League comparison #3: shooting and goals

January 30, 2010

In part one we looked at passing, in part two we looked at crossing. For the final instalment, we will look at what football is all about – goals.

Firstly, a simple statistic – the average number of goals per game in the five leagues:

The only major surprise here is that the Premiership is so far ahead of the other four leagues – and this is an anomoly – the Premiership is seeing a bizarrely high number of goals this season, as detailed here. Indeed, the average goals per game over previous seasons is generally around 2.6 – pretty much average for the other four leagues listed.

Onto the average number of shots per game:

This is the expected result, in that the goals-galore Premiership also sees by far the highest number of shots per game. Three leagues produce exactly the same shots per game score. But, if you’re wondering why the the Bundesliga sees far more goals than Ligue 1 despite the two leagues producing the same number of shots per game, that’ll be the shooting accuracy:

As you can see, 3% more Bundesliga shots are on target than Ligue 1 shots, which helps explain the fact that there are 0.26 goals more scored per game in Germany.In truth, there wasn’t a great deal of correlation between any of these three sets of data. The best way to find a vague correlation between these statistics is to work out the number of shots on target per game. This statistic can be found, obviously, by taking the shooting accuracy percentage of the number of shots per game. You can then compare the percentage of shots on target per game, to the number of goals per game:

And there’s a decent correlation there, although nothing more than you would reasonably expect. What explains Italy and Spain’s position to the right of the line of best fit? Better goalkeepers? More ‘on target, but not threatening’ shots? Using only five sets of data means the results here are highly unreliable, but we can vaguely come to the (perhaps inevitable) conclusion that the more shots you have on target, the more you’re likely to score.

The most startling correlation from all these sets of data on passing, crossing and shooting, however, is when you mix them, and plot percentage of long passes and goals scored per game:

Look at that correlation! Granted, there may be only five figures to plot, but there is an undeniable link between the percentage of long passes, and the number of goals scored.

But this, of course, is exactly the kind of analysis that led Charles Hughes to write the infamous The Winning Formula, which put forward the idea that the best way of scoring goals was to quickly get the ball into the ‘positions of maximum opportunity’ (ie, within shooting distance of the goal) – which ultimately resulted in the now much-maligned long ball game.

One should not, of course, see this graph and assume that the best way to score goals is to punt the ball long. That, in part, was the mistake many managers made having read Hughes’ work. Because this analysis is league-wide rather than looking at individual teams, the ‘goals scored’ figure increasing is not necessarily a positive thing for teams involved. Indeed, a single Premiership team, say, Sunderland, would not be any better off if the goals-per-game ratio in the Premiership suddenly rose to 4.5 – because they are just as likely to be conceding more goals as they are scoring them. Indeed, because a goal scored by one team is always conceded by another, it would be a classic zero-sum game.

A more specific, rigorous statistical exercise within the Premiership would surely not yield similar, pro-long ball results. An educated guess would be that Arsenal hit the ball long less often than any other side – and yet they are the side who have scored (by far) the most goals in the league this season.

This is not The Winning Formula. It is, however, an interesting insight into the vast differences between leagues, and prompts you to ask yourself what you’re looking for from football. Two things often held up as ‘ideals’ in football are (a) plenty of goals and (b) neat close passing. But these statistics insists that they are, at least on a league basis, inversely correlated. So what do you want: long balls and plenty of goals in England, or neat close passing but fewer in France or Italy?

Figures taken from OPTA.

Tags: , , , ,