The methodology – I’ve taken twelve seasons worth of data, and for each team calculated the TSR for the each 9 game span during the season. Games 1-9 are sample number 1, games 2-10 are sample number 2 etc. This gives 30 spans per team season. Across 240 team seasons that gives 7200 spans. First up, a plot that shows the cumulative frequency of a team not achieving at least a given TSR over a course of 9 games.
One thing that I think is important to note – this line crosses 50% frequency at a TSR of 0.486 – I’ll probably come back to that in the future.
Today though I’m interested in how the ’13-14 season is shaping up. So first I’m going to overlay that plot with the current TSRs of the teams in the Premiership this season.
So we have a pretty asymmetric group of teams falling on this line – it’s definitely more weighted at the lower extreme than the upper extreme. Hull, Cardiff, and Fulham are all in the worst 10% of 9-game-TSRs, whereas only Spurs are in the best 10% of TSRs (Fulham are actually in the bottom 0.3% – we’d expect to see about 1.5 of these stretches per season). There are also large gaps between Swansea in 10th and West Brom in 11th, and between Palace in 17th and Hull in 18th.
What happens if we adjust for the quality of competition faced by each of the teams – does that shift where the teams lie?
This isn’t quite an apples-to-apples comparison but whilst there are a couple of exceptions most of the best and the worst teams shift to the left compared to the prior plot. We still have Hull, Cardiff and Fulham amongst the worst 10% of 9 game TSRs, whilst only Spurs sit in the top 10% (this time Fulham have dropped into the bottom 0.1%…). There are definitely more teams in the ‘above average’ category than the ‘below average’ category, but at the extremes we have more terrible teams than excellent ones.
As a final note – there’s the potential here for tight races at both ends of the table, as both have examples of teams that are probably better than their current position but have given others a head-start (the Manchesters, Palace, and Sunderland) and those whose point scoring is likely to regress (Arsenal, Liverpool, Hull, Cardiff, Fulham) whilst the aforementioned hunt them down.