In the early 2000s, the book and movie Moneyball shone a spotlight on the use of data in sports – and highlighted just how important analytics can be in contributing to the success of any individual or team.
Fast-forward to 2019 and it is now commonplace for many of us amateur athletes to collect data about our own performance and vitals, be it through smartphones, smart watches or sensors.
While at the professional level, data and analytics are the cornerstones of any team, when they’re used smartly, they don’t just lead to opportunities on the pitch or the court, they translate into rankings, customer and fan satisfaction. And ultimately, this all means increased revenue streams.
With so much at stake for the clubs and sporting organizations, you might be thinking what difference does all this data make for the performance of the sports teams on the field? Let’s take a look at how data is being used in every aspect of the game, giving a whole new meaning to the idea of ‘total’ football.
Training analysis – how data fine-tunes a team’s strategy
Matches are the fleeting tip of the sporting iceberg we see as football fans – on the terraces, screaming at TV screens, huddled together in public spaces, often with our hands in our mouths, willing in that late deciding goal. But most of the action for the teams actually happens in training.
Players train for five days a week – and a lot of what they do on the training field is tracked via sensors. Data is used to record their vitals – everything from the distance covered in training, their body position, along with any other indicators that give coaches, managers, physios and sports scientists the information they need to improve performance and fine-tune the team’s strategy.
Professional football players don’t just need to be skilled with the ball, they also need endurance, strength, speed, agility and an exceptional understanding of where their teammates are at any one time.
The metrics used to measure these attributes in a player include (1) :
- High speed running distance: the distance covered at the speed above a predefined threshold (that can be tailored to the individual (i.e. speed above 120% of their optimal aerobic velocity or 85% of their maximum speed) or not (i.e. >19.8 km.h-1)
- Mechanical work : representing the number of accelerations/decelerations above an intensity of |3 m.s-2|.
- Force load distribution: force load refers to the sum of estimated ground reaction forces during all foot impacts. It’s assessed via the accelerometer-derived magnitude vector (roughly the area under curve of z force on the accelerometers). fL can be compared between right and left legs during any locomotive actions (e.g. specifically while accelerating vs running at high speed, which is likely related to the use and potential weaknesses of different muscle groups).
Tracking the relevant data and analyzing it during and after training sessions allows coaching staff to make important decisions prior to match day.
Match analysis – data is your most astute coach
During a football match the recorded data commonly comes from video systems which track events as they happen. But there’s also a team of video analysts recording events manually as they unfold.
These two mechanisms combined lead to a rich data collection that can be accessed by clubs during a game but also to carry out pre- and post-match analysis.
Football clubs use their own team’s data alongside data on their opponents to understand team formation and tactics. Plus, they look at individual metrics like a player’s pass completion rate or how many goals he or she scored, above expectations.
So, match analysis takes many forms, ranging from high level statistics summarizing a game, to analyzing player positions on a pitch, and even body position and orientation during passing.
It’s the combination of all this data that uncovers the game-changing nuggets of insight about different teams.
For instance, one team might play very defensively in the first half of a game but then might play offensively to push their opponents in the second half.
Likewise, another team could be highly skilled at passing through and around their opponent’s defenders to give them a better chance of converting a shot on goal – simply because all the defenders have been left behind.
The point is, data makes it much easier to predict all of this. The most important metrics that clubs typically focus on are:
- Number of ball touches in the last 3rd: this is basically the number of ball touches by your player (and opponent’s players) in the last 3rd of the pitch (where your strikers play)
- Expected goal (xG) (2): xG is the probability for any shot to turn into a goal (or cumulative probabilities when we look at xG at the end of the game). In a low-scoring game (such as football), final match scores provide a poor picture of overall performance – these models instead provide clearer indication on performance.
- Pitch control models (3): this is a recurring concept in the analysis of space dominance in team sports. It can be defined as the degree or probability of control that a given player (or team) has on any specific point in the available playing area.
Using data to get an advantage over your competition is possible but it’s not easy. Clubs can purchase the data from providers and while many things can be measured, football is still a game played by 22 humans who bring their own unique physical and mental characteristics to each game.
So how can we ensure we create the right team?
Player recruitment – how data has become your scout’s best friend
Finding the right player to fit into a team is a challenge that player scouts have turned into a profession. And it’s not just about finding a talented player, it’s important to get someone who will get along with the staff and their teammates. Plus, it’s vital that their playing style, unique skills and understanding of the game fit with the team, the game strategy and the combination of existing skills from other players.
It’s no surprise that these considerations can become overwhelming and really time consuming for the scout on the ground. That’s where data comes into play. It means clubs and their scouting team can reduce the long list of options to a shortlist of potential candidates, who they can then focus their visits on to watch in person.
Of course, it’s still important for scouts to use their experience and their ability to read a player when assessing their fitness. But watching a selected list of athletes over a certain timeframe, seeing them in action and evaluating them against a list of criteria all helps in improving the odds of solving this important piece of the puzzle. In short, data allows scouts to prioritize and focus on the most promising players.
And by using advanced clustering methods, clubs are now able to quickly find new players who are similar to other players. This cuts down the number of players clubs need to review, reducing the burden of classical approaches and allowing for deeper analysis of a small number of potential picks.
What are clubs focusing on for this season?
The use of analytics in football has grown significantly in the last decade and it has also evolved. More data has become available. And there’s more automation for data gathering. All of which means clubs now have access to rich datasets and they’re able to extend their analyses further than ever before.
The skillsets of analysts have also expanded and clubs have started to recruit beyond the world of sports and sports science. They’re now also seeking talented data analysts and scientists from other disciplines, such as applied mathematics, finance and behavioural science.
One focus area that analysts are currently exploring is the application of deep learning tools to assess body positions and orientation. Clubs are now interested because tracking this will allow them to fully understand players choices – and most of the time, good positioning is related to position on the pitch (x,y data), space (pitch control models) and body orientation. It’s just one more piece of data analytics which can make a big difference to the success of any team.
As every team craves the same valuable data insights, we can safely say the days of watching a tense cup final with your fist in your mouth aren’t going anywhere. But we wouldn’t have it any other way.
By Eva Murray, Exasol and Mathieu Lacome, Sports Scientist
(1) Player-Tracking Technology: Half-Full or Half-Empty Glass? Buchheit M, Simpson BM.
Int J Sports Physiol Perform. 2017 Apr;12(Suppl 2):S235-S241.
M. Lacome, B.M. Simpson, M. Buchheit. Monitoring training status with player-tracking technology. Still on the road to Rome. Part 1. In Aspetar Sports Med. Journal, 2018.
Training Load and Player Monitoring in High-Level Football: Current Practice and Perceptions.
Akenhead R, Nassis GP.
Int J Sports Physiol Perform. 2016 Jul;11(5):587-93. doi: 10.1123/ijspp.2015-0331. Epub 2015 Oct 9.