Regression in soccer

I want to regress points on goal difference (soccer-related data) over a 15-year period. However, not all of the seasons contain the same number of teams (there were 10, 11 or 12 teams, predominantly 12, in any given season).

The independent variable is goal difference and the dependent variable is points (ie: regressing points on goal difference). The difficulty I have is that there have been occasions when the league has consisted of 10 teams, 11 teams or 12 teams. For example, teams from different seasons could have finished with the same number of points but one side might have played more games than the other. There are three points for a win and one for a draw

How do you adjust for this?

• How are we supposed to answer when we do not know, which role a season plays in your model? What are independent variables, what is the dependend variable and what question do you try to answer with the regression? – Bernhard Feb 1 '18 at 14:28
• Hi Bernhard, the independent variable is goal difference and the dependent variable is points (ie: regressing points on goal difference). The difficulty I have is that there have been occasions when the league has consisted of 10 teams, 11 teams or 12 teams. For example, teams from different seasons could have finished with the same number of points but one side might have played more games than the other. Which is why I was wondering if I need to adjust for this.There are three points for a win and one for a draw – RiobeardR Feb 1 '18 at 14:40
• To be precise: The independent variable is "sum of goal differences of a team during a season" and the dependent variable is "sum of points of a team during a season"? – Karsten W. Feb 1 '18 at 15:19
• Yes, Karsten W, goal difference = goals scored minus goal conceded of a team during a season – RiobeardR Feb 1 '18 at 15:21