Skills
Snitch
- Joined
- Jan 17, 2012
- Messages
- 42,759
Disclaimer - I am not a statistician. Some nerd on here would do most likely do a much better job than my attempt.
Outside of the obvious reason that you need to score a goal to win a game, and winning a game gives you three points which is worth more than 0 or 1 points.
This data is from seasons 12/13 through to seasons 18/19:
All I've done is plot the number of goals scored (blue) and goals conceded (orange) on the y-axis vs the the number of points a team gained over the course of a season on the x-axis. Then using my friend excel I found the R2 of each of those data-sets.
For the entire PL (between 12/13 and 18/19) the data indicates
Now what I've done here is I've normalised the number of points scored/goals scored/goals conceded in each season. I thought this would account for seasons where points totals are lower, and compare goal scored/conceded with the position you will finish.
Basically for each season the team that scored the most points is assigned 1, the team conceding the least is assigned 1 and the team that scored the most goals is assigned 1. Then Points/GF/GA just become ratio.
I.e. 1st place got 100 points = 1. Another team got 80 points = 0.8 etc. Top scoring team scored 100 goals = 1. Another team got 80 goals = 0.8
Long story short - the trend seams to be the same.
Outside of the obvious reason that you need to score a goal to win a game, and winning a game gives you three points which is worth more than 0 or 1 points.
This data is from seasons 12/13 through to seasons 18/19:
All I've done is plot the number of goals scored (blue) and goals conceded (orange) on the y-axis vs the the number of points a team gained over the course of a season on the x-axis. Then using my friend excel I found the R2 of each of those data-sets.
For the entire PL (between 12/13 and 18/19) the data indicates
- a stronger correlation between goals scored and points scored
- than between goals conceded and points scored.
- The difference between the R2s is even bigger if you look at the data for the Top 7 only (i.e. Champions League contenders/pretenders).
Now what I've done here is I've normalised the number of points scored/goals scored/goals conceded in each season. I thought this would account for seasons where points totals are lower, and compare goal scored/conceded with the position you will finish.
Basically for each season the team that scored the most points is assigned 1, the team conceding the least is assigned 1 and the team that scored the most goals is assigned 1. Then Points/GF/GA just become ratio.
I.e. 1st place got 100 points = 1. Another team got 80 points = 0.8 etc. Top scoring team scored 100 goals = 1. Another team got 80 goals = 0.8
Long story short - the trend seams to be the same.
- a stronger correlation between goals scored and how high you will finish
- weaker correlation between goals conceded and how high you will finish
- Holds true for the top 7 too.