In the one of latest articles, I focused on explaining xG metric (you can find it here). Now, let’s apply it into practice and look at some underperformers and overperformers based on this metric. Just to remind you, xG measures number of goals that player should have scored based on the quality of his chances. Therefore, another metric xA (Expected Assists) measures number of assists a player should have had. When we make sum of xG + xA, we get xI (Expected Involvement) = number of goals + assists that player should have registered. In this article, I am focusing on differences between actual involvement (goals + assists) and Expected Involvement (xG+xA) of some Premier League players in order to determine how much more (or less) some players should have scored. I am using xG and xA data from understat website, where you can find a lot of great stats about Expected Goals or Assists. So, the number of assists may differ from “FPL assists” in some cases.
Forwards 18/19 season
First, I am looking at forwards from 18/19 season. Chart below displays values of two variables: actual involvement (goals + assists) on the vertical axe and Expected Involvement (xG+xA) on the horizontal axe.
If there had been no under/overperformers and all players had performed fairly (Exected Involvement = actual involvement), all points on the chart would have lied on a straight 45° line. This hypothetical line is prompted with blue colour in the chart. Every player which lie above this chart is an overperformer (Expected Involvement < actual involvement) and every player below the line is an underperformer (Expected Involvement > actual involvement). The vertical difference between the blue lie and a blue point is the difference between xI and the actual Involvement. Bigger the vertical gap between the point and the line, bigger the margin by which player under/overperformed his Expected Involvement.
Overperformers
The biggest overperformer of the last season from selected forwards is Alexander Lacazette. He scored 13 goals and registered 8 assists (his Involvement was 21), but based on the xG and xA, his involvement should have been just 16,13. Salomon Rondon overperformed his Expected Involvement by 3,95, Sergio Aguero by 3,82. These players overperformed their xI by a big margin. All others above blue line overperformed their xI by less than 2.
Underperformers
Gabriel Jesus was involved in just 10 Man City goals, but Expected Involvement suggests that the number should have been higher, and he should have had hand in more than 15 goals. Mitrovič underperformed his xI by 4,42 and Lukaku by 3,43. Ashley Barnes was superb at the end of season and yet he should have been involved in 2,9 goals more. Joshua King underperformed his xI by 2,81, Danny Ings by 2,69, Jamie Vardy by 2,26 and Roberto Firmino by 2,08. All others above the blue line underperformed their Expected Involvement by less than 2.
Source: understat
