Before the last season started I made a prediction of how the final regular season table would look. This post will revisit that to see how well things panned out.
First of all, a look at the predicted table compared to the final table. Note that the difference relates to the final position achieved by the team in the predicted table. And yes, it could have been laid out better.
With the average being 1.83 places off where predicted, I can hardly call this a success. In fact, the sum of the positional differences for each method were:
- Pythagorean – 24
- Bookies – 16
- Press – 22
- Gut Call – 24
- Overall – 22
But that belies some important considerations. The summary table did predict the play-off contenders correctly, which only the Bookies did in isolation. Similarly the bottom three were correctly identified, although that will be of scant consolation to fans of Worcester as they effectively swapped with London Irish. The mid-table is where the damage was really done, as no model predicted the calamitous season that Gloucester endured, whilst others such as Exeter, Bath and Sale all exceeded or failed to meet expectations.
What conclusions can come from this quick review? Well, not many unfortunately. This is my first attempt at applying various models to the Premiership, and there is a long way to go. It appears that for 2014-15 a degree of weighting (to favour the Bookies assessment) would be the best option, but this is something of a dark art for me and I am unsure as to how I should apply it. Comparing their success rate with the others, increasing their weight by 33% would appear to be logical – although applying that might be interesting.
Similarly even at the outset some elements looked wrong. I would like to claim great foresight when I said that I felt Worcester’s predicted 10th place was too generous, but that would be to ignore the other ‘feeling’ I had that Gloucester’s predicted 5th was too low – how wrong I was!
So for the next post we’ll repeat the process carried out before last season, but with a slight amendment to cater for Bookie weighting (if I can work it out). I am hoping to pay more attention to the season as it is underway, and in particular I am looking to identify any ‘rules’ that might offer an edge in match betting – calculating the home advantage figure might help.