Are Asian Teams More Disciplined?
25/06/10 12:01
As a game fought on a genuinely global scale, football, like no other sport, tends to invite forays into cultural stereotyping. We've all heard them: the English are blue-collar tough; the Italians wily and erratic; the Brazilians rhythmic and carefree; the Germans mechanical and precise. (Which leaves me asking: what exactly are the Americans -- young, naive and idealistic?) Perhaps most trenchant are those about Asian teams and players. Disciplined, hard working, organized. I recall the always lively Jon Champion last season referring to Park Ji Sung -- in the most complimentary way -- as being "so valuable" to Sir Alex Ferguson because he is "so programmable." The same types of adjectives are now being used by English and American commentators in South Africa to describe the North and South Koreans. One heard just this morning a post-game analysis of Japan's landmark victory against the Danes as being a "disciplined" performance.
Can common statistics shed any light on whether these stereotypes hold up? To explore this, I gathered data from Opta (including games through 6/24/10) on a handful of measures that could, in a rough sense, constitute football "discipline". On the offensive side, there is shot accuracy and passing accuracy. On the defensive side, there is a disciplinary points per game measure (yellow and red cards) as well as a tackling accuracy measure. Presumably, if a team is more disciplined on the offensive side, they neither fling shots aimlessly at goal nor spray errant, risky passes. Moreover, a "disciplined" team, I think most would agree, does not make what the English refer to as "reckless challenges" nor do they find themselves getting booked.
The figure below plots the two offensive measures. Here there are some notable differences between the Asian teams. While Japan appears to be an accurate shooting side, their passing accuracy is mediocre; the North Koreans are particularly poor in front of goal. By contrast, it is the Argentines, Brazilians and Dutch who appear to be the most offensively precise, with Germany not too far behind.
And what about the defensive measures? Here we may see something: relative to their amount of errant tackles, Asian teams have not been booked as often. This may tell us two things: 1) when Asian teams miss a tackle, it usually doesn't as often result in a bookable offense; 2) Asian teams are less often booked for offenses unrelated to the "reckless challenge". Now, if you look at the bottom right corner of the graph, it is Span that distinguishes itself here in terms of its defensive discipline. But North Korea is at least in the ballpark -- perhaps warranting the disciplined and organized label it received after their match against Brazil. (Obviously a 7-0 loss to Portugal doesn't do much for one's reputation in this regard.)
In the plot below, I've summarized these findings, adding a fifth measure (goal conversion) and lumping the teams into their respective regions. Along with the clear indication of greater rule discipline, we also see that Asian teams tend to take their chances a bit better than most others aside from the Eastern Europeans. In other words, they have, to date, been more efficient when given clear scoring opportunities. There are some other interesting findings here: the African teams have accurate passers and woeful shooters; the Western Europeans precise tacklers and shooters; the Oceanic teams (notably Australia in that match against Germany) tend to enjoy bookings.
At this point, the hypothesis would seem to be bearing some fruit: Asian teams appear more disciplined in terms of fouls and more efficient in terms of shooting. But can quantitative information go further into helping to explain the cultural differences that are so often alleged? This time, the Castrol Football data offers a response. Consider three stylistic/tactical measures: 1) number of solo runs and 2) number of long balls (as measures of individualism vs. team play); 3) amount of distance covered (as a measure of work rate or effort).
The solo runs measure (plotted below as "y", with number of attacks serving as the predictor "x") is quite illustrative. While we see the Brazilians in their expected position, joined by the Argentines (and note too the Germans), both the South Koreans and Japanese are far below the number of individual acts we would expect given the number of attacks mounted by their sides. Japan seems particularly unique here, with a clear reliance on team play during the build up.
This is revealed further in Castrol's data on passes, which distinguishes between balls played "long," "medium" and "short." Compared to other countries -- particularly the Brazilians, Argentines and Spanish, all of the Asian teams play far fewer long and medium balls. Japan and North Korea appear similar here, although the DPRK's conservative posture is evident by the larger number of short balls.
Finally, we come to work rate. Castrol, usefully offers two metrics here: distance covered (in km) in possession and in defense. Overall, North Korea and Japan rank #1 and #3 in terms of defensive territory covered and Japan ranks #2 among the 32 teams in terms of total territory covered. This is behind the higher mean km for Asian teams as we see below.
But what happens when we plot distance covered in possession as a function of distance covered not in possession? In other words, can we predict the amount of work teams do in attack by the amount of work they do in defending -- thus, giving us some ability to distinguish those teams that work on both sides of ball? When we do, what becomes apparent is that the Japanese and North Korean work ethic is really rooted in a defensive tactical posture, rather than sheer discipline per se. It is the Portuguese and Australians that seem to be working the hardest on both sides of the ball; among the Asian sides, South Korea would be the overachiever. Given what we would expect from their defensive workrate, the Germans, Mexicans and Serbians also seem to be covering a decent amount of territory as well.
Taken together, it seems as if data can, indeed, offer some substance behind the colorful adjectives commentators use to describe the styles of nations (even if we remained concerned about the broader impact of stereotyping.) The North Koreans were indeed hard working on the defensive end and conservative, where Japan has been living up to its reputation as being both group-oriented and, thus far, efficient in front of goal ("clinical" in football parlance). South Korea, the Asian side with the strongest pedigree, is roughly comparable in terms of discipline, build up and overall team play. Moreover, Asian sides appear in the data as nearly polar opposites to South Americans -- particularly the traditional powerhouses. Brazil and Argentina -- and look to be rather different from African teams as well. Japan fared well against Cameroon; South Korea drew against Nigeria. The second round matchups, Japan v. Paraguay and South Korea v. Uruguay, thus offer intriguing possibilities.
And what of the Germans? Thus far, neither their disciplinary records nor accuracy would suggest anything out of the ordinary. They have, however, shown an interesting combination of hard work and individualism that suggests a style that could match up comparably with the top sides from South America and Spain, should they get past the English.
Can common statistics shed any light on whether these stereotypes hold up? To explore this, I gathered data from Opta (including games through 6/24/10) on a handful of measures that could, in a rough sense, constitute football "discipline". On the offensive side, there is shot accuracy and passing accuracy. On the defensive side, there is a disciplinary points per game measure (yellow and red cards) as well as a tackling accuracy measure. Presumably, if a team is more disciplined on the offensive side, they neither fling shots aimlessly at goal nor spray errant, risky passes. Moreover, a "disciplined" team, I think most would agree, does not make what the English refer to as "reckless challenges" nor do they find themselves getting booked.
The figure below plots the two offensive measures. Here there are some notable differences between the Asian teams. While Japan appears to be an accurate shooting side, their passing accuracy is mediocre; the North Koreans are particularly poor in front of goal. By contrast, it is the Argentines, Brazilians and Dutch who appear to be the most offensively precise, with Germany not too far behind.
And what about the defensive measures? Here we may see something: relative to their amount of errant tackles, Asian teams have not been booked as often. This may tell us two things: 1) when Asian teams miss a tackle, it usually doesn't as often result in a bookable offense; 2) Asian teams are less often booked for offenses unrelated to the "reckless challenge". Now, if you look at the bottom right corner of the graph, it is Span that distinguishes itself here in terms of its defensive discipline. But North Korea is at least in the ballpark -- perhaps warranting the disciplined and organized label it received after their match against Brazil. (Obviously a 7-0 loss to Portugal doesn't do much for one's reputation in this regard.)
In the plot below, I've summarized these findings, adding a fifth measure (goal conversion) and lumping the teams into their respective regions. Along with the clear indication of greater rule discipline, we also see that Asian teams tend to take their chances a bit better than most others aside from the Eastern Europeans. In other words, they have, to date, been more efficient when given clear scoring opportunities. There are some other interesting findings here: the African teams have accurate passers and woeful shooters; the Western Europeans precise tacklers and shooters; the Oceanic teams (notably Australia in that match against Germany) tend to enjoy bookings.
At this point, the hypothesis would seem to be bearing some fruit: Asian teams appear more disciplined in terms of fouls and more efficient in terms of shooting. But can quantitative information go further into helping to explain the cultural differences that are so often alleged? This time, the Castrol Football data offers a response. Consider three stylistic/tactical measures: 1) number of solo runs and 2) number of long balls (as measures of individualism vs. team play); 3) amount of distance covered (as a measure of work rate or effort).
The solo runs measure (plotted below as "y", with number of attacks serving as the predictor "x") is quite illustrative. While we see the Brazilians in their expected position, joined by the Argentines (and note too the Germans), both the South Koreans and Japanese are far below the number of individual acts we would expect given the number of attacks mounted by their sides. Japan seems particularly unique here, with a clear reliance on team play during the build up.
This is revealed further in Castrol's data on passes, which distinguishes between balls played "long," "medium" and "short." Compared to other countries -- particularly the Brazilians, Argentines and Spanish, all of the Asian teams play far fewer long and medium balls. Japan and North Korea appear similar here, although the DPRK's conservative posture is evident by the larger number of short balls.
Finally, we come to work rate. Castrol, usefully offers two metrics here: distance covered (in km) in possession and in defense. Overall, North Korea and Japan rank #1 and #3 in terms of defensive territory covered and Japan ranks #2 among the 32 teams in terms of total territory covered. This is behind the higher mean km for Asian teams as we see below.
But what happens when we plot distance covered in possession as a function of distance covered not in possession? In other words, can we predict the amount of work teams do in attack by the amount of work they do in defending -- thus, giving us some ability to distinguish those teams that work on both sides of ball? When we do, what becomes apparent is that the Japanese and North Korean work ethic is really rooted in a defensive tactical posture, rather than sheer discipline per se. It is the Portuguese and Australians that seem to be working the hardest on both sides of the ball; among the Asian sides, South Korea would be the overachiever. Given what we would expect from their defensive workrate, the Germans, Mexicans and Serbians also seem to be covering a decent amount of territory as well.
Taken together, it seems as if data can, indeed, offer some substance behind the colorful adjectives commentators use to describe the styles of nations (even if we remained concerned about the broader impact of stereotyping.) The North Koreans were indeed hard working on the defensive end and conservative, where Japan has been living up to its reputation as being both group-oriented and, thus far, efficient in front of goal ("clinical" in football parlance). South Korea, the Asian side with the strongest pedigree, is roughly comparable in terms of discipline, build up and overall team play. Moreover, Asian sides appear in the data as nearly polar opposites to South Americans -- particularly the traditional powerhouses. Brazil and Argentina -- and look to be rather different from African teams as well. Japan fared well against Cameroon; South Korea drew against Nigeria. The second round matchups, Japan v. Paraguay and South Korea v. Uruguay, thus offer intriguing possibilities.
And what of the Germans? Thus far, neither their disciplinary records nor accuracy would suggest anything out of the ordinary. They have, however, shown an interesting combination of hard work and individualism that suggests a style that could match up comparably with the top sides from South America and Spain, should they get past the English.
England v. USA & What Compels Us about Soccer from a Possessionist Perspective
17/06/10 17:30
Rob Green did not lose the match for England. Rather, England should have had a second. That is, if one is not Jose Mourinho or Helenio Herrera -- and believes, as many of us seem to, that possession and positive outcomes in football are positively linked.
The terrific tools from Scoregrid give us a unique view: showing how possession changes over time. (They also offer possession by territory, which can give you a good sense of how well midfielders and defenders are doing their jobs). Here, I show all of the first set of group matches in South Africa that have, like England v. US, resulted in a draw. What comes immediately into view is considerable variation; contrary to the belief that sharing points is either boring or unsatisfying, a draw can represent many different approaches to a game, some of which might be highly satisfying from an armchair point of view.
In the case that we are interested in here, it shows pretty clearly that England asserted itself as the match wore on. But for some timely blocks and saves by Tim Howard, England might have walked away with the points. Yet, the more frustrated party would have to be Mexico -- who, as we see below, were putting on a Wengerian display of football throughout their match with South Africa, but somehow couldn't take an edge in the scoreline.
But consider other examples of how a draw unfolds. In some instances, like Portugal/Ivory Coast, we see an early surge -- and then 2nd half parity.
In others -- New Zealand/Slovakia, Uruguay/France, Italy/Paraguay -- we see parity throughout, with only minor blips in the action where one side seems to control the ball. These were the matches that did not exactly draw rave reviews from the commentators.
As football, like no other sport, is constantly being assessed for its entertainment/aesthetic value, a question naturally comes to mind: what type of match is the most enjoyable? Putting aside the obvious response (the one where your team wins/your enemy loses), one might wonder whether equal or near equal parity in possession is necessarily the type of football that we like to see. Certainly the teams that nearly always get talked about are those who dominate possession (Inter Milan is an exception, of course), yet doesn't it seem paradoxical that in watching competitive sport that we seem to prefer it when one team controls play throughout a match? Win (Brazil, Germany), lose (Spain--see below), or draw -- possession dominance, it would seem, tends to appeal to us -- in much the same way that we tend to like to watch the big teams win year after year after year.
The terrific tools from Scoregrid give us a unique view: showing how possession changes over time. (They also offer possession by territory, which can give you a good sense of how well midfielders and defenders are doing their jobs). Here, I show all of the first set of group matches in South Africa that have, like England v. US, resulted in a draw. What comes immediately into view is considerable variation; contrary to the belief that sharing points is either boring or unsatisfying, a draw can represent many different approaches to a game, some of which might be highly satisfying from an armchair point of view.
In the case that we are interested in here, it shows pretty clearly that England asserted itself as the match wore on. But for some timely blocks and saves by Tim Howard, England might have walked away with the points. Yet, the more frustrated party would have to be Mexico -- who, as we see below, were putting on a Wengerian display of football throughout their match with South Africa, but somehow couldn't take an edge in the scoreline.
But consider other examples of how a draw unfolds. In some instances, like Portugal/Ivory Coast, we see an early surge -- and then 2nd half parity.
In others -- New Zealand/Slovakia, Uruguay/France, Italy/Paraguay -- we see parity throughout, with only minor blips in the action where one side seems to control the ball. These were the matches that did not exactly draw rave reviews from the commentators.
As football, like no other sport, is constantly being assessed for its entertainment/aesthetic value, a question naturally comes to mind: what type of match is the most enjoyable? Putting aside the obvious response (the one where your team wins/your enemy loses), one might wonder whether equal or near equal parity in possession is necessarily the type of football that we like to see. Certainly the teams that nearly always get talked about are those who dominate possession (Inter Milan is an exception, of course), yet doesn't it seem paradoxical that in watching competitive sport that we seem to prefer it when one team controls play throughout a match? Win (Brazil, Germany), lose (Spain--see below), or draw -- possession dominance, it would seem, tends to appeal to us -- in much the same way that we tend to like to watch the big teams win year after year after year.
FQ Quote of the Day, French Edition
05/04/10 17:30
From Michel Platini,
Adam Wells, Football and Chess, Hardinge Simpole Publishing, 2007, 163The reason why football is loved so much is because it has no definite truth.
The Wenger Curve
16/02/10 11:52
Following a recent loss at Stamford Bridge, Arsene Wenger was quoted as saying
What I found interesting in the response was Wenger's suggestion that possession and success constitute a linear relationship -- that is, the more possession, the better the "demonstration of football". This is a crucial question because, particularly at the international levels, possession has always been considered key to winning.
Does this hold for league football as well? Using match data from the 2008-09 seasons in England, Spain, France, Italy and Germany (note: Bundesliga data are incomplete), we can see whether there is evidence to support Wenger's notion. Below, I've plotted two simple quadratic equations where x=possession (home/away) and y=success (home/away, where 2=win, 1=draw, 0=loss).
The EPL, it seems, is the only league where the hypothesis consistently fits. In Spain and Italy, the more a home team dominates the possession, the higher the likelihood it seems that they will drop points. For away clubs in Spain, success appears to come with more possession, but only once the 50% threshold is met. In France, possession and success appear not to be related at all.
As a second test, let us substitute goals for winning as the dependent variable. Put simply: does more time with the ball mean more goals -- a better "demonstration of football"?
Only in England. And for home teams in Spain. And, to a marginal degree, for away teams in Italy.
Wenger, thus, is borne out -- at least as this point. Given the tendencies of English club football, he had a right to expect a better result than 0-2 based on the possession data alone. The fact that Arsenal had a majority of the ball away at Chelsea is, in and of itself, an accomplishment. In all of 2008-09, Chelsea never lost the possession game at home, just one of two teams (Liverpool being the other) who were able to do so.
I'm continuing to compile data for past seasons as well as 2009-10 to dig deeper into the meaning and value of possession. Clearly, this is something that finds considerable variation across Europe and can help to provide a quick measure of differing styles of play.
Naturally, the quote sparked controversy in the English media -- as anything a foreign manager says or any claim of superiority always does. Arsenal, in his view, had dominated -- and thus deserved to win -- since they held the ball for 58% of the match.You have to congratulate Chelsea. We didn't get a demonstration of football but they were efficient and, at the end of the day, that is what has made the difference.
What I found interesting in the response was Wenger's suggestion that possession and success constitute a linear relationship -- that is, the more possession, the better the "demonstration of football". This is a crucial question because, particularly at the international levels, possession has always been considered key to winning.
Does this hold for league football as well? Using match data from the 2008-09 seasons in England, Spain, France, Italy and Germany (note: Bundesliga data are incomplete), we can see whether there is evidence to support Wenger's notion. Below, I've plotted two simple quadratic equations where x=possession (home/away) and y=success (home/away, where 2=win, 1=draw, 0=loss).
The EPL, it seems, is the only league where the hypothesis consistently fits. In Spain and Italy, the more a home team dominates the possession, the higher the likelihood it seems that they will drop points. For away clubs in Spain, success appears to come with more possession, but only once the 50% threshold is met. In France, possession and success appear not to be related at all.
As a second test, let us substitute goals for winning as the dependent variable. Put simply: does more time with the ball mean more goals -- a better "demonstration of football"?
Only in England. And for home teams in Spain. And, to a marginal degree, for away teams in Italy.
Wenger, thus, is borne out -- at least as this point. Given the tendencies of English club football, he had a right to expect a better result than 0-2 based on the possession data alone. The fact that Arsenal had a majority of the ball away at Chelsea is, in and of itself, an accomplishment. In all of 2008-09, Chelsea never lost the possession game at home, just one of two teams (Liverpool being the other) who were able to do so.
I'm continuing to compile data for past seasons as well as 2009-10 to dig deeper into the meaning and value of possession. Clearly, this is something that finds considerable variation across Europe and can help to provide a quick measure of differing styles of play.
FQ Quote of the Day, US History Edition
07/01/10 16:42
From a Wesleyan University official, writing around the start of WWI,
David Wangerin, Soccer in a Football World (Philadelphia: Temple University Press, 2006), 26.
[Soccer's] only desirable feature ... is that it can be played out of doors after the tennis courts and ball fields can no longer be used because of the weather. It is not a game that appeals to those of our students who are engaged in other forms of athletics. It is not scientific enough.
David Wangerin, Soccer in a Football World (Philadelphia: Temple University Press, 2006), 26.