SkillCorner: Evaluating progressive passing of defensive midfielders

"I think passes completed is the most useless statistic there is. I can have 96% pass completion playing sideways or backwards without being dangerous for the whole game. That means I haven't done anything. What's the point?" (Kevin De Bruyne, May 2023)

Elite teams need players at the base of midfield capable of making passes that penetrate and unbalance the opposition, creating opportunities to score.

In this article, we are going to focus on the ability of defensive midfielders to penetrate an opposition defence based on the movement of their team-mates using SkillCorner's Game Intelligence data. These are the three pass types we are going to use to represent this ability to make progressive passes:

1. Passes to runs In behind
2. Passes to switch play
3. Passes to players in space between the lines

The charts below rank the Premier League’s defensive midfielders for each of these metrics (normalised per 30 minutes with their team in possession):

The limitation of pure counts, even if adjusted for possession differences, is that they can be due in part to a team’s style of play or an individual player’s role. Defensive midfielders in ball-dominant teams tend to have more opportunities to get on the ball, a variety of team-mates offering to receive and therefore more opportunities to play each of these types of passes. Those playing in more transitional or direct teams tend not to.

Similarly, a defensive midfielder playing in a double pivot alongside a deep-lying playmaker may be instructed to recover possession and move it sideways to the playmaker, rather than playing forwards themselves.

The players at the top for each of our three metrics were playing either in ball-dominant teams (Jorginho, Rodri, Enzo Fernandez) or alongside a ball-dominant playmaker (Oliver Norwood, Christian Eriksen, Billy Gilmour). Or, in some cases, both.

So how can we separate a player’s passing output from their team’s style of play and their individual role? SkillCorner Game Intelligence does this by measuring the passes that a player actually made, while using tracking and contextual data to measure the number of opportunities they had to make a progressive pass, as you can see in the video below.

This enables clubs to measure a player’s tendency to play each category of pass by adjusting for the number of opportunities they had to play them. Frenkie de Jong, Joshua Kimmich and Rodri are elite defensive midfielders playing in three of the most ball-dominant teams in Europe (Barcelona, Bayern Munich and Manchester City respectively). All three are in the 80 to 90^th percentile for most of the progressive passing.

However, adjusting these counts as a percentage of the opportunities they had yields some interesting results:

Conversely, João Palhinha and Éderson are below or around average on the pass attempt counts (again, even adjusting for team time in possession), but when their number of opportunities to play each type of pass is factored in, it shows that they’re taking a very high proportion of those opportunities to penetrate the opposition defensive structure.

This comes with an element of risk/ reward and the profiles of de Jong, Kimmich and Rodri are perhaps reflective of controlling play rather than actively taking risk. Clubs can use these metrics to identify players with a tendency to play line-breaking passes when the opportunity arises, as illustrated in the plot below.

The combination of on-ball event data and player tracking data also enables us to measure the difficulty and impact of a line-breaking pass. The plot below displays the percentage of line-breaking passes that were difficult, along with the number of opponents bypassed:

As with any single datapoint, these measurements can only form part of an assessment of a player, and they need to be placed in context of a player’s individual role and the style and situation of their team. However, the additional depth and context of these metrics can generate interesting signals for recruitment staff to identify potential or tendencies missing in surface level data.

SkillCorner’s Game Intelligence product combines and synchronises event and tracking data to create integrated physical, technical and tactical metrics, forming the world’s broadest and deepest football performance dataset for both on-ball and off-ball analysis.

On-ball: Understand passer tendencies and evaluate passing decisions

• Line breaking passes / opportunities
• Dangerous / difficult passes
• Execution under different levels of pressure

On-ball: Identify players that create movement on-ball

• Initiate give and goes
• Receive backward and play forward
• Possession retention when under different levels of pressure

Off-ball: Find players who support the ball carrier effectively

• Options between lines, in space
• Off-ball runs broken down by type and threat
• Off-ball runs that lead to shot and/or goals

Off-ball: Combine tactical and physical insights

• Speeds, trajectory, angles for different types of movements
• Sequences of off-ball runs to create threat

Game Intelligence provides these objective insights and more - for any player across 100+ global competitions – bringing an unprecedented depth of insight in player recruitment and match analysis.

SkillCorner is the leader in tracking data and analytics for football based on a single-camera video source. Its AI-powered technologies provide rich and accurate player tracking data at scale, enabling teams to efficiently identify and benchmark players from all over the world.

More than 200 clubs, leagues and federations around the world trust SkillCorner data to help them make smarter decisions in scouting, player evaluation and analysis.