The Boot Room: Data Analytics with Calvin Charlton

We aim to bring you closer to the club through a new monthly content series where you will get to know more about its goings on as well as reflecting back on times that have helped shape the club into what it is today. To kick us off, Edd Norval gives us an insight into our work with data analytics.

Football fans, managers and pundits can roughly be split into two camps. Those who truly believe in the benefits of data and an analytical, numbers-based approach, and those that don’t. It’s a dichotomy as old as time itself, the exploration of chaos and order. More specifically, it is about whether we believe that we can instil a sense of order onto something as chaotic as a game of football - where twenty-two different humans are trying to make the best decisions whilst physically and mentally exerted.

The rise and rise of data and analytics in sports can be attributed not to football, but baseball, where Billy Beane’s Oakland Athletics side outperformed the giants of the sport. Most of us will recognise this as Moneyball, either as book, subsequent film, or overriding concept. Many were sceptical in its wake, just as those in the boardroom were when Brad Pitt’s Beane first proposed the approach. We’re dealing with humans here, not machines - can we really reduce them to heatmaps and percentages? Should we really sign the player who has had a vintage season, yet vastly overperformed their xG when their performance, in all likelihood, will not be reproduced under different conditions in future seasons?

These are all questions and concepts that Calvin Charlton has asked himself since joining Hibernian six years ago, a young performance analyst having cut his teeth at Portsmouth and Southampton, with the goal of developing a league-leading department at the club. Over those years, with the accessibility of analytic tools and methodologies becoming more prominent, Charlton’s job at Hibs has shifted from seeking out what he could find to efficiently deciphering the plethora of information now available to analysts. At its foundation, Charlton must constantly assess whether each group of insights and calculations have the potential to provide tangible benefits on match-day or if they’re just a superfluous set of numbers.

The next breed of coaches are information hungry.

Calvin Charlton
Head of Performance & Recruitment Analysis

As Charlton told me, “The next breed of coaches are information hungry.” This is in large part to do with younger coaches having worked in this kind of environment during the tail-end of their playing careers. Being a performance analyst is a specialised role, a language that head coaches often aren’t as fluent in as those who work exclusively in that department. Because of this, another layer of articulation comes into the use of numbers. Beyond identifying their value, the findings must be communicated in such a way that a coach will be able to act upon them, either in training regimes, player selection or player recruitment and eventually, be something that will galvanise the player’s, not stump them. In a sense, Charlton is the buffer between man and machine, making arbitrary packages of numbers influence the way that players will perform.

Having said that, miracles are not expected. Numbers cannot tell a full story. A lot of Charlton’s work comes down to risk-management - namely mitigating as much of it as he can. When it comes to understanding performance in these terms, it makes sense, “Data is objective information, free from bias, that can help to support or challenge the decision-making process.” It might sound like a counterintuitive thing to say, but just as many errors can be attributed to think as to ‘not thinking’. Essentially, human error is inescapable, so the challenge becomes making everyone think better, to see more clearly. One way to do that is to circumnavigate the expansive web of complex processes continually firing around our brains by creating efficient pathways to follow.

Coaches do this by repetitive drilling, by only doing things that will impact the matchday performance. The physical aspect of football, literally what it is the body does, is habitual. So too is how players think. Infinitesimally more complex than the body is the mind. This very fact is exactly why sketching out what is ‘good’ thinking should theoretically benefit from being informed by as large a pool of data as possible - to succeed by the law of average that each player and team default to in various scenarios.

Building a portrait of each opponent, each set-piece or each counter-attack by what the numbers say is their most likely behaviour is a crucial part of optimally preparing for what you’re coming up against. You’ll never predict the future, never guess the unknown, but you can map out as much of it as possible by understanding what is most likely and what is the best way to adapt to that. These small details, these marginal gains, are what separates the team finishing third and the one finishing sixth.

It is exactly because the human body falls back on its habits that in-depth analysis can conjure up this ‘likeliest picture’ of how the opposing eleven human beings will act on matchday. Because of the variety in interpreting and implementing information, and the fact we are looking at people and not machines, Charlton is aware of the limits of relying too heavily on data, so rather than stand too far in Camp Numbers making faces at Camp Feelings, he puts a lot of consideration into subjective experience too, integrating that into the delivery and communication of his objective statistics.

There is a point in the game where these two camps coalesce, as Charlton articulates, “set pieces are probably where the old-school & new-school meet – both place high value on set piece efficiency – yet their methods and thought processes are drastically different.” In American sports, which are far more governed by set-piece type situations, thus are more ‘predictable’, a greater deal of emphasis can be placed on analytics. Football “is inherently more dynamic and open,” than sports like American football or basketball, “That’s why I believe you have seen a lot more emphasis placed on set-pieces over the last 10/15 years – because they can be controlled to a larger extent.” That’s also why it is crucial in football not to allow one too much more emphasis than the other.

Where analytics came into mainstream footballing parlance, several decades after the data-heavy approach to American sports, was with the xG (expected goal) metric, which illustrates the probability of a scoring opportunity that goes beyond a ‘chance’ or ‘shot at goal’ - which could be anything from a one-on-one straight into the keeper’s gloves to a 40-yard no-hoper. Both would be considered shots at goal, but there would be a difference in how they are graded on the expected goals model. Charlton thinks that xG “was one of the first statistics that offered contextualisation and a deeper insight to the wider population,” before adding that this has fed into the hunger for fans to learn more.

It’s not just the fans though, data analytics is still a ‘dark art’ to the young Englishman, something that he too continues to try and develop. Knowing and embracing the challenges up ahead, his role as Hibernian’s Head of Performance and Recruitment analysis has made one thing clear - “we need to be smart and innovative in the ways we try integrate data.” This is across the board, from the analysis of upcoming opponents to the way players are signed and trained, all part of an expanding department with a growing team, with latest recruit Euan Fotheringham being testament to the club’s will to continue on this trajectory.

Charlton is now working on establishing a tool within Hibs, particularly regarding talent recruitment, to “allow the decision-makers at the club more clarity and confidence in player trading, as well as providing more insight into less-known markets throughout Europe and the World.” Europe is a powerhouse in the use of data analytics in football. Between Scotland and England, from the top clubs all the way through the leagues, pioneering approaches are being taken to utilise information most efficiently. These nuances are what Charlton finds himself looking out for on a daily basis. What are we doing right? What are we doing wrong? How can we make it better? Who else is doing it better?

Although he won’t give away too much, Hibs are clearly on the right path, focussing time, energy and money into developing this alchemic discipline within the club, an aspect of football that is undeniably ingraining itself within the fabric of the game year-by-year - not only in how clubs operate, but also in how fans come to understand the game.

Many argue that analysts and technology will reduce football to a computer game where decisions are made by robots based on sophisticated algorithms. The reality is, that no matter how far the enhancement of technology and analytics go, football will never be 100% predictable and that’s what makes it the best sport in the world.

Calvin Charlton
Head of Performance & Recruitment Analysis

Those afraid of their beloved sport being consumed by mindless AI needn’t worry yet, as Charlton makes clear, “Many argue that analysts and technology will reduce football to a computer game where decisions are made by robots based on sophisticated algorithms. The reality is, that no matter how far the enhancement of technology and analytics go, football will never be 100% predictable and that’s what makes it the best sport in the world.” For a self-confessed nerd for numbers, he’s as taken by the unexplained in football as he is driven to try and explain it.