Corner Possession Index: measuring a team’s possession quality after an attacking corner.

In the past year, I have been looking more and more at data related to set pieces, and corners to be more specific. It is hard to look at the defensive side of the data because it is primarily focused on attacking numbers. I want to look deeper into the defensive data and I will do that by looking at “next actions”. By doing so I will gain valuable insight into possession actions by a team.

In this article, we will look closer to the Corner Possession Index (CRI). This is a metric that I created for teams. It focuses on the team’s ability to keep possession after a corner with emphasis on the difficulty of non-pass actions.

Why do we need a metric that measures this ability?

The Corner Possession Index (CPI) highlights a crucial but often overlooked phase in modern football: retaining possession after taking an attacking corner. Set-pieces, particularly corners, are pivotal moments that can lead to scoring opportunities, but they also pose a risk of losing control and triggering counter-attacks. CPI measures a team’s ability to maintain possession following a corner, ensuring continued pressure in the attacking phase and reducing the likelihood of conceding dangerous transitions.

Teams that excel in this aspect sustain their attacking momentum, creating more chances and keeping opponents on the defensive. On the other hand, teams that frequently lose possession after corners expose themselves to counter-attacks, forcing them to reconsider their set-piece strategies or implement adjustments to better control the aftermath of these high-stakes situations.

Actionable analysis: Individual Header Rating (IHR) determines choices in blockers vs runners

What do I need to create this metric?

I have to start with the data. The data comes from Opta and is event data from the 2024–2025 season and was collected on November, 8th 2024. This data is raw, but I will use Python to assess the specific metrics that I will use to create something new. There are a few things I need before I can create the CPI:

1. Attacking corner delivery data
2. Direct contact or not
3. What happens with the corner in the next 3 actions: loss/remain of possession within 1–2 following actions.
4. What type of possession is most freqyent

All the teams are featured in the 2024–2025 Primeira Liga and all players are available for these metrics, only exceptions are the goalkeepers. They are excluded because they don’t play a significant role in this part of the game.

Methodology

The process begins by identifying all attacking corner events for a team. After each corner, we analyze the following events to see if the attacking team remains in possession.

First, we will have a look at the corners taken, to assess how many times possession is lost or won from first contact:

I have chosen for Porto in this example, to focus on one specific club to make the visualisation slightly more workable. You can see all corners taken, but now we will have to look at corners that were successful in first contact, and to see what the next actions were:

This allows us to see what the next 5 actions are after successful corner pass. This means that the pass will reach a teammate and he/she will have an action. That action is NextAction1TypeId. And the following is NextAction2TypeId and so forth.

This gives us something work with as we are trying to calculate the CPI, which is the following formula:

Finally, we calculate the CPI score by looking at the ratio of corners with successful deliveries to the total number of attacking corners taken. Of course this formula looks at the sustained possession, so in this case we will look at the next 3 actions and call that sustained possession.

Corner Possesion Index — CPI

When we calculate the sustained possession and calculate the CPI for Porto we get the following information:

59 Corners have been taken of which 30 were successful. That gives a success rate of 50,84%. If we look at the next 3 actions of those successful corners we don’t see anything different in the CPI. Why? That’s because in this case, Porto has retained possession in the next 3 actions when they had successfully taken corner. However we can weigh the different means of possession, for example, if the first action is a pass, I want to filter that out, because that means the corner is taken short or passed to a free teammate

This gives u weight of 0.5 for passes and others have a weighing of 1.

Using AutoRegressive Integrated Moving Average (ARIMA) to predict future shot locations for…

The difficulty of non-pass actions is higher, so that’s why we made this distinction. So if we take that original CPI and weight it, we won’t get 50,84%, but different:

So the corner possession index is 50,84% without weights and with weights it is 44,92 for Porto. CPI for Porto is 44,92

Final thoughts

The Corner Possession Index (CPI) is a new metric that offers a deeper understanding of a team’s ability to retain possession after an attacking corner. By focusing on this transitional phase, CPI highlights a crucial, often neglected aspect of set-piece analysis. Retaining possession after a corner not only increases a team’s chances of creating further opportunities but also reduces the risk of conceding dangerous counter-attacks. We looked at the quality of non-pass actions as they determine how well a team can handle them.

For Porto, the analysis revealed a base CPI of 50.84%, meaning the team successfully retained possession after 30 out of 59 corners taken. However, when weights were applied to account for the difficulty of non-pass actions (weighted at 1) versus passes (weighted at 0.5), the weighted CPI decreased to 44.92%. This adjustment underscores the importance of considering the type of possession actions in assessing a team’s efficiency.