Shooting has been one of the most analyzed and researched parts of association football as it directly leads to goals which determine the score of the match. We take a look at it from a previously unseen perspective and analyze if there are differences between four different levels in German football (Bundesliga, Regionalliga, U19 Bundesliga and U17 Bundesliga) in shooting tendencies and efficiency and explore how these change as players get older. To do that we employ statistical analysis and examine the individual weights of Expected Goals models based on logistic regression. We find that players in higher levels tend to be more risky and aim for corners of the goal and are more predictable in terms of their shot origins. A comparison of headers and kicks show that goal likelihood of the latter is much more influenced by whether a shot has happened after a set piece, whereas goal likelihood of headers decreases more steeply with increasing distance from goal. Analysis also reveals that with increasing level goalkeepers tend to be more reliable saving shots at medium height but have a harder time with shots aimed at bottom corners.
© Copyright 2022 Machine Learning and Data Mining for Sports Analytics. MLSA 2022. Communications in Computer and Information Science. Published by Springer. All rights reserved.