top of page

In Defence of Possession

It goes without saying that the use of possession stats in football will split opinion. There are some that will wax lyrical about a team’s dominance in a game due to having a high percentage of the ball. On the flip side there are those who feel ball retention is underrated, unnecessary and in some cases negative.


Of course, there are merits for both arguments, but generally the focus usually falls on the impact high levels of ball possession has on attacking stats as trends suggest that teams who play a possession based game tend to create more goal scoring opportunities. However, thanks to an inquisitive request on Twitter from @Piethagoram, we will use this article to try and identify any correlation there may be between ball possession and the impact it might have on defensive resilience, and whether it can provide a barometer for defensive performance.


To assist us in our analysis we will be focusing on the 2022/23 National League season as there are a number of diverse styles of play as well as some key outliers, none more so than the likes of Wrexham and Notts County who took this season by storm as well as teams like Boreham Wood, Chesterfield, Woking and Southend United who recorded impressive defensive stats.


All data used is from Wyscout.


The first thing we will do is see how ball possession impacts shots against for each team. This is displayed in the graph below where the ‘y’ axis represents shots against per 90 minutes and the ‘x’ axis the average percentage of ball possession.

Chart | Shots against vs Possession | National League

The first thing we are drawn to is Notts County who incredibly averaged a fraction under 70% ball possession for the season and as such invited an average of just 6 shots against per 90 minutes. This suggests that Luke Williams’ possession based style of play not only ensured they were the highest scorers in the league (117 goals in the regular season), but also helped them record the lowest number of shots against and the second least number of goals conceded, 42.


At the other end of the scale Torquay United recorded the third lowest ball possession of 44.60% and the highest number of shots against, 14.07 per 90 minutes.


On the face of it there is an argument then that high ball possession equals fewer shots faced, and likewise less possession means more shots faced.


Well this appears to be going well, or is it? We identified earlier that there were a few outliers and both Notts County and Torquay fall into that category. However, another we should consider is Dorking Wanderers.


Marc White's team recorded the second highest possession in the league, enjoying 55.90% of the ball. However, they also invited the second highest number of shots against, 12.33 per 90 minutes.


Despite Dorking’s unusual situation, when we add a trend line to the graph it’s clear that overall, the number of shots against reduces as the possession percentage increases and that the majority of clubs fall within a small footprint of this. Maybe we’re onto something!

We want to try and confirm if there is any direct correlation between possession and defensive resilience. One way we can do this is by comparing goals conceded by shots faced. Essentially the higher the shots faced to goals conceded would suggest defensive resilience.


We’ve already identified a trend that the higher the possession the fewer the shots faced. This would indicate therefore that Notts County have a better defence because they restrict the opposition to fewer goal scoring opportunities.


However, when you compare the goals conceded per shots faced, we get a very different result as shown in the chart below.

Chart | Goals conceded vs shots faced | National League

Luke Garrard’s Boreham Wood are well known for their defensive strength, something definitely confirmed by this analysis. Despite inviting just over 10 shots against per game they actually only concede every 12.00 of those attempts on their goal.


If we then revert to our first graph it’s clear that The Wood have very low possession stats which kind of blows a hole in our possession based theory and reinforces that Garrard puts a greater emphasis on how he sets his team up defensively over ball retention.


This contrasts remarkably with Notts County whose defensive resilience is defined by their possession play and despite restricting the opposition to just 6.07 shots per 90 mins they also concede every 6.80 shots against so on the face of it their defensive resilience is very much reliant on their ability to retain the ball for large periods of a game and suggesting that without it they would be less resilient defensively.


Okay, this is proving to be a little more complicated than first imagined.


Is there really any way we can use possession as a barometer for defensive resilience? Well, we have identified trends that allude to this, but it doesn’t detract from the fact that ball possession is a by-product of a chosen playing style and tactical approach. Essentially to make any real sense of this we might be better served by defining defensive resilience in the style of play rather than just possession.


Looking back at the above graph it appears, on the face of it that Boreham Wood are the most resilient in terms of defending as it takes the opposition over 12 attempts to score against them, two more than the number of shots they face per game.


Meanwhile Maidstone United display the least resilience allowing the opposition to score every 5.85 shots, this is very problematic as they faced an average of 12.17 shots per game.


One last thing we can do is look at the impact, if any, possession has on goals conceded and divide the results into four sections to help clearly identify where each team falls.


Chart | Goals conceded vs possession | Nation League

It’s interesting to note that Wrexham and Chesterfield, along with Notts County fall into the ‘High Possession/High Defensive Resilience’ range as these were the three teams that finished in the top three. In fact, Wrexham, who were the eventual league winners, managed a record breaking season with 51% possession, 19% less than County.


Southend did well in restricting opponents to just under a goal a game with 53.1% possession and Gateshead will be hoping the foundations they’ve put in place will stand them in good stead for next season and have backed this up with the recent signing of ball playing midfielder Regan Booty.


In fact there are several clubs in and around the ‘average zone’ that are determined to play a passing game, clubs like Altrincham, Wealdstone, York, Barnet and Halifax, and it would be fair to suggest that if their possession stats were improved slightly they could see themselves conceding less, and in return becoming more competitive next season. However, how they set up out of possession will be just as important as what they do with the ball in possession.


In conclusion it’s evident that if there is any correlation between possession and defensive resilience it’s relatively vague. Yes, in most cases the more of the ball you have the more you restrict goal scoring opportunities for your opponents, but as we found out with Dorking, this isn’t always a guarantee of improved defensive performance.


In one extreme we have seen how Notts County’s high percentage of ball retention has worked well, but we need to bear in mind that despite restricting the opposition to much fewer chances they haven’t necessarily been the most resilient at the back and they were very much the exception to the rule.


By contrast Boreham Wood have less possession, invite more shots against but concede the least goals per game making them appear to be the most defensively resilient team in the league . . . and who would argue with that?


 

About Analysts Bar

Originally founded in 2021 by fellow Notts County fans Colin Sisson and Richard Ogando, Analysts Bar offer data analysis and talent identification services to the footballing industry and related media.


Colin and Richard recently welcomed popular National League Blogger and Analyst Tom Williams to the team ensuring they can continue to provide a comprehensive range of football data services and analysis to clubs, players, and coaches at all levels.


Get in Touch

Twitter: @AnalystsBar


Find us on Fiverr here

Recent Posts

See All

Opmerkingen


bottom of page