Should goal kicks and long balls be abandoned?

Are goal-kicks and long balls useless exercises?

At the outset, my apologies for this rather serious post. But it is football season, and I find it deeply disconcerting to see goalkeepers repeatedly kick the ball aimlessly in the general direction of the other half of the pitch. Don’t they know that the opposition is as likely to take control of the ball as a teammate? Granted that possession is lost far away from the danger zone, but isn’t retaining possession important for winning? Given the amount of money riding on football teams, why isn’t someone doing something about this? Or perhaps I have got this completely wrong? If only to feel less uncomfortable, considering the number of goal kicks and long balls one has to watch during the world cup, I set out to try and understand this phenomenon.

Unsurprisingly, some others have had doubts about the utility of the goal kick as well. One very theoretical analysis, based on data from the American Major League Soccer 2016 season, called for keepers to abandon the goal kick altogether (https://www.americansocceranalysis.com/home/2016/12/7/attention-goalkeepers-stop-kicking-those-long-balls). It would be more satisfying to understand this in tangible, mechanistic terms. So, to my mind, the questions that need to be answered are the following. One, how common are long balls in regular match play? Two, how often does a long ball result in loss of possession? Three, is possession important for scoring goals? And finally, are long balls common enough to have an impact on possession and goal scoring ability? I had to make do with the publically available data for the 2014 FIFA World Cup for my analysis. (https://www.whoscored.com)

Data were available for all the matches played by the 32 participating teams. There were no data on goal kicks per se, but I used the data on long balls instead. Strictly speaking, the goal kick is a variant of the long ball. Wikipedia uses terms such as “relatively aimless” and “bypassing the midfield” in characterizing a long ball. Though crosses were listed separately, it is possible that the data provided could have included even long, but precise passes under this category.

Long balls are common. There was one long ball for every 6 short passes in the 2014 World Cup. I used two metrics for assessing loss of possession from a long ball. The first was “long ball accuracy” (presumably meaning it reached the person it was directed at), and the second (possibly more relevant to the standard long ball and goal kick), “aerial ball won”. Consistent with intuition, an aerial ball was won on the average only 50% of the time. Even long ball accuracy was only slightly better at 58%. Compared to this, a short pass was accurate 85% of the time. Teams were evenly matched and possession averaged about 50%, but teams who controlled the ball less than 45% of the time during a match, score only 11% of all the open play goals (I did not include the goals resulting from set pieces for the analysis). Possession percent correlated significantly with the number of open play goals but explained only 20% of the variation in goals scored. Because the number of short passes and long balls varied widely between teams, I used the ratio of long balls to short passes for analysing the impact of long balls on possession and goals scored. This ratio strongly correlated with possession. The higher the ratio, the poorer was the possession. (Figure) So much so that, if a team plays one long ball for every 5 short passes, it runs the risk of getting its possession dangerously close to the 45% mark. However, there appeared to be no correlation between the ratio of long balls to short passes, and the number of goals scored in open play. In comparison, the number of short passes per game was strongly correlated with possession and explained over 75% of the variation in possession between teams.

This analysis confirms that possession is important for goal scoring during open play, and that the proportion of long balls to short passes adversely impacts the teams’ ability to retain possession. Counterintuitively however, long ball play continues to be used by the major football teams, perhaps because of the persisting influence of the “3-pass optimization rule” postulated by Charles Reep. (http://www.wsc.co.uk/index.php?option=com_content&view=article&id=2502/29&catid=)  However, his assertion that “over 80 per cent of goals result from moves of three passes or less” exemplifies the fallacious conclusions that are drawn from using the wrong denominator. The key question here is, what proportion of three-pass moves results in a goal, which, as others have pointed out, is not as dramatic, because three-pass moves are very common.

As I settle down to watch the Spain-Russia encounter, I am convinced that while there are many reasons why some teams score more goals, kicking the ball long and high, is probably not one of them.