Performance Analytics in Professional Football - Part 1

Monday 16 September 2013

The use of goal line technology has been widely debated by pretty much every news outlet, and following FIFA's commitment to modernise certain aspects of the game, it seems that common sense has finally prevailed.

One topic that receives little mainstream coverage, but will interest many fans, is how professional teams are beginning to introduce and adopt other forms of technology in the quest for marginal gains.

In a global perspective, the use of technology and data in sports is nothing new. Baseball has sabermetrics (I'd advise everyone to read Moneyball by Michael Lewis, don't just watch the movie) and the NFL has a dedicated website giving public access to their entire database.

But football (the one with it's origins in England where you actually kick the ball, not throw it) seems to lie in a murky state where the uptake of technology and data seems to have been comparatively slow.

The difference between football and many other sports is that football is one of the purer team sports, where the collaborative effort of a team normally effects the outcome of a game. In comparison, baseball can be broken down into clearly defined events, such as a pitch, a home run etc. Football is far more fluid than this, thus making it harder to isolate a specific player's contribution to the team. 
Using statistics from Opta, The Guardian Chalkboards are great examples of how player movement and interactions can be tracked through a game

Simple data is the start

Broadly dubbed “Soccer Analytics”, the game is in a period of transition where a rising proportion of clubs (to different degrees) are beginning to turn to third-party providers who can record, compare and present quantifiable information that will help them build a better picture.

Soccer Analytics is formed at two distinct levels: the individual player or the team as a whole. In essence, metrics can be used to look at the contribution from a particular player, or a number of data sets can be used together to look at the team as a collective. There are too many metrics to list on here, but they do cover a huge amount of information.

As I mentioned, integrating specific metrics into football analysis is not a brand new concept - certain data has been available for decades. Starting in 1950, an obscure RAF commander called Charles Reep, began to record and document every aspect of a game whilst watching his beloved Swindon Town, detailing events as discrete as the number of successful passes prior to a goal. Widely acclaimed to be the first 'soccer analyst', it was this data set that led Reep to hypothesise that the most effective tactic was the 'long ball game' after recognising that the majority of goals tended to stem from three or less successful passes.

Arsene Wenger, using his degree in Economics as a platform, is a staunch proponent on the effectiveness of data analysis, to the point where he has a propensity to solely rely on data to form an opinion when scouting a potential transfer target. 

Arsene's rational behind this is straight forward; using a vast array of data points on a group of transfer targets, such as comparing their 'goals to game' ratio or other defined metrics, Arsene believes he can accurately judge the current and potential ability of a player far more effectively than if he were to only watch him play. 

Under this model, if he is able to purchase a player at what he believes is the correct price, and subsequently his predictions of the player's ability turn out to be correct, he is essentially arbitraging the transfer market to exploit market inefficiencies caused by other clubs who have failed to realise a player's true ability/value.

Wenger, a proud 1974 graduate from the University of Strasbourg, is also a major advocate of employing statisticians to help break down Arsenal's matches into small, comparable metrics. According to Slate, he receives a bulky 60 page report filled with numbers after each game. 

Opta, the company behind numerous Twitter personalities such as OptaJoe, are one of the main providers of data to the industry. Boasting a number of lucrative contracts with teams including Chelsea, Opta's database is easily the most detailed and truly is a statistician's gold mine (as a point of reference, approximately 1,500 'events' are recorded in a typical football match). This short YouTube video gives a decent overview of how the data is collected by their analysts.

With access to all of this data and staff who have the capacity and knowledge to form correct judgements, clubs have the sudden ability to scout thousands of players across over 15 leagues. Indeed, it sounds not too dissimilar to Football Manager.

Last week the Harvard Business Review published an insightful article entitled “Ferguson's Formula” that looked at the managerial strategy adopted by Sir Alex Ferguson. Although the article was an excellent read, I found it puzzling that they failed to mention that Manchester Utd, under the direction of Sir Alex, were one of the leaders in the use of analytics.  

Throughout the duration of Sir Alex's tenure at Manchester United the game changed enormously. From the formation of the Premier League in 1992 to the imposition of transfer windows, Sir Alex has spoken publicly about the requirement of managers to adapt to compete. “Most people with my kind of track record don’t look to change. But I always felt I couldn’t afford not to change. We had to be successful—there was no other option for me—and I would explore any means of improving.“

Indeed, fresh-faced David Moyes is not a stranger to the use of Performance Analytics in the sport. Tucked away in the Liverpool suburb of Halewood is Everton's 55 acre training complex that houses their dedicated Analytics Department. Upon his arrival in 2002, Moyes instigated the first moves towards a completely separate management team that would have an input into team choices and training regimes.

Steve Brown, First Team Performance Analyst at the club, worked with providers Prozone and StatDNA to provide Moyes with "extensive information because he's so detailed, thorough and methodical in his work."

"At times we've used certain 'without possession' shapes which have negated the opposition" he said. "At times, the manager (Moyes) has come up with a specific system, making subtle positional changes which has then allowed us to negate the problem, or allowed us to capitalise [on it] -- it's those kind of intricacies which Moyes is brilliant at"


The Manchester United senior management team have recently procured patented video analysis software from a Sydney based company called Sportsec, at the cost of just over £150,000. With Moyes' tradition of extensive researching and seeking input from others at the club, it is becoming quite clear how his management style aligned so closely with that of his predecessor.

Prozone Animation: You'd be forgiven if you thought this was Football Manager

Like with most new concepts, there are always a group of sceptics who will be averse to adapting. Andre Villas-Boas, the current Tottenham manager who was previously charged with leading Chelsea's opposition scouting during Jose Mourinho's first stint as manager, refuses to use any form of data or technology to draw his conclusions. “I have never used ProZone,” he admitted. “You always have to be very, very careful with statistics. For me it's useless but it varies from coach to coach”.

Going forward

The application and effectiveness of metrics can only be judged according to the ability of the club to properly use this information. However, as clubs become more efficient in using the available data, it is likely to provide the biggest benefit for those who operate under a constrained budget. 

At the top end of the spectrum, a rich club can easily identify and accumulate a squad of great players as these players stand out most. "You don't need statistics to spot the real great players or the really bad ones. The trick is to take the players between those two extremes and identify which are the best ones," said the Match Analysis company president, Mark Brunkhart. 

For the 'less endowed' teams, the prospects of suddenly having an improved scouting infrastructure, and the potential of a marginal advantage in the transfer market, may be the difference between assembling a squad capable of competing and one that cannot. 


With this in mind, the power of metrics should not be diminished, but, in my opinion, it should be embraced as the next evolution in the game. With Financial Fair Play regulations now in place, the pressure on management to maximise the potential of their current squad and act prudently in the transfer market is greater. The majority of clubs will no longer be able to 'splash the cash' on an acquisition, only for the acquisition to flop.

Over the coming years I expect that you will read and see more data being used in reports and articles. A number of high profile publications, including the excellent 2013 book entitled "The Numbers Game", have begun to unravel the potential of data analytics and show clubs how they can integrate this information into a wide range of their managerial decisions.

Coming up in Part 2 we will look at specific applications of Performance Analytics - using Opta's vast research database and interviews with industry leaders, we will undertake a comprehensive analysis of exactly how your team is likely to be using the data available.

Don't forget to follow us on Twitter for first access to articles.

Over and out, JW.

6 comments

Adrian Thompson mod

Great article again, looking forward to part 2

Reply
Simon Lynton mod

How much does it cost clubs to get access to Opta, Infostrada etc? Can the smaller clubs afford it?

Reply

Hello and thank you very much

Reply

Very informative and well written piece. Great detail about my favourite sport, football!

Reply

Hello and thank you very much

Reply
Anonymous mod

part 2 plssss

Reply

Post a Comment