Football Analytics Demystified | #1

Football analytics is hot topic. Every match that is being played at the World Cup in Russia gets analysed in detail by football experts on TV, in news papers and online. How does analytics contribute to a better understanding of 'the beautiful game'? Time to make an analysis of this analytical work so you know what is going on.

In this series of blog posts "Football Analytics Demistified" I will shed some light on all the different types of analysis that are being used and how these relate to each other. Let's start with looking at the different types of football analytics.

3 main categories of football analytics

We can distinguish three categories:

1. Descriptive Analytics: what happened?
2. Predictive Analytics: what will happen?
3. Prescriptive Analytics: what should happen?

In this first post I will briefly describe each type and give a couple of examples for each.

Descriptive Analytics: what happened?

This type of analysis is used to describe facts. It is helpful to coaches, players and fans to get an understanding of the performance of teams and individual players during a single match and over a series of matches. These facts allow for answering questions like "What is the track record of a team in away games" and "Which player has the highest shot accuracy". Nowaday, visual information like passing and running options can be added to video footage.

Examples

• Rankings: top scorer's lists, team rankings, yellow and red cards, ...
• Infographics: pixel perfect summaries of statistics with nice drawings and photos;
• Video augmentation: add visuals to video footage to support communicating about football situations;

Predictive Analytics: what will happen?

What is going to happen? Fans all from over the world spend hours and ours of reading, discussing and speculating about the expected outcome of an upcoming match. This anticipation and the inherent difficulty of predicting football may be part of the reason that this game is so immensely popular.

However, more and more time and money is invested in analytical tools and techniques to make the best predictions possible. There is reason why: there is (a lot of) money to be made.

Examples

• Sports betting: predicting match outcomes (big business!);
• Injury prevention: prevent players from getting injured due to physical overload by recording their physiology with sensors;
• Scouting: determining which young talents have the biggest potential to become great football professional;
• Robot football: robots that play football need to predict how the game will evolve so that they can determine their next move;

Prescriptive Analytics: what should happen?

How should football be played? Prescribing the behavior of a team for any kind of situation is the basic task of a coach. However, everybody has an opinion about this. TV pre- and post match discussions show experts that highlight scenes and compare a team's actual tactics with how they should play.

These 'football recipes' are typically highly opinionated as there are not many universally accepted football laws. What is 'good' and 'bad' football is generally speaking a matter of taste. However, if an audience embraces a particular football style, like 'tiki-taka', one has to obey the rules. That is where prescriptive analytics comes into play.

Examples

• Tactical analysis: in-depth analysis of team strategies and formations;
• Video analysis: similar to the descriptive analytics example, but now a theory about the desired game play;

I the next post, I will discuss descriptive analytics in more detail.