About a month ago I took a look at a way to predict the end of season table based on an individual team’s form over the course of the season. The results weren’t necessarily comforting for Fulham FC supporters concerned with relegation, but nothing to get overly worried about just yet either. With only five games remaining in the season, it's time to take another look at the same model to see what has changed and what has stayed the same.
For a quick refresher, I used the total points each team had after each game to draw a line of best fit and used a computer program named Gretl to forecast the line to the end of the season. From there you can look at confidence intervals, standard errors, and all of those good math-ey things. Here is Fulham’s graph updated to their latest win against Cardiff:
Just looking at the red line on graph, it’s pretty cool to see what the past four games have done for the season. We’ll stay with our fingers crossed that red line continues on it’s way up, but in the meantime we have statistics to drag us back down to reality, but that might not be as bad as it sounds.
Using the same model described above across the bottom twelve teams, here is what the *updated* predicted end of season table looks like:
Now there are one or two "anomalies", shall we say, to take note of. The first is a big one. Nottingham Forest have taken a spectacular single point in their past six English Championship matches. Given that this model only takes into account form across the entire season, and doesn’t put weight on current form, so the math still thinks that they will end up in 15th place. I’m not so confident that they will be able to get the eleven points that is the difference between them and the predicted end of season table, so a 17th – 18th place finish is probably more realistic, assuming that they are able to turn their form around.
The next and most interesting to us is Fulham. While the math remains confident that the team will finish in 19th place, I’m confident enough to make the prediction that Fulham are going to hover around 17th place and maybe even finish slightly higher judging by the remaining games and current form (which is awesome.)
So yes, there are some flaws to this way of predicting results. My goal for this summer is to come up with a better model that takes more explanatory variables into account, such as goal difference, team wage budget, etc. to try to be a little more accurate. We’ll see about that.
I want to end this article in the same way I ended the last one where I predicted the final table, which is by saying that it is still anyone’s guess what is actually going to happen. Especially since there are six teams in the present table that are at least within a single point of one another. Having said that, I remain as confident as I was last month that Fulham will beat the drop, and I have the math to back that up.