How Much Should Data Influence Transfers?

The numbers only work when backed by contextual knowledge


In January, after months of poor results and worse performances under American coach Bob Bradley, Swansea City found themselves without a manager. Sitting second-bottom, the January transfer window represented a crucial period in the history of the Welsh club. A strong month and their Premier League status could (and eventually would) be retained, a poor month and relegation was a real possibility. As a response to the managerial confusion, the Swansea board appointed Dan Altman, founder of North Yard Analytics and ‘Moneyball Expert’, as a transfer consultant.

Swansea had been recruiting poorly for years, at least that’s what the numbers tell us, and Altman got to work immediately on correcting a misfiring team. Over the course of January, Swansea brought in Norwich’s Martin Olsson, Tottenham’s Tom Carroll, and Aston Villa’s Jordan Ayew - three players that had a positive impact on team performances almost immediately, and ultimately kept Swansea up. When you consider that two of the three were recruited from the second tier of English football, and Carroll was all but a reserve player at Tottenham, the genius behind the signings comes into sharper focus. Altman’s influence in these transfers shouldn’t be underestimated, and his focus on the numbers behind the players ultimately salvaged the club’s season.

Most analytics teams at the top level of sport will tell you that their clubs have a data-driven approach to player acquisition. Where scouts would be sent to form an ultimately subjective opinion on a player’s quality, managers are now presented with objective, cold data to inform their dealings in the market. Experienced scout Rob Mackenzie told Sky Sports: ‘Data is neutral, reliable and it allows you to assess a significant number of players in a time-efficient manner. It provides a platform for you to identify players and benchmark expected performance levels. It also allows you to compare similar profiles, help establish what else is available in the market and therefore work out who the most valuable players are to your club.’

There are, of course, some players that speak for themselves. A world class player is going to be considered world class until their numbers have suggested otherwise for a considerable period of time. Most aren’t afforded this wiggle room, though; for the average top flight player, the numbers are everything. Players no longer only have to worry about their key statistics (wins, goals, assists, interceptions), they now have to consider everything from passing success rate to number of intense sprints per game. These are metrics the manager may well have noticed before data came into play but, with the numbers laid out bare for all to see, there is absolutely no hiding. For a young player vying for a big move, a poor pass completion rate could mean not being seen by a scout at all.

You might, then, assume that the data scientist could, ultimately, replace the scout, but there are plenty of reasons to be skeptical of data’s efficacy as a standalone. Put simply, the numbers can lie. There are numerous examples of statistics warping perception of a player’s ability, but let’s take a simple one to illustrate the point. As a substitute and bit-part player in Chelsea’s title-winning Premier League campaign, Michy Batshuayi featured in just 236 minutes of football. His five goals in that time give him a goals-per-minute ratio of one to every 47 minutes, the best return in the Premier League.

The necessary context here is that Batshuayi often came on late, against teams Chelsea would be expected to beat, and scored two against the defensively abject Sunderland on the very final day of the campaign, when neither team had anything left to play for. Ultimately, context is key. For effective player acquisition clubs need both a knowledgeable scouting team and the data to support any decisions and minimize the risk involved. Together, they can create an effective machine.

Data is also influencing player sales, as well as player acquisition. A prime example of this would be Portuguese club Benfica, who are just as successful at developing young players as they are at winning games of soccer. In fact, there are few better clubs in world football at sniffing out talent at a young age and employing data science and advanced training techniques to develop them into valuable assets. Take David Luiz, sold from Benfica to Chelsea in 2011 for a fee of €25 million. Just four years earlier, the Portuguese club had snapped him up from Brazilian side Vitória for just €1.5 million. Atlético Madrid keeper Jan Oblak followed a similar trajectory, moving to Benfica for just €1.7 million in 2010 and leaving four years later to the tune of €16 million.

The secret to Benfica’s success in the transfer market begins at their youth training facility - the Caixa Futebol Campus. The facility has enough accommodation for 65 youth players, with almost every aspect of their lives closely monitored and analyzed to ensure optimum development. Everything from their sleeping habits to their spring pace and recovery time is logged and stored in one huge database for analysis. With this information, the coaches can put together individualized training regimes that focus on eliminating weaknesses and ensuring that injury doesn’t hamper development. The data is both heavily protected and shared with the players themselves - a truly modern take on bringing the youth through.

Data shouldn’t and can’t run transfers for a club on its own, nor can it develop youth players into marketable assets. What it can do is provide already established and effective teams with yet more tools to make the best decisions. As data visualization and data pooling techniques improve, the likelihood is the conversation around data’s effect on sports will actually lessen. It will become more of an embedded tool and less of a revolutionary game-changer. Figures like Altman will become another cog in a scouting department rather than a caretaker brought in to fix and ailing side. 


Picture credit:

Sports tech

Read next:

5 Sports Tech Startups To Watch Out For In 2017