Based on the performance of teams in the English Premier League so far in the season, I’ve developed a model which predicts how each of the remaining matches in the league will go. While predictions on individual matches might be shaky and hence not very accurate, I want to publish my prediction on what the league table will look like at the end of the season. If my predictions turn out to be right, I can claim later on to be a master sports predictor and hope for some business to come my way because of this. Of course I’m taking the risk of putting my predictive reputation on the line by coming out with these predictions, but sometimes such risks need to be taken.
I last ran my model on April 1st, and this is my prediction of the final tally of the Premier League.
Some notes
1. This is a purely statistical model based on goals scored so far in the season. For each time I’ve modeled how many goals they’ll score given the number of points their opponents have racked up in the league so far
2. I know this is wildly optimistic for Manchester United as the model thinks they will win each of their next 8 games. I don’t think this is going to happen. So disregard that prediction
3. Interestingly the top of the table according to my prediction is identical to the top of the table now.
Wouldnt it be better to also take into consideration defensive records? Perhaps then the model can learn from more than one feature. It is probably markovian at some order, so previous game histories would also possibly factor in. The real advantage is none of the top 3 have anything to concentrate on but the league, so too arsenal, but, chelsea have the added baggage of the europa league to look at.