Separating offensive and defensive performances

What makes Heracles and RKC the two extremes of the Eredivisie? What area does Marco van Basten need to work on at Heerenveen? Why does PSV stand out from the rest of the title contenders? And why is Vitesse not a true contender for the title?

The recently introduced method of computing Relative Shot Rates allows us easy answers to all of these questions. This method allows to separate offensive from defensive performances by looking at relative numbers of shots created and conceded.


Why are shots better than goals?

The simple answer is that shots are nearly ten times more frequent than goals. This means that it takes much more time to collected enough data when studying goals instead of shots. The average amount of goals scored by an Eredivisie team is around 1.6 per match, while the average amount of shots is around 13. In other words, the amount of shots during the 5th match round equals the amount of goals at the end of a 34-match season.
We know that converting shots into goals and preventing your opponent from doing so is a skill that is hard to repeat. Superior teams don’t separate themselves from inferior teams with regards to shots conversion and saves percentages, but simply by creating more shots and conceding less.


Why the relative amount of shots?

So, better teams create more shots and concede less. It’s no secret that some teams excel by creating shots, while other specialize in conceding less, but the it’s only the top teams that combine this skill.
However, by only looking at the absolute amount of shots, another problem creeps up.
Suppose that your team is a relegation favorite and you want to compare their total shot rate (TSR) to a relegation threatened rival team. Now, your team faces the league leaders early in the season, while the rivals only meet them near Christmas. It’s easy to imagine that your team’s shot rate over much of the first half of the season looks worse than their performance actually is. Only when the other team has also been thrashed by the league leaders the comparison becomes fair again. The Relative Shots Rate (RSR) compares the performance of a team with the performance of other teams that have played the same opponents. Details of this method were explained earlier.

The X-axis of the plot (horizontal) depicts the teams’ offensive performances, while the Y-axis (vertical) depicts the teams’ defensive performances. Note that the offensive performances show a wider range of variety than the defensive performances. For clarity, the defensive record is shown inverted, so that better defenses (that concede fewer shots) are higher in the graph. Better offenses are on the right hand side of the graph, so the upper right hand corner is the area of optimal performance, while the lower left hand corner shows the struggling teams.

It is immediately clear that PSV sticks out positively. They are matched in defensive terms by title rivals Feyenoord, and at a small distance Twente and Ajax, but PSV’s offensive record is second to none and allows them five shots per game more than even the best of the rest. The media runs a nice narrative of PSV’s defensive problems, but their defensive record shows that they are up there with the best of the Eredivisie.

An interesting position in this graph is RKC in the upper left hand corner. Erwin Koeman’s team takes a completely different approach compared to lower right hand corner Heracles. RKC matches the top teams in terms of limiting their opponent’s number of shots, but they pay for it by sacrificing their offensive chances to the level of relegation strugglers Roda, NAC and Willem II. Heracles, on the other hand, are second in terms of shots creation, but concede way too many shots in the process.

Van Basten’s Heerenveen has a situation that is quite comparable to Heracles. They can boost the third best offensive record, but only four Eredivisie teams give up more shots, which illustrates the imbalance in their team. Expect teams like Heerenveen and Heracles to be involved in high scoring matches, like Heerenveen’s 4-4 draw today, and expect RKC to play in low scoring affairs.

So, why is Vitesse not a true title contender? See for yourself, their offensive record is slightly above average (+0.25), but they concede 0.67 shots more than the average Eredivisie level. Expect Vitesse to regress as the 2012/13 Eredivisie comes near its end. Their overachievement so far will probably propel them into a play-off position, but be cautious to expect more on the basis of their present shot records.


In the end

We will regularly revisit this Relative Shots Rate concept, as it seems the best way to asses team performances. Studying shots rather than goals provides a much more solid base by generating nearly ten times higher numbers. On top of that, the Relative Shots Rate eliminates the bias that is otherwise always introduced due to differences in strength of schedule.

Next up: separating home and away performances…

2 thoughts on “Separating offensive and defensive performances

  1. Jeff Jefferson

    “Superior teams don’t separate themselves from inferior teams with regards to shots conversion and saves percentages, but simply by creating more shots and conceding less.”

    Is this really the case? Can the over performance of Vitesse not simply be explained by the fact that they have certain players (MostlyWilfried Bony) who are sustainably capable of maintaining a higher conversion rate than other teams.

    1. 11tegen11 Post author

      This is very much a counter intuitive phenomenon, but there is little doubt that this is true.

      The bottom two graphs in this post by James Grayson illustrate very well that both shooting percentage and saves percentages don’t correlate with points in the league standing at all…

      This post on 11tegen11 analyzed the Eredivisie teams of 2010/11 and 2011/12 with respect to PDO, which is nothing more than conversion and saves percentage combined. The graph at the bottom leaves no doubt about the fact that no correlation exists at all between two seasons.

      This post about Bony, written a few weeks ago, will demonstrate that conversion rates as high as his 70% (14 goals from 20 shots on target) are not sustainable. In fact, since that post was written, Bony has converted one of his six additional shots. His conversion rate of shots on target has now regressed from 70% to 58% and it is likely to drop further towards the usual numbers of around 35%.


Leave a Reply