Case study: spreadsheet scouting

When writing Soccermatics I spoke to several analysts and scouts in professional football who were starting to get to grips with using numbers. At that time, one of the players who was seen as a data-scouted player was N’Golo Kanté.

In the 2014–15 season, N’Golo Kanté topped the table of successful tackles per game for the whole of Europe. He was then playing for French Ligue 1 side Caen. Sending a scout to check him out was a no-brainer, and lots of clubs did exactly that. It was Steve Walsh at Leicester City who was the most insistent with his manager about signing Kanté. Leicester bought him for a reported £5.6 million, won the Premier League, and sold him on to Chelsea for £32 million the next year. The tackle stat was simple, but it didn’t lie.

One analyst I talked to described this as ‘spreadsheet scouting’. The scouts have large spreadsheets where the columns include tackles made, interceptions, passes and dribbles. They sort the spreadsheets by the column they are most interested in. Row A shows the best player, Row B the second best, and so on. It is from there that they start their search. When Steve Walsh took over as director of football at Everton he signed Row B on his spreadsheet – Idrissa Gueye from Aston Villa. Like Kanté, Gueye also came from France to the Premiership in 2015, and during his first season ranked second only to Kanté in tackles and interceptions (Gueye later transferred to PSG and in 2022 is back at Everton).

Teams don’t rely solely on spreadsheets to buy players. All professional scouts agree that it is important to see a player in action before signing them. Footage of matches is valuable, but there is no substitute to sitting at the edge of the pitch and watching how a player reacts to the movement of play around him, and how he responds to teammates. A proper evaluation involves visits to several matches and, if possible, the opportu- nity to speak to the player and watch him train. The real question for a club is about getting a good balance between initial statistical screening, the use of scouting networks, video analyses, watching matches live, and using data to double check decision-making. Accounting for all of these inputs makes the job of picking players difficult.

The difference in outcome for Aston Villa and Leicester City during the 2015–16 season is a perfect illustra- tion of this point. Both teams followed the same principles of initiating a search using statistics. Both of them found under- valued players in France during 2014–15. Leicester signed Kanté and Riyad Mahrez and won the Premier League for the first time in their history. Villa signed Gueye and three other players from the French league, but were relegated with their lowest-ever points total. The biggest difference between Leicester and Villa appears to lie in the mutual trust between the coaching staff and the analysts. While Leicester had found a way to integrate statis- tics into all aspects of their operations, the recruitment staff of Aston Villa and manager Tim Sherwood could not agree on how the stats should be used.

When I spoke to Rory Campbell, technical scout and analyst at West Ham, he emphasised the need for a club-wide analytics-driven strategy that combines all aspects of player assessment: from statistical analysis, through an understanding of personality and attitude, to seeing the players’ strengths and weaknesses on the pitch. Rory’s own background personifies this combination. As a schoolboy, he played for Arsenal until he was 16. He then went to Oxford University where he studied philosophy, politics and economics during the day and was a successful high-stakes poker player by night. After university, he took his coaching badges, working his way up from training his old secondary school team through jobs at Barnet and Cardiff.

These experiences have allowed Rory to understand how to combine football, analytics and psychology in decision- making. ‘Whether its poker, economics, betting or football, it’s about using all the information required to make correct decisions under pressure,’ he told me.

Graduates from top-flight universities are becoming more common at club employees. Henry Newman studied philos- ophy and economics at the London School of Economics. He used spare time during his studies to learn everything he could about football, shadowing managers and taking all of his coaching qualifications. Since then Henry has worked in first-team and academy roles for Barnet, Charlton, Brentford and West Ham. He told me ‘I can’t say I have taken anything directly from my academic education and applied it in foot- ball, but it has shaped everything about how I approach the game.’

Rory and Henry have brought their analytical way of thinking to the clubs they have worked for. But it is their experience in all aspects of the game that allows them to apply this way of thinking.

Rory believes that many clubs still have a long way to go before they are properly exploiting data. ‘The whole process of recruitment is the wrong way round at some clubs,’ he told me. “Recruitment is often initiated by an agent and the clubs are then simply reacting to that. Of course agents are required, but clubs should work internally to identify players.’

Far too often, agents approach the club with names of play- ers who could be interested in joining them, and then there is a discussion between data analysts, video analysts, trainers, the manager and the chairman about the qualities of the proposed player. At West Ham, Rory wants to be proactive instead. The club should decide on a style of play, and then set up their scouting system around that long-term strategy. To achieve this, Rory is developing algorithms that automatically iden- tify candidate players who fit the club’s long-term goals. By discussing the results of this analysis with scouts and agents, in footballing language, he expects a consensus to build up around players that is based both on statistics and conventional methods.

In Major League Soccer (MLS) in the United States and Canada, the process of integrating analytics into scouting has gone further than in Europe. MLS teams tend to be open to new ideas, especially if they come from other major American sports. Atlanta United are an interesting case study, because they have been built from scratch over the last five years, with the aim of starting playing in the MLS during the 2017 sea- son.

Lucy Rushton, Head of Technical Recruitment at Atlanta, has used statistical analysis throughout the process of building up the team. The coaching staff decides together what type of players they would like to recruit. Sometimes the process will start with Lucy performing a statistical search through the data for a player that matches requirements, and then a scout will be sent to look at the player. In other cases, a scout will have seen a player and Lucy will create a video montage of all of the actions performed by that player in a specific situation. At Atlanta, and many other MLS clubs, there is no division between stats and scouting.

Read more in Soccermatics: Pro Edition