I’m quite sure that somewhere in a parallel universe of some sort, there are two equally impressive candidates preparing to duel. The manner of their contest, however, is more than a little unusual. This is not the FA Cup final. Nor is it Wimbledon. In this totally-not-made-up encounter, mental experiences are to be pitted against one another. Victory will be gained by whichever experience has most frequently been described by befuddled punters this world over as: “that’s so weird!!”
The top two seeds, frankly in a class of their own in this epic contest, are set to slug it out once again. This cognitive El Clasico sees déjà vu in the red corner, and the Baader-Meinhof phenomenon (a.k.a. the frequency illusion) in the blue. Both, I’m sure you agree, are truly bizarre and somewhat disconcerting to experience.
If the umpires of the tournament are as diligent as Michael Oliver (seriously, what a ref!), then, on my account, they’ll have chalked up a point for the blue corner a couple of weeks ago. On the evening of the 18th November, a tweet by Alex Olshansky found its way onto my Twitter timeline. It regarded a quote by Chile’s coach, Jorge Sampaoli.
Whilst some were left reeling by the sexist overtones of Sampaoli’s quote, I was in a state of shock courtesy of the witchcraft that is the Baader-Meinhof phenomenon. Despite Sampaoli’s being in charge of one of the most exciting international sides on the planet, I hadn’t actually properly registered his existence until he cropped up in a random article about Chile’s cracking Copa America campaign that I’d read that very same morning. And then, just a few hours later, there was his name again. Literally mental.
Once I’d got over this reminder of the complexity of the human brain and of the subjugation of women in the twenty-first century, Sampaoli’s metaphor was simple. Possession, apparently, doesn’t really matter for much in deciding outcomes in a football match. The international director of Sanchez, Vidal, Vargas et al. is not alone in having expressed these thoughts, as shown by the following ‘insights.’
Many analysts have politely begged to differ with these perspectives. On an individual game by game basis it is absolutely possible for a team to win having had much less of the ball, with two such instances influencing the afore-listed tweets. Yet, to again quote Olshansky (this time from an interesting 2014 StatsBomb piece on offensive and defensive possessions), “there is a relationship between possession % and points.” The Numbers Game makes a similar point (2014, p. 156), “It seems natural to assume that more possession should lead to more wins and fewer losses. And it’s quite right.”
Let’s now swerve back into the analytical field in which I predominantly operate – the world of goalkeepers. Teams that opt for a possession-based style obviously have a tendency to play the ball out from the back when possible. Those that opt for ‘Route One’ long-ball stuff, unsurprisingly, will instead be more likely to lump long passes forward. So, with goalkeepers plainly placed in a position of extreme responsibility within these two contrasting tactical models whenever the ball finds itself at their feet, it makes sense to compare keepers by their passing style. I have made an attempt to do this (though it took me aaages!) by counting every pass made by a goalkeeper within the Bundesliga, La Liga and the Premier League this season, as available on a game by game basis on the Statszone app and website. Statszone, following the Opta data by which they are powered, divide the pitch into thirds. These thirds, handily, may again be divided into three by the helpful pitch stripes. This gives 9 sectors, as follows:
By multiplying the percentage of each keeper’s pass destinations within each sector by the number of the sector, we end up with a number between 1 and 9 which represents the destination sector of a typical pass of theirs. If a keeper ridiculously opted to pass exclusively to the 1st sector, they would ‘score’ 1. If they only lumped long into the 9th sector (they’d need an absolute cannon of a boot and probably a hefty following wind to achieve this!) then they’d score 9. An even 9-way split, meanwhile, would result in a score of 5.
I’ve recently posted bar charts denoting the passing styles of these leagues’ keepers on Twitter, but here they are again (passing ‘scores’ determine the left to right order of the charts):
So what do these tell us regarding the value of possession, or lack thereof? Whilst the bar charts are interesting as intra-league comparisons of one keeper relative to another, we must revert to the passing ‘scores’ to determine who appears to be right, as far as keeper passing can be considered a proxy for team style, in the battle of Sampaoli, Ogden & The Mirror v Olshansky & The Numbers Game.
Utilising clubs’ ELO ratings (from clubelo.com) – a metric intended to be universal across leagues – as a measurement of team quality, and plotting keepers’ passing ‘scores’ against their side’s ELO indicates a reasonable relationship between short passing and quality. Whilst Barcelona’s Claudio Bravo and FC Bayern’s Manuel Neuer exemplify this, not one of the 11 clubs with an ELO <1800 has a keeper with a passing ‘score’ higher than 5.
Interestingly, the relationship is by far the strongest in the Bundesliga:
Of a somewhat intermediate level in La Liga:
And very weak in the Premier League, with table-topping Leicester City’s long-ball merchant Kasper Schmeichel doing his best to disrupt proceedings:
Although the Premier League thus appears, at first glance, to do a good job of restoring some credibility to the claims of the ‘possession means nothing’ brigade, with passing short apparently of little avail to Boruc, Guzan, Krul & Ruddy, it is notable that the league’s top 7 teams on ELO remain camped on the left-hand side of the chart.
At risk of causing déjà vu for those who have recently read/re-read Chapter 5 of The Numbers Game, it seems that, in general, better sides really do tend to prefer playing the ball out from the back.