23 Jul What Does an Off-Ball Run Force the Defence to Do?
An attacker can change a defensive structure without touching the ball.
A winger moving inside may force the full-back to engage. A forward dropping between the lines may become the responsibility of a midfielder and then a centre-back. A run behind the defence may oblige one defender to leave the line and continue tracking the attacker.
The pass may never arrive.
Yet the defence has already had to react.
This raises an important question for coaches and analysts:
Can we measure how attacking movements redistribute marking responsibilities across a team?
Our recent study proposes a tracking-based model that detects temporary marking relationships between opponents and analyses how those relationships are distributed across players and teams.
The key idea is simple:
Teams defend space first. Attacking movements activate marking responsibilities inside that spatial structure.
Zonal defending still creates individual marking relationships
In open play, most teams in LALIGA defend primarily through zonal principles.
Defenders do not normally begin with a fixed opponent whom they follow throughout the pitch. They protect spaces, maintain distances with teammates and move collectively according to the position of the ball.
A back four basculates, advances or retreats as a line. Its first reference is the collective structure.
But zonal defending does not mean that defenders never mark individual opponents.
When an attacker enters a defender’s area of responsibility and becomes an immediate threat, that defender must control him. If the attacker continues into another zone, the responsibility may pass to another teammate. In some situations—particularly a run behind the defensive line—the original defender may have to continue following the opponent beyond his usual zone.
There are also exceptions in the match plan. A team may assign an individual marking to a specific player, use a mixed system or introduce man-oriented behaviours in particular areas or phases.
However, the general interpretation remains:
The defender protects a space. The movement of the attacker determines when a direct marking relationship appears.
The model should therefore not be understood as a representation of permanent man-to-man assignments. It detects temporary moments in which a defender is close enough to an opponent and moves in a sufficiently similar direction to indicate direct defensive control.
How does the model detect marking?
A marking interaction is identified when two conditions occur simultaneously:
- the defender is no more than two metres from the opponent;
- both players move in a sufficiently similar direction, with a maximum difference of 30 degrees.
The attacker does not need to be in possession of the ball.
The framework was applied to optical tracking data from 99 LALIGA matches from the 2019/20 season, recorded at 25 frames per second.
This makes it possible to reconstruct continuously:
- which defender is controlling which attacker;
- how long the relationship lasts;
- which opponents repeatedly enter each player’s responsibility;
- and when several defenders are controlling different attackers at the same time.
Most detected episodes were very brief, frequently lasting less than one second.
This supports a dynamic interpretation. Many of these relationships are not permanent individual markings, but short defensive responses to players entering, leaving or moving between zones.
The important question is not only who marks whom
Once these interactions have been detected, the model allows us to examine three football questions.
How many defenders are activated at the same time?
Several attackers may simultaneously occupy spaces close to the defensive line.
One may pin the left-back. Another may position himself between the centre-backs. A third may attack the opposite half-space. Each movement activates a different defensive responsibility.
The model measures when several teammates are controlling opponents simultaneously, even when they are not marking the same player.
This does not automatically mean that the team is defending well. Several simultaneous marking relationships may indicate:
- effective control of multiple threats;
- strong occupation of the defensive structure by the attacking team;
- several defenders being pinned;
- or a defence that has become excessively focused on opponents rather than space.
The data identify the configuration. The video and tactical context explain what it means.
In the study, these simultaneous marking relationships represented approximately 0.5% to 5% of individual defensive time, with defenders showing the highest values.
Their duration declined rapidly as the number of players involved increased. Configurations involving two or three teammates were much more frequent than situations in which five or more players were simultaneously controlling opponents.
This creates a useful attacking question:
Which movements and positional structures force the greatest number of defenders to assume direct responsibility at the same time?
That could help analyse overloads, occupation of the last line, rotations, movements between zones and attackers who pin defenders without receiving the ball.
Which defenders repeatedly deal with the same attackers?
Two defenders may control the same attacker at different moments.
A winger may initially be controlled by the full-back and then become the centre-back’s responsibility after moving inside. A striker may move laterally across the defensive line and interact successively with both centre-backs. A midfielder dropping towards the ball may pass through several defensive zones during the same possession.
The model identifies which teammates repeatedly deal with similar groups of opponents.
The strongest similarities were found between defenders.
This should not necessarily be interpreted as evidence of shared individual assignments. In a zonal defence, it may reflect attackers moving through neighbouring spaces and responsibility being transferred between defenders.
From a practical perspective, this can reveal:
- attacking players who affect several defenders through their movement;
- recurrent movements between adjacent defensive zones;
- relationships between full-backs and centre-backs;
- and parts of the defensive structure repeatedly activated by the same opponents.
The analytical question therefore changes.
Instead of asking only:
Who marked the forward?
We can ask:
Which defenders and which defensive zones did the forward activate during the match?
Which players deal with the widest variety of opponents?
The model also distinguishes between players who repeatedly control a limited number of rivals and those who must continually adapt to different opponents entering their area.
The study found that midfielders dealt with the widest variety of opponents, while defenders generally concentrated their interactions on fewer players.
This result reflects an important tactical difference between positions.
A centre-back usually protects a relatively stable sector and interacts mainly with forwards operating near the last line.
A midfielder works in a more fluid environment:
- forwards drop between the lines;
- wingers move inside;
- attacking midfielders rotate;
- full-backs advance;
- and different players arrive around the ball.
The midfielder may therefore accumulate less total marking time than a defender while having to respond to a much more diverse set of opponents.
This distinction matters for performance analysis and scouting.
Two players may record a similar volume of defensive interactions but perform very different roles:
- one repeatedly controls the same direct threat;
- the other constantly adapts to different players moving through his zone.
A simple total cannot distinguish between them.
The model also measures attacking influence
Although the framework was developed to study defensive marking, every interaction is produced by both teams.
The defender responds, but the attacker’s movement activates the response.
The model can therefore offer information about offensive influence without the ball.
In the study, forwards accumulated the greatest amount of time being marked, and every player in the top ten for total marked time was a forward.
That finding is unsurprising, but its tactical meaning extends beyond direct possession.
A forward may influence the defence by:
- pinning a centre-back;
- forcing two defenders to share or transfer responsibility;
- moving across several defensive zones;
- threatening depth and preventing the line from advancing;
- creating space for a teammate;
- or activating several defenders while the ball remains elsewhere.
A movement does not need to receive the ball to have value.
An off-ball run can reorganise the defence before the pass is even considered.
The model does not yet determine whether that movement creates a genuine advantage. It does, however, make the defensive response measurable.
Practical applications for coaches and analysts
The main value of the framework is not to replace video analysis. It is to identify which moments deserve closer examination.
It could help analysts locate:
- attacking sequences that activate several defenders simultaneously;
- players who repeatedly move through different defensive zones;
- areas where marking responsibilities change frequently;
- defenders exposed to an unusually wide range of opponents;
- pairs of defenders who repeatedly deal with the same attacking threats;
- and prolonged followings that differ from the team’s usual zonal behaviour.
It may also support comparisons between opponents.
The same defensive line may produce very different patterns against:
- a positional attack with relatively fixed roles;
- highly mobile forwards;
- inverted wingers;
- overlapping full-backs;
- frequent exchanges between striker and attacking midfielder;
- or teams that consistently overload one side.
The model can show how those attacking structures alter the distribution of defensive responsibility.
This may help answer questions such as:
- Which opponent forces our centre-backs to leave their usual zones most often?
- Which movements repeatedly activate the full-back and centre-back together?
- Which attacker affects the greatest number of defenders without receiving?
- Which sector of our structure is occupied most frequently?
- When do we stop protecting space and begin following opponents?
What the model does not determine
The framework does not measure the complete quality of a zonal defence.
It cannot decide automatically whether:
- the line basculated at the correct moment;
- the distances between defenders were appropriate;
- a player should have followed or released an attacker;
- a transfer of responsibility was successful;
- several simultaneous markings represented defensive control or disorder;
- or the attacking movement created a real tactical advantage.
A well-organised defence and a disorganised defence may both produce several simultaneous marking relationships, but for very different reasons.
The model measures who became directly responsible for whom, when and for how long. Tactical interpretation still requires information about:
- the ball;
- the positions of teammates;
- the height and width of the block;
- defensive cover;
- available space;
- and the subsequent outcome of the action.
Future developments should connect these interactions with ball progression, entries into dangerous areas, recoveries, shots and goals. The paper itself identifies the need to relate the measures to defensive effectiveness and match context.
From protecting space to assuming responsibility
The most useful contribution of the model is not simply that it detects marking.
It helps describe the relationship between attacking movement and zonal defensive responsibility.
The defence begins by protecting space.
The attackers move through that space.
Those movements force different players to assume, transfer or maintain direct control.
The model makes it possible to quantify:
- which defender becomes responsible;
- how long that responsibility lasts;
- which teammates repeatedly deal with the same opponents;
- how many defenders are activated at the same time;
- and which players must adapt to the widest variety of attacking threats.
This leads to a different way of evaluating movement without the ball.
The question is no longer only:
Where did the attacker run?
It becomes:
What did that movement force the defence to do?
Read the full paper
The complete study, “Beyond marking networks in soccer: Coordination, similarity and entropy,” by Calero-Sanz, López del Campo, Resta and Buldú, is available in Chaos, Solitons & Fractals. The paper provides the full methodological framework, network definitions and detailed results derived from tracking data from 99 LALIGA matches.
Full paper: https://doi.org/10.1016/j.chaos.2026.118819