How Expected Goals helps us learn about the efficiency of players in European soccer
This blog post is part of the Udacity Data Scientists Nanodegree Program. The analysis with the required code is posted in here.
Introduction
Expected Goals (xG) is a new measuring method that describes the probability of a goal shot resulting in a goal. This method is based on parameters such as distance to the goal, angle to the goal, position of the opponent and position of the goalkeeper. Another measurment is the expected assists (xA) which gives the probability that a given pass will become a goal. Therefore, paramters like the strength, the endpoint and the distance of the pass are considered. In this article I want to examine if the goal probability is higher for an offensive player than for a defensive player on the basis of the expected goals and who is the most efficient player. The data that is used is the “In-depth soccer statistics” from Kaggle (see Resources). All players, goals and expected goals of the English Premier League, the Spanish La Liga, the Italian Seria A, the French Ligue 1 and the German Bundesliga are examined.
Question 1: Does an offensive player score more goals than a defensive player?
Before we consider the efficiency with the help of xG and xG, we just look at the scored goals. It seems intuitively clear that an offensive player scores more goals per season than a defensive player. Can statistics prove this assumption right? Defensive players are players who have played a defensive position (defensive midfielder or defender) at least once in a season. All other players are offensive players. Goalkeepers were not included in the statistics. In the following analysis, the seasons of the last six years were taken into account. A total of 28.757 goals were scored in all five leagues. As expected, most of the goals have been scored by offensive players. The plot shows the average number of goals scored by one offensive and one defensive player per season.
Question 2: Do offensive or defensive players have a higher quotient of goals per xG?
Next, the ratio of goals scored and expected goals for offensive and defensive players was calculated. One could expect that this ratio is higher for offensive players as they should be more efficient in front of the goal. In the last four years, a tendency towards the offensive players can be seen, but there is not enough data to make a reliable statement. It is noticeable that the value is often below 1 for both offensive and defensive players, which either means that the efficiency has decreased in the youngest past or that the method is not reliable. As a conclusion, the question initially asked in this paragraph cannot be answered due to a lack of data.
Question 3: Who are the ten most efficient players of the seasons considered?
Finally, the most efficient players of the seasons 2014–2020 were considered. For this purpose, the scoring is taken into account. It consists of goals and assists, which were both divided by the expected goals and the expected assists. In order to avoid outliers, only players with at least 20 scorer points were taken into consideration. The higher the scoring, the higher the efficiency. In the top ten are defensive players.
Conclusion
In summary, we can say that although offensive players score significantly more goals, the probability of scoring a goal is not much higher for the same shot on goal quality. This is also reflected in the most efficiency list of the top ten players.
References:
In-depth soccer statistics: xG, xA and more: https://www.kaggle.com/jashsheth5/indepth-soccer-statistics-xg-xa-and-more
Github: https://github.com/fabieck/Write-a-Data-Science-Blog-Post.git
