My after season team performance analysis continues.
This week I’m taking a look at Clearances. A much used individual statistic that many rely on to rate the value of central defenders. But does it add real value? Are defenders with more clearances better than those with fewer clearances?
My research looks at clearances from two different perspectives:
- What is the relationship of clearances by the opponents’ relative to a team earning points, and
- What is the relationship of defensive clearances by your own team relative to your team earning points.
For Major League Soccer, 2016, the number of clearances each team had, and their opponents’ had, was counted per game.
I then did a simple correlation on the number of clearances (per game) relative to the points earned (per game). Pictured below is a summary of the two perspectives relative to their correlation to points earned.
- Diagram 1: The average number of opponent clearances (per game – right part of the diagram) has a (-.30) correlation to points earned.
- Diagram 1: The average number of defensive clearances (per game – left side of the diagram) has a (+.30) correlation to points earned.
- Diagram 1: It’s pretty clear that the correlations vary (considerably) from team to team.
Average of opponent’s clearances per game versus points earned:
- Sporting KC gained the most benefit from lack of opponent clearances throughout the season; their correlation was (-.48). In other words Sporting KC were more inclined to earn points when the opponent had fewer clearances. This seems reasonable, especially since Sporting KC offered up the second most crosses (19 per game) this year. The less likely the opponent was in clearing those crosses the more likely Sporting KC had in converting those crosses to goals scored.
- DC United got the least benefit from lack of opponent clearances; their correlation was (-.08). In other words the number of clearances by their opponent’s, throughout the season, had little to no overall impact in DC United earning points. This also seems reasonable since DC United offered up the third fewest crosses (14 per game) this year. With not many crosses offered it seems reasonable that this mode of creating scoring chances was less likely to occur.
What’s that mean?
- For me, I would offer it means the number of defensive clearances an opponent has, per game, isn’t really a strong team indicator.
Average of defensive clearances per game versus points earned:
- San Jose gained the most benefit from defensive clearances (.57); meaning San Jose were more inclined to earn points when having more defensive clearances per game. This seems reasonable as San Jose faced an average of 20.5 crosses and over six corners per game; tied for 8th most in each category across the league. A higher volume faced should result in a higher volume of clearances.
- New York Red Bulls gained the least benefit from defensive clearances (-.01); meaning the Red Bulls were just slightly more inclined to earn points when they had fewer defensive clearances (per game). What is unusual with New York is they averaged a greater number of defensive clearances (21 per game) but faced fewer crosses and corners than San Jose.
What’s that mean?
- For me, I would offer it means (again) the number of defensive clearances a team has, per game, doesn’t greatly determine the outcome of a game.
- If neither opponent defensive clearances per game, nor your own teams’ defensive clearances per game, don’t have a strong correlation to points earned then the individual player statistics – that make up those clearances’ statistics won’t have much value either.
- If anything – given the wide variation in clearances’ value, relative to points earned, a players’ individual clearances (per game) should be weighted relative to that game – and that game only. Recognizing that the ‘weight’ of those clearances is subject to change every single game.
- Perhaps what’s really missing here is the volume of “clearances not made” instead of “clearances”?
- Finally, as a ‘giggle check’ if-you-will, I did take a look to see if the correlation of clearances was over .50 relative to the number of opponent crosses and corners offered – it was. The average correlation across the league was .71 – quite strong… see Diagram 2 below.
- So our own common sense is supported by data analysis.
- Said differently; “common data sense” shows the volume of clearances are related to the volume of crosses and corners.
- Therefore… (in my view)
- If “the common data sense” (shown in Diagram 1) does not show the volume of clearances having a strong relationship to earning points then our own common sense should follow that view.
- Again reinforcing that individual defensive clearances, as an effective individual statistic, does not add real value at all.
COPYRIGHT – All Rights Reserved. Trademark: PMP