose at home.

    Which tends to demonstrate that ball possession has no impact on the score.

    However, here are the teams sorted by possession over the first 37 matches of the season*:

    Marseille : 57.3%
    Bordeaux : 56.3%
    Saint-Etienne : 55.1%
    Lyon : 53.2%
    Lille : 52.4%
    Rennes : 50.1%
    Sochaux : 50.1%
    Lorient : 49.7%
    Caen : 49.7%
    Monaco : 49.4%
    Valenciennes : 48.8%
    Grenoble : 48.6%
    Nice : 48.3%
    Auxerre : 48.0%
    Toulouse : 47.9%
    Nancy : 47.3%
    Le Mans : 47.2%
    Paris SG : 47.1%
    Nantes : 45.0%
    Le Havre : 43.1%

    The clubs showing the greatest gap between their standing in possession and points are Saint-Étienne, Sochaux, Paris SG et Toulouse.

    Next class we will watch The Exorcist and read Proudhon, after which the subject of possession will hold no secrets for us.

    * my note: the conclusion being that there is a pronounced correlation between possession and final place in the standings (below). But which leads to the other? The only real value of the examination in my opinion would be as an aid in analyzing the play of the teams falling outside the norm.

    Bordeaux
    Marseille
    Lyon
    Toulouse
    Lille
    Paris Saint-Germain
    Rennes
    Auxerre
    Nice
    Lorient
    Monaco
    Valenciennes
    Grenoble
    Sochaux
    Nancy
    Le Mans
    Saint-Etienne
    Caen
    Nantes
    Le Havre
     
  18. revelationx

    revelationx Member+

    Jun 5, 2006
    London
    I previously posted this link about match analysis software.

    https://www.bigsoccer.com/forum/showthread.php?t=600399

    You could get this but it is really expensive. I would be surprised if possession stats were not measured as the position of the players and the ball is recorded at every moment.
     

Share This Page