1. Introduction

This Monthly Report continues the reflection started in the previous edition in determining the technical profile of players. To do this, we have again based ourselves on match data produced by the specialist company InStat.

The sample is made up of 6,111 outfield players from 32 leagues of UEFA member associations who had played at least 1,000 domestic league minutes (injury time included) during the 2021/22 season for the club employing them on the 24th March 2022.

Figure 1: study sample

The profiling was carried out using eleven game variables: six offensive, four defensive, and the passes. The actions selected have the advantage of being frequently carried out during matches, with a good distribution between players, which guarantees a more robust profiling. This is even more so when we take into account a sufficient number of match minutes as we have done in this study.

Figure 2: match variables taken into account in the profiling

2. Domains and profiles of play

For all the variables taken into consideration, we have calculated the gap between players’ performances and the average measured at the level of their team. This approach allows us to isolate the roles footballers fulfill within their clubs. These roles are not only linked to the position occupied, but also to the personal characteristics of players.

Insofar as the technical gestures selected are not distributed in an unequivocal manner between the players of a team, some being more concentrated than others, the differences from the team’s average have been transformed into standard normal distributions. This procedure allows us to identify without statistical bias the actions for which the players really stand out the most from their teammates.

As figure 3 shows, the eleven indicators taken into account were sorted according to the area of the game they belong to. For example, the crosses and dribbles are both linked to the "take on" dimension. Or the key passes (passes for goal opportunities) and assists (passes for goals) refer both to the domain of "chance creation".

Figure 3: game domains of reference of the variables selected

For more finesse in the definition of roles played within a team, we have also taken into consideration the second variable for which players stand out the most from their teammates. With some additional groupings for atypical playing profiles, for example players who are very active in shooting but otherwise defensively oriented, we finally retained 15 technical profiles.

Figure 4: technical profiles retained

3. Players’ distribution

Some technical profiles being more common than others, the players analysed are not evenly distributed among the fifteen categories identified. Nevertheless, all profiles comprise at least 3% (i.e. the defensive shooters) and at most 14% of the total number of players (i.e. the ground-to-air blockers).

Our approach makes it notably possible to distinguish between the roles played by footballers in the same playing area of a team. For example, the attacking trio of Paris St-Germain is divided between the profile “shooter creator” for Lionel Messi, that of “shooter infiltrator” for Kylian Mbappé and that of “infiltrator creator” for Neymar Júnior.

Concerning the usual starting 11 midfield trio of Real Madrid, it spreads between the “defensive shooter” profile for Toni Kroos, the “air blocker filter man” profile for Casemiro and the "playmaker creator” profile for Luka Modrić.

Figure 5: players’ distribution between technical profiles

Distinct profiles can be identified also for footballers playing in the same position in different teams. Liverpool FC’s centre back Virgil van Dijk, for example, has an “air blocker filter man” profile, while his counterpart from Chelsea, Antonio Rüdiger, has an "air blocker playmaker” one, and Manchester City’s centre back Rubén Dias plays a more common “ground-to-air blocker” role.

Figure 6: technical profiles for three clubs

At least 1'000 domestic league minutes, season 2021/22

4. Conclusion

The role-based approach presented in this report is particularly useful for determining the technical profile of players based on actions carried out in comparison to their teammates, independently of their position on the pitch.

For sure, the position played influences the roles carried out, but it does by far not account for everything. Our profiling method thus allows us to refine comparisons between players, both within a team and between different teams.

In parallel to the methods of statistical proximity and the k-medoids algorithm presented in the previous Monthly Report, and in combination with the experience capital method detailed in the 70th Monthly Report, this approach is a precious tool available to clubs to fine tune their strategies and recruitment techniques.