The FIFA Football Language

Previously, a team’s performance analyst was responsible for everything video- and data-related, but we are now witnessing a surge in specialisation. When we look at specialist data professionals, we can now observe an array of data scientists, data analysts, data engineers and software engineers being appointed by teams to utilise and create insights from the ever-increasing amount of data now available. With so many more specialists contributing to the creation of intelligence, the key is finding the right balance between data insights and technical interpretation, making sure there is always football context.

With talented and experienced data-literate professionals coming into the game and more and more new technologies coming to market, football professionals must always know exactly what FIFA is looking at. Each provider has its own definitions and calculations, while some are more accurate than others. Because of this, even the most basic comparisons between two different data sets will give you different answers to the same question.

The appointment of Arsène Wenger as Chief of Global Football Development brought a new stimulus to increase the level of understanding of the game. Wenger’s vision is:

“To improve football understanding and experience by creating new Enhanced Football Intelligence (EFI) through the combination of technical expert observation and football data analytics”.

FIFA now has a team of technical experts, performance analysts, data analysts, data scientists and data engineers who are committed to implementing this vision. For this team to be able to investigate how the game is evolving, FIFA first needed to look backwards to provide a basis for its analysis moving forward. To avoid data discrepancy issues and to be able to work from a data set that provided more than just on-the-ball data, FIFA made the decision to create a unified language and an aligned performance data set. This language was designed not only to support analysis at FIFA, but to be implemented as an educational tool for the world of football in order to better understand how the game can be viewed and broken down.

The FIFA Football Language is the blueprint for how FIFA wants to analyse football moving forward. It has been developed over the last two years through collaborations with multiple high-level technical experts, performance analysts and data scientists to break each area of the game down into fine detail while maintaining football context. The basis of the language is the individual player and the actions they perform with the ball, but also their actions in relation to the ball when they are out of possession. The FIFA Football Language captures the behaviours of the team and the individual while they are in possession, but also off the ball, analysing areas such as “Offering to receive” and “Defensive pressure”. Each action, where relevant, is considered in terms of the impact it has on opposition team shape, providing another layer of context to player actions. In essence, the language considers player actions while they are on the pitch and not just when they are in possession.

The language is broken down into four distinct sections: In possession, Out of possession, In contest and Goalkeeping. Each element has a sequential flow that is collected around each individual player action. For example, from an in-possession perspective, the language considers the actions carried out by a player in an attempt to receive the ball, the actions carried out with the ball and then the impact of their distribution. The sequence would look like this:

1. Did the player make an offer to receive the ball?
2. Did the player perform a FIFA movement type to receive the ball? This is a data point that describes the type of movement a player performed in relation to the opposition team shape. For example, did the player run in behind the defensive line?
3. If the player did receive the ball, which previous action did they receive the ball from?
4. Where did they receive the ball in relation to the opposition team shape?
5. What pressure from the opposition were they under when they received the ball?
6. What did they do with the ball?
7. What pressure were they under when they attempted to distribute or progress the ball?
8. Did the distribution action penetrate the opposition’s shape?
9. What was the outcome of the possession sequence?

Within each of these sections, additional information is collected to provide more detail on the relevant action. If we look at “Offering to receive”, for example, the language covers a player making an offer to receive the ball and where that offer was made in relation to the opposition team units and shape. The term “units” refers to the number of opposition defensive lines the attacking team or player faced when executing the offer. Team shape relates to the location of the offer in relation to the opposition’s positional structure, specifically whether the offer was made inside or outside the opposition team shape. A sample sequence would provide the following information:

This would mean that a player has actively signalled, changed body shape or made a clear movement to receive the ball in behind the opposition’s final unit and outside the opposition team shape. This information would be collected even if the player does not receive the ball.

In the following scenario, a player has actively signalled, changed body shape or made a clear movement to receive the ball between units 2 and 3 and inside the opposition team shape, constituting an offer to receive.

Although much of the Enhanced Football Intelligence data set is collected by a team of highly trained football analysts, FIFA also automates elements through a series of algorithms and models created by FIFA’s data scientists and engineers. An example of this would be how FIFA measures team units. FIFA has created a unit-detection algorithm based on player tracking data that effectively clusters the players on the pitch into lines. From this algorithm, FIFA enriches the manual data collection by adding the opposition units at the time of the action. For example, let us imagine that a team is playing in a 4-4-2, with a defensive mid-block in an organised state; a pass is made, and a player receives the ball behind the first unit attacking line. The data collection process for the player receiving the ball would involve the football analyst manually inputting that they received the ball from a pass, and the automated algorithm would enrich the manual data by adding “between units 1 and 2” and “inside”.

Following the data collection process, by utilising other data sources such as automated tracking systems, FIFA can generate analysis and insights as seen in the video below. An enrichment pipeline and process have been developed to allow the data collection to be enhanced even further. Another area in which we employ an enrichment process is in calculating phases of play. Each in-possession phase (build-up, progression, final third and counter-attack) is automated. In the same way, the out-of-possession phases (high block/high press, mid-block/mid-press, low block/low press, counter-press and recovery) are all automated based on player tracking data, which looks at how many players are located in certain areas of the pitch and the speeds at which they are travelling. Automated data collection ensures consistency in the way data points are captured. This means that when FIFA looks to compare, for example, the time spent in a low block, there is confidence that the data has been collected in the same way across multiple matches. 

With the creation of the FIFA Football Language and the assembling of the technical analysis team, FIFA now has the capacity to progress with deeper and more insightful analysis, having developed a blueprint and framework for observing and analysing the game moving forward. The detail and depth provided by the FIFA Football Language allows for very specific questions to be asked. The mission is to improve football understanding and ascertain why certain events are important within the game. Our analysis will be delivered here on the FIFA Training Centre, combining data, video and technical interpretation to help build and grow technical, tactical and physical understanding within the footballing world. With so many specialists contributing to the creation of intelligence, the key is to find the right balance between data insights and technical interpretation, making sure there is always a clear football context.