I’ve written previously that points scored by teams promoted from the Championship are essentially meaningless when it comes to predicting how many points those teams will score in their first Premiership season. Shots do do a better job, but I no longer track TSR in the Championship, so when making predictions I just blanket predict that each newly promoted team will score the average number of points scored by the previous newly-promoted teams. It’s a pretty dumb model but it saves me a ton of work for what I suspect would be a marginal gain in accuracy.
Anyway, the point of this post was to show this plot, which I was convinced was on the blog somewhere but I now can’t find – I feel like it’ll be a useful resource to point to in the future. It shows how many points have been scored by newly promoted teams in each Premiership season since ’01-02.
The average here is 113.8 points per season go to the newly promoted teams, or 38 each. And for the curious amongst you the standard deviation of the sample is 14.0. And using this blanket method will lead to predicting the points scored by the newly promoted teams to within 10 points 65% of the time, which is really pretty good for saying how truly dumb of a model it is.