Analyzing the Beautiful Game
Sports analytics, no matter the field’s renegade posturing, has now been around long enough to have its own pieces of conventional wisdom. Baseball’s cognoscenti know all about the primacy of on-base percentage over batting average, and they have also come to realize once-treasured strategies like bunting and stealing bases are best used sparingly. In basketball, the mid-range jump shot is slowly being phased out as an inefficient relic of antiquity. Spreadsheets are shaming football coaches into rolling the dice more often on fourth downs.
But for many American fans tuning into the World Cup, soccer’s nuggets of analytic insight remain as foreign as the game itself. There are set pieces to orchestrate, attacking strategies to plan, areas of the defense to exploit — and it isn’t always apparent which tactics are best. But analytics has clear advice on how to do some things right.
Soccer analytics is very much viewed as a discipline in its infancy. And the sport itself is often described as especially resistant to the pull of number-crunching, whether due to its fluid nature, its sportocratic establishment culture, or a fear that the un-sentimentality of data will rob the Beautiful Game of its celebrated elegance.
There’s not much truth to that. Off and on, people have been tracking relatively detailed soccer data in some form for more than six decades, up to and including the modern companies that exhaustively log every event on the pitch. (2)
The beautiful game has always been resistant to numerical representation, at least beyond the almighty scoreboard.
But that's changing. Slowly yet surely, numbers have crept into how we talk and think about soccer. Once, there were goals. Then assists came along. Now there are saves, clearances, blocks, interceptions, punches, punts, the always-controversial stat of "possession" and more.
All over the world there are analysts, managers, players and entire think-tanks devoted to decoding any given 90 minutes of soccer. At the Major League Soccer (MLS) for example, every touch is tracked by Opta. MLSsoccer.com boasts a pair of columns – OPTA Spotlight and Central Winger – devoted to diving into those reams of data and pulling meaningful numeric representation out every week.
Likewise, the teams themselves are in on the game, hiring data analysts to give them a leg up on the competition. We're hoping to pull the cover back a little bit, to give the fans a chance to understand where our understanding is going. And make no mistake: We're maintaining a healthy level of skepticism ourselves; it took baseball 100 years to realize batting average was less important than on-base percentage, after all. In order to take a truly scientific take, we need to be wary of our own biases and blindness.
There are still years and years of data that need collecting.
But we're getting there. We're learning. That's the point of science, after all. (3)
This literature review considers whether there are statistics that can prove a soccer team to be more successful than others by scoring more goals and winning games by answering the following questions:
Does increased possession equate to a better chance of winning?
Can statistics even be used to analyze the possibility of winning matches?
Does increased possession equate to a better chance of winning?
Possession has become one of the most contentious of the newly available range of soccer statistics that are regularly quoted in live match analysis. It is generally assumed that enjoying more possession is a positive, but does this assumption hold water, and to what degree should bettors use possession stats in handicapping a team’s ability to win games?
The success of sides that take a possession based approach – such as Barcelona at club level and Spain at international tournaments – coupled with the aesthetically pleasing brand of passing football practiced by these teams, has made domination of possession an aspiration for many teams.
In addition, Liverpool’s Brendan Rodgers’ often repeated quote that “if you can dominate the game with the ball, you have a 79% chance of winning” is also taken as an endorsement of possession based football.
The success enjoyed by a short passing, possession orientated approach is much more mixed than advocates of this approach imply
However, if we take a more measured view of recent competitions, the success enjoyed by a short passing, possession orientated approach is much more mixed than advocates of this approach imply.
Holders, Spain bowed out of the 2014 World Cup at the group stage, despite having over 60% of the possession in defeats to first the Netherlands and then Chile. Also their success at Euro 2012 owed much to a semi-final penalty shootout win against Portugal, a side that had enjoyed only a minority share of possession throughout the tournament.
Barcelona conquered Europe in 2008/09 and 2010/11 with tournament possession figures in the mid to high 60% and pass numbers averaging around 700 per game. But sandwiched in between were wins for Inter Milan, 45% overall possession in the competition, barely a third in the final and just 400 passes per game in 2009/10, followed by Chelsea’s 47% overall possession in 2011/12.
Numerous high profile head to head meetings have fallen to the side that shuns possession, most notably Chelsea’s aggregate win over Barcelona in 2011/12 gained with just 20% of the ball and Real Madrid’s 5-0 aggregate trouncing of a Pep Guardiola led Bayern Munich at the UCL semi-final stage this year with less than 30% overall possession.
Each of these examples is cherry picked and small in sample size, but they do appear to contradict Rodgers’ precise assumption, and the assertion that all a side needs to do to stand an excellent chance to win a match is to keep the ball for longer than their opponents.
The major flaw with raw possession as a useful indicator is that it is a secondary statistic that has been derived from a composite of many other primary stats that are much more fundamental to a team’s success.
Teams get possession by being good at making tackles or interceptions, they keep it by passing well and they use their possession to create chances. They score goals by converting these chances and they win games by scoring more goals that they allow.
So we have a chain with match result at one end time of possession at the other, but in between are statistics that are ability dependent – often score line dependent – which are better indicators of match success. Possession in itself is not a strong indication of how well a side has performed. Of more importance is what a side is able to achieve with the possession it has.
Swansea, in their early Premier League seasons, typified the confusion that simple possession figures can cause. The Welsh side were unconvincingly compared to Barcelona, primarily because their possession figures rivalled those of the Catalan side.
However, whereas Barcelona possessed the ball high up the field in their opponents final third and combined quick, intricate passing sequences – known as tika-taka – to create chances for world class players, Swansea’s use of possession was much more conservative, rooted firmly in their own half of the pitch, often comprising square or backward passes and designed as a defensive tactic centered around ball retention to prevent the opposition from scoring.
In the same 2011/12 season that Swansea were using the third best possession statistic in the EPL to create 472 chances, Barcelona were topping La Liga’s possession charts and creating 626 opportunities. The four sides immediately surrounding Swansea in the EPL possession table used similar levels of ball retention to create an average of 681 chances over the season, almost 50% higher. At the end of the season, Swansea finished a worthy, but mediocre 11th.
Swansea’s choice of how to use possession in a defensive, rather than attacking fashion was a result of their limitations as a recently promoted team. They protected a game start point in much the same way as more talented teams sometimes use possession to protect a lead.
Stoke City had travelled a similar route in 2008/09, with an inverted approach, this time by conceding possession and relying on chance creation from set pieces and long throws to generate their goal scoring opportunities, whilst making scoring difficult for their opponents with a packed defensive third.
More illustrious sides, notably those managed by Jose Mourinho, also appear willing to sacrifice the ball in return for defensive solidity and the chance of scoring from counter attacks or set plays, especially when faced with technically superior opponents.
A game is sometimes won by an opponent’s mistake. Errors, in Mourinho’s view are more likely to be made by the team in possession of the ball and he is therefore happy to give them the opportunity to present his teams with a gift and at the same time attempting to ensure that the game remains stalemated.
Mourinho’s Chelsea demonstrated this tactical aspect of possession when his side went to Anfield as underdogs in 2014, enjoyed just 27% of possession, but defeated Liverpool 2-0, with goals from a Gerrard mistake and a swift counter attack, effectively killing the Reds’ title ambitions.
Although happy to possess the ball and dictate the course of the game against inferior sides, Mourinho is prepared to be adaptable when faced with a sterner challenge. The likely possession statistics for the Liverpool-Chelsea meeting were virtually determined before a ball was kicked, and therefore, should have been anticipated by bettors.
Playing without the ball can often produce unattractive football and negative media reaction, but in many cases sides revert to these tactics to give themselves the best chance of achieving a favorable result.
Stoke City under Tony Pulis were technically inferior to many of their Premier League opponents, so an overtly defensive stance, both at home and away, coupled with long balls into the opponents half followed by brief bouts of pressing, was a pragmatic approach to maximize their chance of gaining points. Virtually every one of their 56 victories achieved under Pulis came with less than 50% of the possession.
Possession is only of minor interest, the real value lies with which primary events took place during those bouts of possession or non- possession
Possession is only of minor interest, the real value lies with which primary events took place during those bouts of possession or non- possession. How many chances did a side create or concede and just as importantly, were the chances of high quality or not? Since these types of statistics can be easily gathered and their correlation to goals and ultimately match result is strong, the need to quote possession diminishes and the temptation to try to draw conclusions merely from possession should be avoided.
The true indication of a team’s ability is how efficiently they use their possession in creating chances and how well they can stifle their opponent’s use of the ball when they have it. A side may out possess a side and even create more chances, but those scoring opportunities may be of poorer quality, due to their poor shot location and potentially higher levels of defensive pressure – something discussed in more detail here.
Chances created on the counter attack are converted at significantly higher rates than similar open play shots
Chelsea’s counter attack goal at Anfield and Ronaldo’s strike on the break for Real Madrid in the UCL second leg in Munich, easily beat an exposed goalkeeper. These goals from high quality counter attacks aren’t typical of all such chances, but using shot location data, it appears that chances created on the counter attack are converted at significantly higher rates than similar open play shots, which often involve slowly crafted build ups and allow the defense to become organized.
Therefore, it is wiser to attempt to draw conclusions from fundamental statistics such as shots, placed in their proper context, rather than assuming that dominant possession automatically indicates a superior side and any loss was unfortunate.
Low possession teams often have alternative ways of scoring. For example, in La Liga, the home of possession football, lower possession teams tend to score proportionally more of their goals from set pieces and counter attacks than do sides that have followed the Barcelona model.
This year’s champions, Atletico Madrid, scored a third of their goals from set pieces and counters, whilst enjoying 49% possession. That figure fell to 45% in the UEFA Champions League, where over half of their goals were either from counter attacks or set pieces. In the final they came within three minutes of defeating their cross city rivals, Real Madrid.
Possession figures tell you little of how effective a side may be, although possession based teams may also be good teams, creating an illusion of a general widespread correlation. It merely helps to inform you about how they tactically chose or were compelled to play the match. http://www.pinnaclesports.com/en/betting-articles/sport/possession-in-soccer Can statistics even be used to analyze the possibility of winning matches?

The graph below shows the relationship between average possession per game (along the bottom) and shots per game (along the side) for the 98 sides in the Premier League, Serie A, Ligue 1, La Liga and the Bundesliga.

As you might expect, there’s a fairly obvious correlation – the more possession you have, the more shots on goal you’re likely to attempt, which is hardly a revelation.
The graph is interesting, however, for two reasons. First, because there are clear differences between the five separate leagues. Second, because there’s a handful of sides that don’t fit the pattern, and a lot of variation amongst the sides who see a lot possession.
The sides who are significantly ‘higher’ on the graph compared to the line of best fit are particularly efficient with possession – they have more shots than you’d expect for the amount of the ball they enjoy. Those who are significantly ‘lower’ are less efficient – they see a lot of the ball but record relatively few shots on goal.
Of course, being more or less efficient is not necessarily ‘better’ – because the sole purpose of possession is not to score a goal. Possession can be used as a defensive tactic to play out time when a side is ahead, and can be used to tire the opposition, before attacking more directly later on. The intention here is not to ‘rank’ sides, but to show their different styles.
Here’s a similar graph breaking the 98 sides into their five respective leagues:

The most interesting pattern here is the Premier League, because so many sides are on the more direct side of things. Swansea, the black dot towards the bottom-right corner, are unique within the division, which explains why there’s been so much discussion of their patient passing game.
A line of best fit for each league also shows the differences:

But what about individual teams? Here, some sides are significantly higher or lower than what we’d consider standard. The most efficient sides average one shot from every 3% of possession, the least efficient teams need 5% to have an attempt.
Again, the purpose is not to indicate that a particular style is better – simply that there are big variations. Swansea see more possession than Manchester City, yet have six fewer shots per game. Manchester United have as many shots as Barcelona, despite 12% less possession per game.
In general, more possession will mean more shots – but it’s not always the case. http://www.zonalmarking.net/2012/05/04/the-relationship-between-possession-and-shots/ The statistic of pass completion % is one that I have discussed in the past as having limited relevance - see the post here for example. The key issue is that pass completion % does not tell you anything about who won the game or who scored the most goals, because its very situation specific (at certain times of the game a team may be ceding possession and allowing a team to complete a large number of non-threatening passes, for example). One key factor that pass completion % does not take into account at all is pass difficulty. Whether a team is banging the ball around in the defensive backfield or its making quick one-touch passes in traffic, pass completion % treats them all equally. About the best you might be able to do at the player level is to compare players at similar positions on this metric, but as we know, no two players (even in the same position) play exactly the same role.
We decided therefore to try to partially address the problem by running a regression to determine pass difficulty and then adjust passing skill based on the difficulty of passes attempted. We ran a regression on over 100,000 passes from the Brazil Serie A, and also looked at several subsamples (passes in attacking 1/3, only passes on the ground in the attacking 1/3) and all roads lead to one conclusion: after adjusting for difficulty, pass completion % is nearly equal among all players and teams. Said another way, the skill in executing pass is almost equal across all players and teams, as pass difficulty and pass completion % is nearly completely correlated.
Before summarizing or concluding any further, let's discuss a bit more the analysis that was done. We took completed pass as the dependent variable in a logistic regression that included the independent variables of level of defensive pressure on the passer, pass distance, direction passer is facing, whether the pass was one-timed, and if the pass was with the head or foot, and if the pass was hit on the ground or in the air. We also used the field zone of the next touch (whether the pass was complete or incomplete) as a proxy for the level of pressure on the recipient of the pass, because we know that defensive pressure tends to increase as you move up the field and towards the goal. We needed to do this because we can't really measure pressure on the intended recipient on an incomplete pass. All of the coefficients we tried were extremely significant and the regression had a very strong fit.

https://web.archive.org/web/20120119153242/http://blog.statdna.com/post/2011/05/04/-Differentiation-in-passing-skill-between-players-and-teams-is-non-existent.aspx

References http://sloanreview.mit.edu/article/what-businesses-can-learn-from-sports-analytics/ (1) http://fivethirtyeight.com/features/what-analytics-can-teach-us-about-the-beautiful-game/ (2) http://www.mlssoccer.com/numerology/news/article/2012/06/18/numerology-brave-new-world-lies-ahead-mls (3)