The tweeted musings of Will Smith’s 17-year old son Jaden are a curious bunch. The Daily Nexus, the student newspaper of the University of California, Santa Barbara, has pondered whether the young rapper and actor’s tweets – which include the existential gem “How Can Mirrors Be Real If Our Eyes Aren’t Real” – are the work of a merely childish ‘Twitter Sensation’ or the thoughts of a ‘Potential Philosophy Professor.’
On the 7th of August 2013, the then 15-year old surpassed himself by asking: “If A Book Store Never Runs Out Of A Certain Book, Dose That Mean That Nobody Reads It, Or Everybody Reads It.” Misspellings and a rather irksome refusal to budge from capitalizing Every Single Word notwithstanding, I genuinely believe that Jaden may have been on to something with this. Just because two scenarios appear identical does not mean that their circumstances are. A book store never running out of a certain book, for example, may be either the result of rampant success – wherein everybody wants a copy, so the shop orders in more and more and thus never runs out – or of a dismal failure – if the original stock simply never shifts.
So what on earth has this got to do with goalkeeping analytics? Call me crazy, but Jaden may well have inadvertently hit the nail on the head regarding a particularly pressing problem facing the analysis of punching statistics. If a goalkeeper is recorded as having punched the ball, we do not know if it was a ‘good’ punch – such as may occur where a cross is impossible for the ‘keeper to catch, so punching the ball away is the best available option of alleviating defensive pressure – or a ‘bad’ punch – wherein a cross was eminently catchable, and possession is limply handed straight back to the opposition.
I think that’s enough about the underlying existential issues that surround quantitative analysis, with the answer to whether punching is a good or a bad thing a long way beyond the publicly available data that I use. Delving into the data at hand does, however, allow us to other questions. Do some goalkeepers routinely punch a higher percentage of crosses and balls into the box than others, and if so who are they? And what about shorter ‘keepers? Does a diminutive stature force a ‘keeper to punch a higher proportion of balls? It’s time to find out.
If we look at data for the 2014/15 Premier League season (sourced from Squawka – with Punch% calculated as Punches/(Punches+Claims+Failed Claims), and only goalkeepers to have played <1400 minutes analysed), it becomes clear that some ‘keepers certainly seem to like to punch more often than others. Although the vast majority punched on fewer than 15% of the occasions that they decided to deal with a ball into the box, with ten of the twenty-one goalkeepers somewhere between 9% and 14%, there are certainly some players – Simon Mignolet, Hugo Lloris and Joe Hart, I’m looking at you – who dramatically exceeded the mean rate of 14.2%.
This is, of course, just one season’s worth of data, and the Save% and Claims% returned by ‘keepers in any single season have been shown to be fairly unrepeatable. By spreading my net to Europe’s Top 5 Leagues (EPL, Bundesliga, La Liga, Serie A, Ligue 1), I have been able to compile a dataset going back three seasons in order to assess the repeatability of Punch%. There were 27 ‘keepers in this dataset who attempted to deal with a minimum of 100 balls into their box in both 2013/14 and 2014/15. When the Punch% of each season is plotted against the other in a scatter graph (2013/14 is on the x axis and 2014/15 on the y), and a regression line drawn, it is clear that Punch% appears to be a fairly repeatable stat (R² = 0.31) as far as goalkeepers are concerned – suggesting that those who punch a lot or a little in any given season are likely to have done so deliberately as a result of their personal preference or team tactics rather than just through random luck. Indeed, to draw on two examples of Premier League ‘keepers in the sample, Julian Speroni recorded a low Punch% in both 2013/14 and 2014/15, with Mignolet’s 2013/14 Punch% not quite as high as his 2014/15 effort, but still well in excess of 20%.
If we look, then, at data for ‘keepers in Europe’s Top 5 Leagues over the past three seasons, we may gain a strong indication of which ‘keepers – like Mignolet – appear to have something of a penchant for punching, and of those – like Speroni – who sit at the opposite pole. If we consider only ‘keepers to have played a minimum of 4050 minutes, the equivalent of 45 games, across the three seasons, it is clear that Mignolet, Lloris and Hart’s dedication to the art of punching as displayed in 2014/15 was not anomalous, with Lloris Europe’s ‘punchiest’ ‘keeper across the period.
The least ‘punchiest’ table, meanwhile, also contains a great deal of Premier League interest, with West Ham’s former custodian of the gloves, Jussi Jaaskelainen, apparently particularly averse to punching.
So now that we have established that Punch% is, as far as goalkeeper stats go, fairly repeatable, and also which ‘keepers appear to have continually displayed a strong preference either for or against punching, let’s go on to consider whether this is in anyway linked to their height. Whilst I established in my debut blog post that there is no link between Claims% and a goalkeeper’s height, perhaps shorter ‘keepers maintain a Claims% equal to their taller peers merely by virtue of punching more? If we assess the Height and Punch% of the 40 goalkeepers to have dealt with the most aerial balls across the past three seasons in Europe’s Top 5 Leagues (a sample wherein every ‘keeper has dealt with at least 260 balls), we may produce the following scatter plot.
Whilst the trendline indicates that shorter keepers may express a slight preference for punching, the R² value of 0.01 is pitifully small – suggesting that only 1% of variation in Punch% is accounted for by a goalkeeper’s height.
We may conclude, then, that punching a lot or a little is merely the personal preference of a goalkeeper rather than a random distribution from season to season, with a goalkeeper’s preference apparently unaffected by factors such as height.
The real question at hand now is whether Jaden Smith would be a Lloris or a Jaaskelainen. Hopefully he’ll make his views known soon – but in the meantime, I’d like to hear what you think. All feedback is welcomed, so let me know on Twitter @sam_jackson94, or in the comments below. Thanks for reading!
‘Hugo Lloris punches clear’ photo credit: Getty Images. http://www.gettyimages.co.uk/detail/news-photo/spurs-goalkeeper-hugo-lloris-punches-clear-from-nikica-news-photo/459518622