Following up on a promise I made earlier this week, here is the final Pythagorean for this season's English Premier League.
|West Ham United||38||47||66||35||41||-6|
I thought there would be a huge discrepancy between the projections and the point totals, but I looked at my code and found out that I was calculating win/draw probabilities for a team that scored up to five or six goals! That wasn't sufficient for this season so I adjusted the maximum score upward, which greatly improved the results.
So the top four performed in line with their expectations. Tottenham were about a point better, Manchester City about a point or two worse, but that was the difference between Champions League or Europa League football next season. Liverpool and West Ham underachieved in a big way if you looked at the statistical expectations, and if you talked with supporters of either side they would probably say the same thing!
I used points won on the field to compare with Pythagorean estimates. Portsmouth would have gone down even if they hadn't been assessed a penalty for going into administration, but they underachieved during the season. Wigan, despite their propensity for suffering heavy defeats, performed seven points better than expected. Those home wins against Arsenal, Chelsea, and Liverpool ensured their survival, and that was reflected in their Pythagorean. It was a combination of failing to cobble together enough points from winnable matches and Wigan Atletic's run of upset victories that doomed Pompey to the Football League Championship — not an administrative ruling by the FA.