Part 15 of the still growing season review. Last time we looked at possession in it's simplest form. Now we shall look at possession to see how it correlates to numbers like, points, win %, shots and shots on target.
Join me after the jump to see how it all shapes up.
2011/12 EPL Possession By Win % And Points
In the table above we see EPL possession (blue), Win % (red) and points gained (yellow). The teams along the bottom of the graph are in order (l to r) of possession gained.
We see possess ion slowly decline as we move left to right. Win % and points gained correlate with each other extremely well, as would be expected. ( points average is 52 per team, Win % is 37.5%)
What we are interested in is; Does possession correlate with win % and points? What we see is a general decline in win % and points as we move downwards in possession. There are 4 positive spikes on the win % line, being Man City, Man Utd, Newcastle (and Chelsea, to a lesser extent). These teams are the outliers, the exceptions to the general pattern of decline.
Swansea, Liverpool, Wigan and Wolves are teams in the top 10 by possession whose points total is lower than their possession % number. They are the exceptions until we reach teams further along the possession scale, like, Bolton and QPR.
Here is the same graph as above but by a logarithmic scale (best fit). Possession (blue), Win% (yellow), Points (red)
This gives a slightly better indication, by eye, of what is happening. We can see a general trend that indicates the best teams make the most of their possession, this is registered in the win % and points curves. We see a gradual decline of those measures against possession as we go further to the right (and decline).
Possession by Shots for and Against
Here is the graph for possession against shots on target for and against. Possession is listed in descending order, by team, left to right.
From the above chart we see that the shots for totals (which matches each teams possession number) declines, overall, as possession declines (l to r). We see the decline in the original line, but I included a linear line to emphasize the point. It's nearly a perfect fit but for Swansea. Despite their 3rd place possession, the club had the16th lowest shots for total and that throws of the graph slightly, but the pattern is there to see.
More possession equals more shots for, with a couple of exceptions, Swansea, Everton and West Brom.
As for shots against we see that the more possession a team has (and maybe the better a team is, whatever that actually means in a real sense) the less shots against they concede. As possession declines as we go from left to right we shots against totals rise. Less possession means more shots against and we know where that leads us.
The teams that stand out from the gradual rise in shots against are; Swansea, who conceded a lot, Wolves also conceded a high number and Everton who conceded a low number of shots for the amount of possession they had. Stoke are out of sync at the very end of the graph. They don't have a lot of shots for or concede that many in proportion to their (low) possession.
Possession by Shots On Target for and against
Here we also see a general decline with shots on Target for against possession. From nearly 6 shots on target per game for the best team to around 2 shots per game for the worst team. Swansea and Wigan are the teams that dont fit in with uniformity, both teams are under-performing in terms of how much possession they had.
Again, as expected, shots on Target against rise as possession declines. Less ball time-more risk of shots against. As always, there are exceptions to the rule. Swansea, Wolves, Norwich and Bolton were conceding more than the general trend of possession decline suggested they should have
Possession by Goals for and Against
Raw data, above.
And here we see the same numbers in a different display.
In the first graph, which features the raw data, we see that although there is an overall decline in goals for, and, an increase in goals against as possession declines, there are many peaks and troughs along the way, and goals aren't a perfect fit. The general increase and decrease by possession is there for us to see though.
Now, correlating goals to possession isn't really a useful exercise. Why? There are too many variable to the scoring of a goal-shooting inefficiency, goalkeepers, luck etc. I thought it a fun graph to put at the end of the piece. Again, as we saw before, we see as possession per team declines, that goals for also declines (this is more to do with shots, and more importantly shots on target). We see the same also for goals against. Less possession means more goals against, again other factors involved.
Thoughts
The above data is only a one year, 380 game study and there are risks of small sample size effects here. But it is what it is, a 380 game look at what effect possession has on simple stats like shots for, shots on target, win % and points (goals as a fun bonus).
The effect is relatively clear. I speak generally here, not exactly, and I believe that the more possession a team has the more shots it is likely to have, that in itself leads to more shots on target.
The less possession a team has the more shots and shots on target against are conceded.
We then extend the possession link and say the more possession-the more points a team gains, again I speak generally. All the links to points and possession are indebted, I believe, to the rise or fall in shot totals, and specifically shots on target totals.
Thanks for reading. Part 3 of possession in the EPL in a day or so. It's good beach weather here!