I don’t always understand my models, so I just want to take a quick look at the data and see what I find. These are the parameters that I currently use for the model. Come on, let’s explore!
Kicker scores by player
Points allowed to kickers (by opponent)
Points allowed to kickers by home team (i.e., stadium)
Kicker scores by team score
Kicker scores by Opponent score
A few thoughts:
Opponent score has absolutely no effect. Team score has an effect. Home team and opponent are closely related. JAC gives up a lot of points. KC and SEA do not. There really isn’t any strong effect at the Superdome (I have a memory of someone saying it’s a good place to be a kicker). In fact, stadium seems like a weaker predictor than just opponent.
I’ll keep playing with the data, but I want to post this so I can stare at it on the train and hope that some idea for a new analysis comes to me.