Broadly speaking, football players perform their actions with the intention of increasing the chance of scoring a goal or decreasing the chance of conceding a goal. Hence, a natural way to measure the effect of an action is by computing the action’s impact on the scoreline. To do so, we compare the chance of scoring and conceding a goal from the match situation before the action with the chance of scoring and conceding a goal from the match situation after the action.
Our Contribution Ratings model values any type of player action based on its impact on the scoreline. That is, an action valued at +0.03 is expected to contribute 0.03 goals in favor of the team performing the action, whereas an action valued at -0.03 is expected to yield 0.03 goals for their opponent. For example, while a pass valued at +0.03 might not have resulted in a goal in one particular situation, it is expected to yield three goals if the action were repeated 100 times in highly similar match situations.
Let's take a look at this tackle by Virgil van Dijk in the match against Tottenham Hotspur. Dele Alli is entering the box and about to shoot when Van Dijk stops him from shooting with a brutal tackle. Before this tackle the chance of scoring for Tottenham was 0.20, which is similar to the expected goals value of shooting from such a game situation. After Van Dijk's tackle the chance of scoring for Tottenham has drastically decreased to 0.01, which means that Van Dijk's tackle gets a contribution rating of 0.19. If Liverpool would have gained possession, the contribution rating of this tackle would be even higher as Tottenham’s chance of scoring would have dropped below 0.
For every player, contribution ratings of all individual actions of a certain action type are summed up and normalised for the minutes he has played (per season). By doing so, the module is able to show how each player rates on different aspects of the game, compared to all other players in the same position in the same league for that specific season.
When we consider action type Crosses, for example, we obtain the cross contribution for a player by first summing the contribution ratings of his/her crosses in the season and then normalizing the obtained sum per 90 minutes of play.
The odd one out: Defensive Positioning
The aforementioned explanation applies to all “on-the-ball actions”; actions in which the player under consideration actually touches the ball (or the opponent in case of a foul). So this concerns all action types we provide on the platform, except for Defensive Positioning.
Instead of looking at the chance of scoring and conceding a goal from the match situation, we examine the contribution of a direct opponent in case of defensive positioning. Specifically, a player receives positive credit when their direct opponent has a lower impact in terms of Contribution Ratings for offensive actions than can be expected based on earlier performances.
To give you an idea of how these Defensive Positioning computations work, we provide an example computation of the contribution on Defensive Positioning of Séamus Coleman (Everton FC) in the Premier League match against Crystal Palace on February 8, 2020. In this match Coleman’s direct opponent was Wilfried Zaha. In recent matches, Zaha reached an overall offensive contribution of around 0.4 per 90 minutes. In the match against Everton, Zaha only reached an overall offensive contribution of 0.3, which is -0.1 compared to what we would expect based on recent performances. Therefore, we consider the contribution on Defensive Positioning of Séamus Coleman to be +0.1 for this match. Note: this is a simplified representation of the actual calculation. In practice, the model also assigns a part of the defensive contribution to surrounding players (Yerry Mina, Morgan Schneiderlin and Theo Walcott), and accounts for other factors, such as substitutes and positional switches too.