A statistical analysis of a managers' sartorial and tactical choices.

I am a big fan of the excellent stadio podcast, hosted by Musa Okwonga and Ryan Hunn. One subject that Ryan in particular enjoys to discuss is sartorial choices of managers, a topic recently brought to prominance with Julian Nagelsmann’s choice of suit in the Champions League semi-final.

Reuters

In a recent podcast Ryan said (and I am paraphrasing here) that there was no way to prove that an managers’ clothes actually make a difference, which I took as a challenge!

StatsBomb have an excellent open data resource, including data from all matches from the 2018 World Cup, which I chose as a decent starting point for my investigaton. My first task was to go through google images and match reports and classify each manager’s predominant outfit choice.

Outfit n
Polo 1
Shirt 6
Suit 20
T shirt 2
Tracksuit 3

We have a solid variety of outfits worn in the World Cup, with 5 distinct outfit categories. Honourable mentions go to Hervé Renard for combining his Jamie Lannister looks with a (very) tight shirt and jeans, Sir Gareth for his waistcoast (which has been classified as a suit) and Gernot Rohr for repping a polo all tournament.

Simon Hofmann - via Getty Images

The first cliché around managers’ clothing and their tactical preferences is the age old argument that suited and booted managers play more possession based football than those that prefer to don a tracksuit. Indeed this is so ingrained that it even pops up in Football Manager press conferences! Early era Pep vs every era Tony Pulis, for example. Lets take a look at the numbers…

It seems that this trend does holds some weight. Teams whose managers wear suits (which for this analysis has to include a jacket, tie is optional) complete more passes than other sides. In fact, the inclusion of a jacket over just a shirt leads to almost 200 extra completed passes per game, so you really should soldier on with a blazer despite the weather! Teams managed by a tracksuit manager seem to favour route one football, completing on average almost half the passes that a suited managers’ side would. But we all know that possession doesn’t mean anything if you don’t create chances, I’m looking at you here Quique.

xG, or expected goals, is an excellent barometer of how many chances a team creates, and on the quality of those chances.

This follows broadly the same pattern, with some key differences to note. Firstly, t shirt managers seem to be guilty of the aforementioned offense: lots of possession, no chances created. This is primarily down to Jogi Löw’s interesting tactical decisions in this tournament, although Sampaoli’s Argentina side suffered from the same issue, so perhaps the t-shirt was to blame all along. The stacked bar chart shows that coaches that favour a tee also don’t permit any high balls, whereas all other teams have a broadly similar passing profile.

Secondly, despite sharing a similar completed pass rate with Polo shirted managers, the teams managed by tracksuited coaches produce a significantly higher number of quality chances. This may be down to the gegenpressing philosophy of the finest tracksuit manager we’ve seen in recent years, Jürgen Klopp.

That checks out too! Tracksuit managers’ teams press significantly more than any others. You would expect more pressure further into the oppositions’ third from a Klopp team, but I think that this at least proves the hypothesis that tracksuited managers demand more pressing from their team. The next cliché that automatically jumps to my mind regarding tracksuit managers is reliance on set pieces, perhaps down to being a 24 year old QPR fan fondly watching highlights of us under Warnock…

This one does not match up with my expectations! xGA relates to the xG value of the shot that follows a pass, so this plot shows the reliance of each team on set pieces for creating chances, and it seems that the more formally dressed managers’ teams actually create a higher proportion of their chances through set pieces - a triumph for the love train?

EPA

There are so many intricacies to look at on this subject, for example how does the presence of a tie impact the suited managers statistics - is the impact as great as that of a jacket? But in terms of the initial analysis of the 2018 World Cup I think we can determine that there are clear trends (in the admittably small sample size) on xG, passes completed and pressing intensity. I hope you enjoyed this article, and I hope to do more analysis on whatever subject takes my fancy (as long as I can find open data on it)!

Please let me know if you have any questions, or if there are any other (no matter how bizarre) theories you may have that I can investigate through the power of stats, through my twitter or the comments below.

Robbie Still
Robbie Still
Data Scientist

I am a data scientist, I work in ecology and enjoy looking at football statistics in my spare time, mainly in R. I also love reading and watching anything fanasty, and I currently live in Brighton with my girlfriend, Coral.

comments powered by Disqus

Related