Whilst the last four seasons have provided a new winner of the Premier League as well as a couple of surprising relegations, there has been a stability to the upper half of the table which might have gone unnoticed to many.
In this four-year period, a whopping seven teams have made it to the top ten every season (Man. Utd., Chelsea, Arsenal, Man. City, Tottenham, Liverpool, and Everton). Six of them also inhabit the upper half in the current season. In fact, only 15 clubs have entered the 40 spots provided in the upper-half during these four seasons. Honorable mentions go to Aston Villa and Fulham, who made it in three out of four attempts.
This must indeed be a disturbing read for the people of the the West Midlands, the Leeds-Bradford conurbation, or Tyneside-Wearside-Teeside. Currently only the Merseyside area is challenging the domination of London and Manchester clubs in the Premier League, and neither of them look like serious contenders for the bucket at present.
The domination of the top clubs makes for an interesting study, though. Since these clubs compete so consistently for the top honors, or one of those desirable four spots that grant access to the Champions League, it is interesting to look at how the clubs fare in the so-called ‘six-point’ matches that we love and fear so much. The idea is that their consistency makes them more statistically interesting, especially if we look at them combined.
So, is this a good idea? Well, there is only one way to find out.
Already at first glance the results seem to contradict the notion that the internal battle between the seven most consistent clubs will relate strongly to how they finish in the league. The chart below shows the total amounts of points harvested by the teams internally over the last four seasons (2012/13 not included). This totals 168 matches, of which 84 – obviously – are home games.
[table id=92 /]
Clearly then, this does not grant my “internal league” a lot of explanatory power, since Liverpool come out almost on top, despite their plunge into mid-sea fishing in the latter seasons? And what about Chelsea ranking fifth, despite the fact that they have finished in the top three in 75% of the four last seasons? The strongest team amongst the seven is, not surprisingly, Man Utd, with an average of 1.83 points per game. What might come as a bigger surprise is that Liverpool fare second best, averaging at 1.5 points per game. This is in fact only 0.2 points less than their total average for all games in the four seasons, which stands at 1.7 points per game. Chelsea on their side only manage to bring home 1.19 points per game to match Everton against top opposition, a huge plunge from the 2.0 points they have taken home on average in the last four seasons against all opposition.
Well, apart from cementing the myth that Liverpool lack consistency against opposition from the lower end of the table, and coining a new narrative that Chelsea create their success by preying on that same bottom half, things get a bit more interesting when we start digging into the numbers by chopping them into smaller chunks.
The tables to the right look only at the final top-four ranking, and breaks it into separate seasons. The first column (ILP) shows how each team ranked internally amongst our seven selected teams in the same season. The second (DEV) how much their actual league position deviated from this internal league position (ILP).
It might not sound much, but remember that we only looked at 168 matches (11%) of the total 1520 matches played during these four seasons.This I find interesting. It might not seem like it at first glance, but these numbers are quite precise. Indeed, for the 11/12 season the numbers are more or less spot on, only missing number four by 1 spot (ranking Liverpool above Tottenham internally).
Adding it up, we get this table, which shows the power of the predictions given by just looking at these seven clubs. The bigger the number, the bigger the difference between the internal ranking and the actual league position. No value means the prediction was spot on:
[table id=94 /]
As the table shows, the first place winner was predicted accurately in two of the seasons. All over, even with the messy 2010/2011 prediction, this method had predicted position 1-4 for four consecutive seasons within a margin of 1.31 places. In 25% of the cases the prediction was spot on, and 44% was within 1 spot of the actual result.
So what does it tell us? Well, first of all, and not surprisingly, it seems that the “six-point” matches have a huge impact on how the teams fare at the end of the season. Second, it shows us that the consistency of the upper-half teams can be used to one’s advantage statistically when placing one’s bets or attempting to determine wether setting fire to that jersey can wait another couple of weeks…
As for how they fare so far this season, with about 1/4 of the internal matches played:
[table id=93 /]
I will leave it up to the reader to decide how much to put into it at this early stage. Keeping an eye on the further development amongst these seven teams should however give us a good clue on how the end table will look.