None of the views expressed here represent the views of the Coming Home Newcastle Blog
Note: I thought maybe a few of you guys might be interested in some stats that I had lying around, So here goes. I write for Bitter and Blue.com on SBnation network
Welcome all to part 16 of my Premier League review. Today we'll be looking at possession with a twist. In Part 1 we looked at the raw possession numbers. In Part 2 we looked at how possession correlated to goals and points, shots and shots against.
Today I shall be looking at each EPL team and their points total to see if they under achieved or over achieved in regards to their possession total.
Join me after the jump where I'll list every EPL team and show you if your team had more or less points than their possession number expected them to have.
So the first question we ask is; How can we judge if a team under or over achieved when talking of points gained?
Firstly, let me use an example
Arsenal 38 games. 70 points. 58.76% possession.
They are good numbers and Arsenal had the highest season-long possession total.
Arsenal had 1.19 points per 1% of possession (70 points/58.76%) and this is the important number.
The league average possession was 50%, and League average points was 52.5 which equals 1.05 points to 1% of possession. 1.05 points is the league average points per one % of possession. Now that we know the average we can judge whether teams gained more points per their individual possession number, or less points.
This will be called actual points versus expected points.
Using the example above, we can see that Arsenal overachieved against the league average points per 1% of possession. Arsenal's actual points per 1% beat their expected points per 1% by 0.14.
We know Arsenal had a high possession number, 58.76%. Using the League's average points per 1% of possession (1.05), Arsenal were expected to gain 61.69 points, we obviously can round that number up.
Arsenal's expected points total was 61.69 (58.76% x 1.05 lge average points per 1% of possession)
Arsenal's actual points total was 70 points from their 58.76% possession. Arsenal overachieved by 8.3 points, in terms of points gained, with their possession totals.
Still following? I hope so, as there are some cool things via a few graphs I shall show you now
Points By Possession
The above chart is the raw data, At the bottom of the chart we see team name and individual possession number. We are going right to left in descending order. We see those teams and their individual points total as represented by the bar graph.
The Manchester clubs are bringing the mean line up to 52.5 thanks to their respectively excellent seasons. & teams above the mean line.
Teams Points Per 1% of Their Possession
Now, we saw only 7 teams above the points mean line in graph 1. Here, once we have divided the teams individual points total by their individual possession, we see some different results. Remember points per 1%= (teams points/teams possession %)
Fulham and Stoke rise above the league average of 1.05 points per 1% of possession.
Man Utd lead the EPL with 1.54 points per 1%. City are 2nd. Newcastle are 3rd with 1.37, which showed us that they earned a lot of points despite not having a majority share of possession throughout the season. Spurs are 4th, Arsenal, Everton and Stoke are one and the same after the leading group.
Points Per 1% Of Their Possession +/- Differential (actual points per 1% vs expected points per 1%)
This is the first (really) important graph that we have come across. The above shows us a +/- chart in which we can see if an individual team had more points per 1% of possession than expected. Some teams did, some didn't, there were a couple of shocks.
Swansea, the purists darlings. Barca-esque under Mr Rodgers, well, they certainly kept the ball well (EPL's 3rd best possession) but that didn't necessarily mean that it was converted into points. Swansea were down 0.24 points per 1% possession of the ball.
Wolves were down 0.53 per 1%. This, we expected in a way. Wolves were terrible and far worse than their individual expected points per 1% estimate. The real shock was that Wolves were a decent possession team in the 2011/12 EPL, recording 48.21% of possession over the 38 games.
Actual points vs Expected points
The above graph is the all important one. The other graphs featured above are important, but more so as to show us how we got to this graph.
This is actual points vs expected points. This is the one which tells us what a teams expected points total was and how they performed against it.
(expected points=the individual teams possession% x the league average points per 1% of possession-- 1.05 points)
We can see when we use possession and points to come up with an expected points total that the teams with more possession (roughly) exceed their individual expected points total. But there are exceptions.
Actual Points vs Expected Points (differential)
In the above graph we see which teams, when using their overall possession% number x the league average points per 1% of possession, exceeded their expected points totals.
Arsenal, despite having the best possession% gained only 8.3 points more than expected.
Man City and Man Utd were the dominant teams in 2011/12 and they exceeded their expected points totals handily.
Swansea, despite their excellent possession, were 13.9 points worse than expected. Liverpool were -6. Wolves were -25.6. Newcastle overachieved and we then see the a pattern, the poorer possession% a team had, the more likely they were to underachieve.
Differential Per Game
This is what actual points vs expected points looks like when we do a graph for +/- per game.
The Manchester clubs were up by a staggering 0.7 points per game. It is simply a sensational number. Is it sustainable? We don't know. Swansea down 0.4. Wolves down by 0.7 points per game and we also see Stoke, despite the worst possession % in the EPL up by o.1 points every EPL fixture. Why was this? The next 6, yes 6 posts, will show us why.
Upcoming on the 2011/12 season review.
Expected shots for
Expected shots against
Expected shots on target
Expected shots on target against
Expected goals for and against
Thanks for reading
I write for Bitter and Blue.com. Twitter