Last weeks article showed how inefficient crossing is, especially from open play due to the low accuracy levels across the league. The focus of this article is the relationship between possession and goals scored, conceded and both combined. I will look to compare all three, as well as focusing on the correlations depending on league position.
Firstly, as expected the scatter graph below shows there is a significantly positive relationship between possession and goals scored throughout all Premier League teams r=0.81, P < 0.01. Predictions can be made as to which teams score the most goals based on possession stats. (Anybody unfamiliar with correlations there is a brief explanation of what the r and p values indicate below)
For people unfamiliar with correlations I will briefly explain. The r value indicates how strong the correlation is. The closer to ‘one’ the greater the correlation. The closer to ‘zero’ the weaker the correlation. The same applies for negative values so the closer to ‘negative one’ the stronger the negative correlation. If the P value is less than 0.05, this means it is a significant correlation. So it’s less likely the results are due to chance.
The most efficient teams for scoring the most goals in relation to possession lie above the line of best fit. They are Reading, West Brom, Newcastle United, Tottenham, Chelsea, Man United and Wigan Athletic. All of the other teams are more inefficient, lying below the line of best fit.
Likewise, a significant correlation was found between possession and goals against, however this was a negative relationship r =-0.56, P< 0.05. This correlation is a lot weaker, as shown in the graph below, than the one for goals scored however it still shows that generally higher possessions can predict fewer goals conceded.
For goals against the more efficient teams are those that lie below the line of best fit. Everton, Stoke City, Man City, Chelsea, Man Utd, Arsenal, Tottenham, West Brom, Sunderland, West Ham and Norwich City make up these teams. They all have the best ratio in terms of possession per goal conceded.
Our final graph based on all teams combined shows a highly significant positive correlation between possession and goal difference r =0.77, P <0.01. This shows a steeper gradient than the other two however it has a lower correlation than goals for. The results are also more scattered with many outliers so looking specifically at goals for/against may be more reliable.
Moving on from looking at the league as a whole, I direct my attention to looking more specifically at teams. For the graph below, I have split the teams into four groups. Each consisting of five teams. So the top five makes up one group (A), the next five make up another (B), next five (C) and bottom five (D).
The higher teams statistically have a greater relationship between possession and goals for than the lower teams. (A: r = 0.96, P< 0.01. B: r =0.87, P< 0.05. C: r =0.54, P= 0.35. D: r = 0.28, P= 0.64.) Only groups A and B had a statistically significant positive relationship (Top half teams). Both C and D had small positive relationships, but neither was significant, so predictions can’t be made due to a high risk the results are due to chance.
A more efficient team in terms of possession related to goals cannot be concluded to be the greater team. Possession could be used as a defensive tactic when leading a game, or teams may have greater efficiency when build up play is slowed down and not rushed.
[quote]All of the stats from this article have been taken from the Opta Stats Centre at EPLIndex.com – Subscribe Now (Includes author privileges!) Check out our new Top Stats feature on the Stats Centre which allows you to compare all players in the league & read about new additions to the stats centre.[/quote]