Redefining the keeper position

There’s nothing new about a sweeper-keeper. A goalkeeper who plays as a sweeper and supports the team in possession is quite common in modern football. Most teams demand technical profiles while recruiting keepers who are calm in possession right from the moment the ball is rolled into play and the opponents press them well into the team’s own defensive third. But Tim Walter takes it a step further and uses his keeper in a role similar to the powerplay phase in futsal. I take the case of Daniel Heuer Fernandes, the goalkeeper at Hamburg, to explore this positional innovation deeper.

A look at Fernandes’ involvement depicts how far he drifts outside his penalty area leaving his goal open to participate in possession. The players’ average positions during the game shows the goalkeeper’s average position well outside the box.

Meanwhile from his passmap, we see the number of passes he plays at the base of his defensive structure, often slotting in as an extra defender during build up. The passes are played from almost all areas outside the box in the defensive third, and occasionally, even the middle third.

An extreme moment was against Bremen with Hamburg under pressure to equalise, had their GK play key passes in the opponent’s half in the final minutes.


The principles of play involving the keeper is based on the rotations that are trademark of Walter’s teams. Both at Kiel and Stuttgart, his teams build up from the back using dynamic rotations that constantly create positional superiority in the midfield and create a free man to play out of. Understanding the movement these rotations create along with its underlying principle sheds more light on the specific demands of the GK as a libero in certain moments.

Let’s take a situation where the CB on the right side is in possession of the ball building out of a 4-2-3-1 and the opponents have picked a moment to press them in a ball-oriented manner closing all near options and denying them an opportunity to switch sides.

The opponent CF closes down the ball carrier, the LM the RB and the CM the pivot. The CB could play the ball to the other CB in the adjacent channel, but the AM is not far off to close him down either. In most situations, the CB would either look to play the ball long to the forwards or play back to the GK to regain superiority at the cost of moving backwards.

However, in Walter’s rotations, the CB plays the ball sideways and then pushes up. By doing so the CB immediately creates positional superiority behind the first line of press. Simultaneously, a player from the midfield, the CM or a fullback who had previously pushed up, drops lower to become the free man. These rotations are quite similar to the 4-0 or 3-1 rotations in futsal where positional superiority outweighs numerical superiority since the area is smaller and it’s always 4v4.

Notice how in this movement, the GK slides over to the right to create a better angle in possession. GKs by instinct try to always protect their goal and hence position themselves always along a line connecting the nearest threat and the centre of their goal, the shortest route back to safety in case of an error in possession. But in this game model, the GK has a more sophisticated role to facilitate possession by creating the right passing angles. By sliding to the right, the GK slowly shifts up into c3 providing numerical superiority.

Following this rotation, the opponents have two choices – either to continue pressing high or to hold and protect the centre. If the opponents press high as in this case depicted below, the space in the middle between the lines opens up. Here, the AM positioned in c4 is free to receive a line-breaking pass and turn towards goal to attack. Numerical superiority is held in c3 with a 3v1 against the opponent CF.

If the opponents fall back with the aim of protecting the spaces in the centre and denying room to receive between the lines, the build up continues through the free man in the rotation and Hamburg are able to push higher up using their GK. Numerical superiority is still maintained and the CM can always pass it back to the GK for safety.

The objective is to win superiority in each zone – be it numerical, positional or qualitative. Had the GK been positioned deeper in his area, the opponents would have been more encouraged to press the players in c3 despite the movements to generate the free man, forcing the play back to the GK. However, since the GK is higher up in c3, he immediately provides that superiority won through short passes. Instead of having the players protect the goal and maintain possession, in this philosophy, it’s the superiority generated that does the job. Therefore, it doesn’t matter where the GK is positioned as long as the team wins superiority.


The keeper often supports the central defenders by acting as a sweeper between them. This libero role, seen in the game model of many other teams, enables the CBs to split wider apart and push higher up in search for a progressive pass without the risk of being outnumbered by the opponent forwards who press. Hamburg also create a back-three while building up with Fernandes playing libero between his CBs.

The distinctive attribute though, of Fernandes’ role as a sweeper-keeper, is how high he pushes out of his box, almost into the middle third as we see below, leaving his goal wide open. This allows forward players to occupy strategic spaces giving positional superiority.

The build up is maintained using rotation of players at the back creating constant triangles or diamond structures to outnumber the opponent forwards who press.

Even though the keeper leaves his goal unattended, the belief of the team is that there isn’t a threat as long as they have possession of the ball. With this belief, every time Hamburg is in position, the goalkeeper quickly transforms into an outfield player and plays with his feet outside the box.


A known fact is that the team in possession has the possibility of using their keeper as a passing option, but the team out of possession cannot use their keeper as a pressing option. Hence the team in possession always holds overall numerical superiority. This logic has been used in building out of the back with the keeper to ensure a clean ‘exit’ out of the opponent’s high press. Tim Walter takes this a step further and involves the keeper not only in the build up phase but also the construction phase in the centre of the field. This effectively results in an overall numerical superiority for Hamburg in the most important part of the pitch.

Involving the keeper in the build up in pivotal role almost as an extra CB means one player can push further up to occupy a positional role. This is a huge advantage as the extra player has the liberty to position between the lines or allow one of the forwards to pin back the defensive line instead of dropping back to help out in the build up. The fluid rotations allow the players to exchange roles dynamically through their movement.

In this example above, the GK helps win 3v2 superiority in zone b3 while the players in c3 are engaging in rotations during build up. Further up Hamburg have three players pinning back the defensive line, two offering width in the wide channels and two players positioned at different heights between the lines. Having the extra player really helps the team to hold positional superiority further up giving them a big advantage in possession.

It definitely carries the risk of inviting a moment of high man-oriented press where a loss of possession almost certainly results in giving away a cheap goal. However, the positional advantage is quite big for Hamburg, and if built up well, the opponents struggle to press one player short and are constantly pinned back having to defend all the spaces in the middle.


The need for goalkeeper to occupy the right spaces in possession to facilitate better passing angles is usually limited to the penalty area when they are attempting to play out of a high press. Once the team is further up, the GK usually plays as a sweeper to collect loose balls played out. In Hamburg’s system the positioning of the goalkeeper is even more crucial to be involved so actively in keeping possession outside the penalty area. We’ve seen evidence of this in his passmap earlier as well.

Take this case below, as his CB picks up the ball in the outside channel, the GK needs to occupy the adjacent channel by drifting further wide and high rather than deeper as a GK would conventionally do.

This positional behaviour, similar to what a CB would do, is important to facilitate the right passing angle to build out. It also allows the outfield players like the RB to stay higher up and wider in the outside channel. Had the GK sat deeper, it would have encouraged the opponent to jump right away.

Positioning also allows to create more advantageous situations of numerical superiority in general. It’s intuitive to understand that a 3v2 superiority holds more advantage if the players are positioned in a triangle rather than a straight line, or a 2v1 if the players are at a good distance and not directly behind each other. The same logic applies if the keeper is one of the players creating the superiority.

In the example above, since the CB is in one of the inside channels, the GK creates more advantage by occupying the other vertical channel rather than if both are positioned in the same channel. Thus the GK has to move wider like another CB to make a better passing angle to surrounding teammates.

We see a similar case in the game against Bremen as well, where in a very unorthodox manner, the GK leaves his spot in front of goal to offer the right passing lane in the adjacent channel and help build up with a 2v1 superiority.


Opponents hesitate to accept the invitation to press with an open goal ahead because the extra player positioned somewhere deeper can immediately create a threat in transition with one vertical line breaking pass.

Look at this situation below where the opponents have a potential 2v2 pressing situation in b3. The two forwards can jump and press the GK and the CB with an open goal in behind.

But if you notice in the second frame, the CF still scans to check the space he leaves behind if he presses because when playing against a team having that extra player somewhere, you simply can’t allow passes through the middle as the entire back line is pinned back by forwards.

We see the consequences of leaving spaces in the centre to jump in the next example. Here the opponents decide to engage in a high man-oriented press with the GK outside his area.

Hamburg manage to play out of the press and a few seconds later engage in a an attacking transition with the opponents’ back line pinned 4v4. The costs of being unsuccessful in the press is conceding crucial spaces in the centre with Hamburg’s forwards in position to transition with numerical and positional advantage.


Walter’s system is a fresh innovation that redefines the role in which goalkeepers have been used in the game. Although teams that like to dominate with the ball are already familiar with the ball-playing keeper who is comfortable with the ball at his/her feet, Hamburg’s employment of the keeper in the attacking phase opens up a Pandora’s box in how spaces can be exploited when they have the ball. They tend to dominate most of the possession, and having the extra player further up immensely helps in pinning back the opponents, who are reluctant to jump out of their structure to press.

In Hamburg’s case, the results do not correspond to the game model they like to play with because of lack of finishing when in front of goal or getting caught out of organisation defensively. Regardless of this, Walter’s model at Hamburg offers plenty of ideas for teams looking to dominate possession in the same way, especially during build up and construction and are worthy of greater attention in tactical circles.

The dynamic nature of space on the pitch

Space and time are the two most commonly utilised words in the vocabulary of football. The game of football involves players with personalities and physical attributes who are constantly moving, and their positions affect the free spaces which the opposition looks to recognise and exploit in order to progress the ball towards the goal. Therefore, success in the game is attributed to the accurate interpretation of these spaces by players during the game. 

The coach has the responsibility of conceiving a game model through which all the players of the team can easily recognise those spaces. The interpretation is however not merely an individual objective. Success in executing the game model or plan also demands that a group of players, or ideally the whole team interpret the same spaces in the same way. This means that at any moment in time, two or more observers in different positions on the pitch must interpret the nature of the same space with fairly high consistency. As a coach or an analyst, your role is to see spaces in the same manner that players do, despite having a very different vantage point off the pitch. If you want to convey your ideas using the lexicon of space in football, it is imperative to understand the nature of those spaces to achieve similar consistency in their interpretation. 

Football analytics has progressed to the point that using tracking data and pitch control, it is able to formulate the mathematical space occupied by every player on the pitch to great accuracy. Nevertheless, when it comes to analysing how the game is perceived in the mind of a player, current mathematical models have a lot more variables still to consider to accurately interpret spaces the same way that players see them on the pitch. 

Abstractness of space beyond geometry

The issue with attempting to achieve an objective analysis of space in football is its abstractness. Space on a pitch doesn’t exist the way it is usually depicted by coaches on a tactic board. Every player who has played the game understands this. A coach’s tactic board represents the players as 22 identical circular objects and their positions define the space available to play. For example, space between the lines is represented by the area of a trapezium bound by the positions of the four defending players (two defenders and two midfielders). When a coach illustrates this space at half time in the dressing room, it is a static two-dimensional eucledian space determined only by the cartesian coordinates of four identical circular board magnets.

During the game, space in terms of its physical properties is not static. Space is dynamic. It is not only bound by the positional coordinates of the players at that particular instant but also their momenta. The speed and direction along with the body profile of each player influences each individual player’s momentum which subsequently influences the space held by the players. Furthermore, apart from the spatial variables, there is also the variable of time. Space in football exists only at a certain moment in time and is rapidly changing in relation to time. This dynamic property of space means that it is always transforming. Hence, to understand spaces, it is necessary to understand how the spaces are transforming with time and not just their geometry at any instant in time. 

There are other variables that also affect the space such as the momentum of the ball, and the weather conditions and the pitch conditions. Subsequently, if the player looks for a teammate to pass the ball, the momentum of the teammate and the body profile are also taken into consideration to calculate the space to play a specific type of pass. During the game, players constantly perceive spaces on the pitch taking into account all these variables intuitively and simultaneously.

We call this “phase spaces” and is defined by: where the ball is, what situation it is in, where the opponents are, the distances between the ball and the opponents and our own players, the trajectories made by each player and each opponent and the ball, the way the game is oriented, the organization that the game has. And all of this constitutes only a game situation that lasts a tenth or two tenths of a second. The moment the ball changes from site, change the players and a new situation appears. And likewise successively.

― Paco Seirul·lo,
on his theory of ‘Phase Spaces’ (Espacios de Fase)

In the first example with the help of tactical illustrations, we see a situation where the CB who has received the ball from the left side of the pitch is looking to switch the play to the right. Positioned in the inside channel, the CB has two players in the outside channel, the RW and the RB.

The game model requires the RB to create superiority behind the line of press by moving into a position between the lines. For the RB to offer a passing option to the CB, the space between the lines needs to be interpreted. In this situation, we commonly define the space between the lines as the area of the trapezium whose vertices are the four opponent players – a, b, c and d. This is how we expect the coach or analyst to illustrate this space to explain the game model on the tactic board.

However, during the game, the players do not perceive this space as a static geometrical shape. It is perceived dynamically by also considering the momenta of the four players – pa, pb pc and pd respectively. Note that the teammates, the CB carrying the ball and the RW are also in motion and have their individual momenta pCB and pRW respectively, which influences the behaviour of the opponents a, b, c or d and subsequently affects the space for the RB to receive the pass.

The momentum of an object is conventionally represented by the letter p. It is the product of two quantities, the object’s mass (m) and its velocity (v): p=mv

Thus, we see that the space between the lines the RB and CB perceives is a dynamic entity in motion that undergoes instantaneous transformation. For the RB to receive the ball successfully, we have to consider how the space changes in relation to the movement of the RB with a momentum of pRB and the motion of the ball, pball, when the pass is played. The space changes if the opponents react to the actions of the RB and CB.

Hence, these two need to pick the right moment in time to execute the action while the space remains relatively constant due to the inertia of the moving opponents. (Newton’s first law of motion or law of inertia states that an object in a state of rest or motion continues to do so unless acted upon by an external force). Note that the position and movement of the RW in this situation is necessary for partly maintaining the space as constant for a longer time period by pinning back the defensive line formed by c and d

Now let’s look at this tactical illustration one frame further, after the pass has been played. The manner in which the actions of the CB (the passer) and the RB (the receiver) influences the nature of the dynamic space and their perception of it, determines the actions that follow. Depending on how the RB perceives the space, he/she may decide to receive it with a touch and play it directly to the CF facing the same direction of momentum, or turn and look to play a deep pass into the space in behind on the outside channel for the RW. 

As with the spaces between the lines, bound by the positions of the opponents, the spaces in behind are also influenced by the momentum of the players. If the previous example demonstrated how the motion of players of the same team influences the space, here we see the case of space being influenced by two opposing players. For the space behind the opponent d to be viable, it is necessary for d to jump with a momentum pd (press leaving the position in the defensive line) and for the CF to be in diagonal momentum pCF to attack the space. The sum of these opposing motions (along with the position of the last defender d) impacts the viability of the space. Similarly, the space on the outside channel is influenced by the momenta of pc and pRW, both who are competing to win the same space. In this case, the one who is quicker to recognise the space and react by accelerating to it wins the space. For the sake of consistency of using the jargon of physics, here acceleration (or more accurately – Force, f) is defined as the rate of change of momentum with time. Players respond to changing situations by rapidly accelerating and decelerating using the forces generated by their musculature.

The representation of players as circular objects having the same physical properties leads us into believing that they are able to move in all directions with the same magnitude and quality. Although it serves for simplicity while explaining tactics, I posit rather that the players are better represented having an anatomical orientation. This way it is easier to understand through intuition that a player can accelerate in a forward direction faster than sidewards or backwards. It also factors the law of inertia which provides a better interpretation of spaces in tactical illustrations.

The representation of the same example as above, but with anatomical orientation of the players conveys more information about the nature of spaces during the game

I dive into another example to illustrate how body orientation and motion affects the nature of space. Let’s take a 2v2 situation of a CM progressing with the ball in attacking transition against defenders a and b, with a teammate LW on the left.

The CM targets to play a pass through the defenders targeting the wide space on the outside channel. He/she predicts that the LW, despite being onside while the pass is played, has sufficient forward momentum to meet the pass in the targeted space. Let’s assume that the CM, right footed, decides to play the pass on the ground with the inside of the foot that curls inwards into the path of the LW. 

Although the defenders a and b are separated by a physical distance of 10 metres, the space to play the pass through isn’t defined by the mathematical value of 10 metres. In the first case, we look at what happens when the defender a is tracking back with a momentum of pa and the defender b who is facing the oncoming CM decides to step forward with a momentum of pb. The space to play the pass through is closing in and effectively less than 10 metres considering the time it takes for the ball to pass through the defenders with a momentum of pball.

In another case, we see the defender b tracking back the run of a third player CF with a forward momentum of pb. Here, the space to play the pass is expanding although the distance between the defenders at this moment is still 10 metres. Considering the momentum of the ball pball and the defender pb, the pass will pass through with greater success because of a bigger gap. Thus, we see how the body orientation of the players and their directional momentum influences spaces on the pitch.

Contemporary football is familiar with the use of Voronoi diagrams to mathematically calculate the space occupied by each player at any instant. In mathematics, a Voronoi diagram is a partitioning of a plane into regions based on distance to points in a specific subset of the plane. Using tracking data of players positions during the game, coaches and analysts hope to objectively identify spaces in opposition structures that they can exploit and adjust their players positioning during different phases of the game better. 

Above is a Voronoi diagram of two teams, the blues in attacking phase playing entirely in the half of the reds who are currently defending. I would like to zoom into a specific part of the pitch on the left where the left winger from the blue is positioned on the shoulder of the right back from the reds. 

The mathematical Voronoi generated suggests that the left winger occupies a major area of the outside channel including the space behind the defence compared to the right back, a. But if we look at the body profiles of the players along with their directional momenta instead of viewing them as circular points (from which the Voronoi was generated), we see that the LW is only able to gain superiority in the space in the outside channel with a forward body orientation and an initial forward momentum pLW which gives the LW an advantage over the defender a with a back/sidewards momentum pa.

It is easy to understand this scenario intuitively from the perspective of the CM who is on the ball. For the CM, the space behind the defence is only viable if the LW is in forward motion towards that space. On the contrary, if the LW is oriented backwards without initial motion, facing in the same direction as the RB, the CM cannot play a pass to gain the same advantage as before.

In this case, the CM doesn’t view the space in behind in the same way, although the positions of all the players are exactly the same as they were earlier. As the body orientation of the player is different, the subsequent spatial configuration should also change so as to favour the defender a than the LW. Here I have manually adjusted the geometry of the Voronoi to portray the disadvantage for the attacking team purely because of the body orientation and momentum factor.

The Voronoi diagram does not take into account the position of the ball, the speed of the players or which way they are facing, which invalidates any conclusions that are drawn from it. Pitch Control, an alternative created by William Spearman, Lead Data Scientist at Liverpool FC, is much better at estimating the probability of successful passes, and evaluating the passing options of a player.

It is undoubtedly a great tool to view how the players occupy spaces in any instant and analyse their probabilistic pass options. Nevertheless I find it still far from predicting how players interpret spaces during the game. The science calculates the variables that take place on the pitch as they happen, but players anticipate what will happen in the future time. A forward can make a double movement and the space he/she wins as a consequence is not in the direction of the initial movement, but rather the second movement. The player has seen this space and anticipated the movement long before executing it.

How players conceive space-time models

So far I attempted to explain the dynamic nature of space and how players perceive them during the game. But the game is played by displacement of the ball. As the players are mostly occupied with tracking and technically manipulating the ball, they aren’t able to observe this dynamic transformation of spaces the way you have just visualized previously as an analyst. Instead, through periodically scanning their surroundings, they construct a panorama in their heads with multiple images. This panorama helps them conceive a space-time model. A space-time model is a prediction of how space-time will transform in the near future. The closer the model is to how the space-time is in reality, the better the player is able to choose the right action during the game.

One can never know with perfect accuracy both of the two important factors which determine the movement of an object — its position and its momentum.

— Heisenberg’s uncertainty principle

As a player if you scan something in your surroundings moving at a particular momentum p, even when you look away you are able to predict its position after a certain time t. A player is constantly predicting the positions of everything in the surroundings – teammates, opponents and even the ball – while not observing them. Assuming that everything on the pitch obeys the laws of classical mechanics, the simple kinematic equations of motion helps calculate the same results that players do intuitively without any background in physics! Furthermore, the results are not as important as the ideas of space-time that are conceived during the game.

Kinematic equations of motion in physics

Let’s look at an instance of the blues looking to progress the ball up the pitch with the CB carrying the ball out of the back.

The CM is currently a free option to pass the ball to. For the CM to recognise that he/she could be a progressive outlet for the CB, the CB needs to estimate the nature of the space by scanning the surroundings.

Upon scanning, the blue CM observes the opponent red CM starting to press with an initial momentum pCM. This pressing action in the surroundings effectively changes the space that blue CM initially had to receive the ball. 

Although the blue CM has scanned and identified the oncoming press from behind, while looking away towards the ball, the CM needs to predict how much space he/she will have by the time the ball arrives. This would depend on both the momentum of the marker pCM and the ball pball and their respective distances.

By conceiving a space-time model that predicts what will happen by the time the ball reaches the CM, he/she will decide whether to play the pass back first touch, or to turn and progress the ball further to the AM. The space-time model also gives a sense of where the pressure is coming from which helps orient the CM with the right body profile. The accuracy of this model ensures a successful action that is both safe in terms of keeping possession and advantageous to the blue team. This is also an example of contracting space as a result of markers closing in. 

As another instance of space-time modelling, I take the instance of the RB in possession of the ball on the right side of the field. The opponents engage in a ball-oriented press trying to forcing the RB to either play the long ball or play backwards. The CM on the right outside channel steps forward to provide a narrow outlet for the RB in the space bound by the surrounding markers. 

On the inside channel, the AM is man-marked by the red CM. So, when the AM tries to move behind the line of pressure in front of the red defence, the CM follows closely. This creates space centrally in the inside channel behind the red CF who angles his/her run to press the situation on the right. The blue CM on the right who steps up, scans and observes the motion of the AM clearing up the space and the other CM on the left who is moving forward into that space behind the red CF. 

When the blue CM on the right receives the pass from the RB, he/she can securely play the pass behind the press of the red CF into the space being created. The space-time model predicts the motion of the players affecting the space and presents a viable opportunity for the CM to play the pass to the fellow CM in the centre. The initial scan detects the momentum of the fellow blue CM and predicts his/her position in the next frame where the pass can be played to break through the press. Although in the previous frame the space appears to be smaller, the movement of the AM and the marker away from it and the angled motion of the CF shows how the space is expanding.

Players use predictive space-time models to manipulate the perception of space-time by surrounding players. Take the example of la pausa, where a player delays his/her action for the space to become more viable and advantageous. This skill is a demonstration of how space-time modelling enables a player to predict how the play will develop in the near future and thereby manipulate the positioning of the surrounding players and their subsequent perception of space-time.

The relative nature of space-time

Relativity of space-time accounts for the idea that space is subject to the perception of an observer. The players who play the game are the active observers and thus their respective interpretation of space has an element of observer bias. How much space and time a player perceives he/she has is influenced by a lot of factors, but most predominantly by his/her scanning behaviour. Each scan gives an opportunity to capture an image of the surroundings where a picture of the surroundings reaches the observer at a speed of light. 

Space and time are relative — they depend on an observer’s speed. But the speed of light is more fundamental than either.

― Einstein’s theory of special relativity

Since the observers (players) are in motion, the perception of space and time is relative to each observer. Constant, efficient and active scanning enables players to model their space-time predictions as close to reality as possible and consistent to one another. Ultimately, when a pass is played between two players it is necessary for both of them to interpret the spaces similarly for the action to be effective.

The technical abilities of the player also plays a big role in the perception of space and time. Technically better players perceive having more time and space than technically weaker players in similar situations during the game. Perception of space is also dependent on player profiles and positions. A central defender tends to perceive more time and space being on the outside of the positional structure than a central midfielder or a forward. Similarly wingers or fullbacks perceive the spaces on the outside channels differently. By virtue of players getting habituated to constantly perceiving familiar spaces because of their regular positions, they tend to shape their playing profile and instincts accordingly. A midfielder might sense the need to make a quicker decision and switch the ball to the other side because of contesting for tight spaces among opponents and focus on collective superiorities, while a winger tries to buy time and isolate opponents in wide areas to take them on in 1v1s more frequently. 

Fundamentally, I believe that the ball is the only thing that is absolute in the game. Everything else is relative, subject to the interpretation of the players who play the game. Although the geometric standards of the field, and it’s canonical elements like the dimensions of the penalty box or the boundaries of the goal are also absolute and irrefutable, the spaces within them generated by the players’ positioning and their subsequent actions are interpreted relatively.

This explains the technical challenge of scoring a goal versus playing a pass. A pass, even a long one, can be off by a couple of metres and still be counted as accurate because the receiver has the freedom of space to adjust and adapt to the margin of error and still receive it successfully. A shot on goal, however, if off by even millimetres results in rebounding off posts or a fingertip save by the keeper. Even the referees and linesmen cannot be a hundred percent objective in their interpretation of the players actions on the field, since they all are in motion, and are limited by their respective vantage points. Therefore I believe that only the ball is absolute, much like the elusive truth we constantly seek in our lives. Everything else ― Space, time, actions, tactics, justice or motivation ― is just an illusion. On the pitch, I, as a player, am trying to sell you the illusion that I have more space than you, the referee my perspective of justice, my coach a belief that the tactics are working and the spectators the motivation of collective triumph through a game of football.

Without the ball there is nothing; the ball is the mother, the source of life in football. What’s the goal there for? For the ball to go in it. Without the ball, nothing has any meaning.

― Juanma Lillo

Chaos or order is just a matter of perspective

I believe that the abstractness and relative nature of space should help us appreciate the power of perspective in the game, and in the universe. I reflect upon an anecdote a good friend of mine once shared: He said, “In our daily lives, we appear to be living in chaos and the nature of events manifest a great deal of unpredictability. But if we zoom out to the level of planet earth, we observe a relative order how the earth spins on its axis each day so consistently. Then we think about the geoclimatic changes, the great land masses and oceans moving on the planet and find another level of chaos, but if we zoom out to the level of the solar system, we find more order in the structured orbits of all the planets. And we can extrapolate this nature further on to the Milky Way or the expanding universe and notice the same phenomena. The variables of space and time are ultimately a question of perspective.” 

Similarly, playing the game of football is like being in the midst of chaos. The events happening around the player are so dynamic and unpredictable, and no player recounts having experienced the game in ‘real time.’ But sitting in the stands high above the level of the field, we question why the pass wasn’t played to the winger who has been positioned in free space for the last two minutes. Even further away, watching the game on a screen sitting on our couches we notice far less of what happens during the games, doused by the relative order of the same attempts to attack and score that our team has been attempting for the past hour.

The objective of a game model is to offer players cues to recognising the same spaces efficiently through common principles of play. To develop as a coach or an analyst, we must embrace the chaos that the game presents at the level of the players who play it and not hide behind the order of observing the game that a privileged perspective offers. Understanding spaces in the game as dynamic entities rather than static geometry helps us appreciate how the game is actually played and what makes it beautiful.

Brighton’s beautiful game

Brighton and Hove Albion is a team that has clearly caught my eye this season in the Premier League. Under Graham Potter, the Seagulls have shown leaps of progress in the previous season already. Some of their commendable performances this season were the second halves against Manchester City and Liverpool, two of the strongest sides not only in the English Premier League but also in Europe. Brighton fought hard to achieve a draw against Liverpool, and despite conceding a couple of goals against City in defensive transitions, they were electric in attack often putting the heavyweights on the backfoot. The game against Leeds United away was one of the most exciting games of football with plenty of chances despite ending in a draw.

Just as the manager himself was, I was quite surprised too to notice the reaction of the Brighton fans after the game for the kind of football they play and their current position in the league table. Although the lack of goals and wins are evident, his side is highlighted by a clear set of principles that define their playing style. They are a possession-based team that build up with superiorities and press aggressively upon losing the ball. In this analysis I attempt to dive deeper into some tactical aspects of Brighton under Potter referencing moments from their last draw against West Ham United.

The attacking phase: patient buildup and construction

In the attacking phase they build up patiently. In fact, Brighton are the slowest team in the league at progressing the ball up the pitch, as they average just 1.04 metres per second of progression.

They are tactically flexible in terms of formations. They are comfortable playing either 4–2–3–1 or 3–4–2–1. The common aspect of both systems is the way they construct their attack. The goalkeeper is the first attacker and is comfortable with the ball at his feet. In many situations he drives outside his box with the ball at his feet as a deep libero. This allows his two CBs to drift wider. This is a very important principle because the wider and higher the CBs are able to move, the better they are able to connect with the advanced players through line breaking passes. They are also better connected to the fullbacks with short passes, and in general this ensures better width to the structure coming out of the back.

The first passing option out of the buildup is the two pivots who are crucial to the Brighton system in orchestrating the attacking phase. The two central midfielders occupy the two inside channels and try to be positioned at different heights. From here they can form triangles with the FBs and the CBs and advance the ball further up the pitch through short passes that break lines of press. Yves Bissouma is a fundamental member of the two-man pivot system, and he is usually paired with either Jakub Moder, Pascal Groß or sometimes Adam Lallana.

Against many teams, the two pivots are tightly marked, in which case Brighton still want to maintain their fundamental of patient buildup through short passes instead of launching the ball long. They just look for other options. Sometimes in a 4–2–3–1, the FB drops to buildup with 3 players which is a familiar system anyway when they play 3–4–2–1.

In most cases, Brighton prefers to have their fullbacks higher up and wide to provide width. If the pivots are not an immediate passing option, they look for alternatives to play the short passes through the centre. Here comes into play the dynamic role of the attacking midfielders and forwards. Having so many players in advanced positions, one of them can easily drop deeper to offer an outlet through the inside channels while forming triangles with surrounding players who can act as the third man to get the ball out beating the first line of press.

In both systems 4–2–3–1 and 3–4–2–1, Brighton have atleast two attacking mids and one centre forward who play between the lines in front of the opposition defence. Neal Maupay, alongside Leandro Trossard or Alexis Mac Allister aren’t tall physical forwards who hold up defenders physically as target men. Neither are they the quickest sprinters who target the space behind the defense with runs in behind. Instead, they prefer to play between the lines making complementary movements that make space for each other with their positioning. The three forwards in the centre form a triangle. If one makes a run against the defensive line forcing them back, the other drops between the lines making space to receive a pass, while the third moves out wide and overloads the wide channels with the FBs.

The FBs bear the responsibility of connecting the attack and the defence with short passes, overloading the wide channels in attack offering width and immediately squeezing the space with the counterpress during defensive transitions. Having an engine like Cucurella is very handy in this system whom you find in one instance pressing outside his box, and in the next instance making the run on the far side for the switch of play.

The CBs are an important part of the possession style of play in both attacking with line breaking passes and defending with close range support. Brighton have ball-carrying CBs who are comfortable pushing up into the midfield if they find space.

The defensive transition: counterpressing

The key principle upon losing the ball is to immediately apply pressure. Brighton have the third highest successful pressure rate of 30.8%. Having a numerical superiority in every zone with players positioned to help each other out with short distances from each other plays a massive role here. In fact, the longer Brigton’s defensive transition lasts, the more dangerous the situation turns out for them with most such events leading to shots on goal conceded bringing the keeper into action. Brighton’s PPDA (passes allowed per defensive action) is around 11.2 which is fifth highest in the league behind City, Liverpool, Leeds and Aston Villa. With the CB positioned so high up the pitch, they prefer to win the ball back as far as possible from their own goal.

Brighton’s tackles attempted shows they prefer to counterpress higher up the pitch and away from the centre or around their penalty area (source:

Overloading the midfield with both pivots and forwards dropping deeper and CBs pushing higher enables them to create the superiority they need in specific zones. The other reason they prefer to win the ball back early is because they require their FBs and wingers positioned high and wide. A defensive transition forces the wide players back with the task of defending.

Brighton prefer to exploit players like Cucurella for his attacking qualities rather than his defensive ones. During a defensive transition, if Cucurella has the liberty of holding his position higher up because his central players are most likely going to win the ball back, it helps Brighton transition better into attack should they win back possession since they can always switch the play with a diagonal long ball.

With the goal of pressing they opposition high, they also pick specific moments defined by cues that trigger a high man-oriented press with the entire formation pushing up. The objective is to force the opponents to lose the ball or clear it long.

Positioning in zones

Brighton can be described as a team that likes to play very positionally both in the way they attack as well as defend. The players make sure the distances between one another are short creating lots of triangles for fluid passing networks in possession. They try to minimize their defensive transitions to the extend that the subsequent phases can be looked at as an extension of the existing phase.

In the zone of intervention, where the action around the ball takes place, the players are tasked at immediately putting tackles in to disrupt the opposition gaining control of a possible attacking transition. The players around the zone of intervention are positioned close to each other in a zone of mutual help to offer support through short passes in case possession is won back. Through these short passes, Brighton reestablish their rhythm and can then look to move the ball into the zone of cooperation where players are positioned strategically to advance the ball further up into the final third.

The fluidity of Brighton’s attacking system also has a structure to ensure its stability. For example, the two pivots who are always required to support the zone of mutual help by moving closer to the action. The FBs always must provide width in attack by staying wide. The forwards need to constantly drop between the lines to offer free passing lanes but at the same time be positioned behind lines of press to create superiority during possession. The CBs have to move wide to form the zone of cooperation for the ball to be switched, in which case the GK forms a link between them as a deep libero.

What is lacking?

Brighton’s disappointment this season is most definitely the final third. For the style of football they play and the chances that they create, they simply don’t convert them to score enough goals. This causes viewers and their own fans to question the practicality of their system of play. The lack of being clinical in the final third causes an understandable uncertainty in having all that amount of possession. They need to find more ways of arriving into the final third both centrally and wide. They also need to find ways to attack the space behind the defence quicker when the defensive line is higher up the pitch.

The other Achilles heel has also been longer stretches of defensive transition. They leave huge spaces behind to defend and this causes them problems against oppositions with quick, physical forwards who can exploit moments of counterattack and convert them. Although I’ve only analysed their open play, Brighton also concede from set-pieces which definitely needs some homework to sort out.

At the end of the day what matters in football is results regardless of the style you play. Currently Brighton play far better than what the results reflect about their performances. Graham Potter’s principles would be definitely better represented if they pick up their goalscoring form as well. Nevertheless, they are an exciting team to watch for their possession-based style and I sincerely hope that the pressure to achieve better results doesn’t cause them to sacrifice their flamboyance on the ball.

The 3–3–1–3 formation: Its audacity and its frailty

For anyone diving into football tactics, getting a grip of the 3–3–1–3 is a temptation that is hard to avoid. After all, it was popularised by an unorthodox Argentine nicknamed El Loco who, many top managers like Guardiola, Sampaoli and Pochettino, regard as their mentor and the greatest coach of the game ever.

Marcelo “El Loco” Bielsa — The coach who popularised the 3–3–1–3

It’s a hipster tactic. It screams in the face of conventional formations in modern football. When teams pick a safe strategy with which they can park the bus, the 3–3–1–3 looks like a jet plane. You rarely find guidelines to play a 3–3–1–3 during your coaching pathway. Anything if at all would be found on the internet written by some FM geek who experimented with it or self-proclaimed tactical buffs like me. At this point, however, I admit that I actually know very little of it despite having been obsessed with it like everyone else.

I gave it a practical shot myself in my early days of coaching a women’s team few years ago. It was a cup quarterfinal against a team that plays in the UEFA Women’s Champions League and I felt adventurous. There was nothing to lose. The game was open, end to end. But by half time, my pivot (and my best player) was fuming in the dressing room: “This is a disaster, I feel completely outnumbered in the middle and I can’t find one key pass. Our defence is all over the place.”

After watching Klich against Southampton this Tuesday, I can imagine why. Bielsa has unleashed the 3–3–1–3 at every club he’s been since the very first game — Bilbao, Marseille and Lille. But at Leeds where he has served the longest of all, he took a more cautious approach since the beginning with the 4–1–4–1 being his stock formation for most of this time in Yorkshire. The few games in which he has used the 3–3–1–3 has been a palpitating episode to say the least for a Leeds fan.

3–3–1–3 Overview

The 3–3–1–3 employs:

  • 3 centrebacks that stick quick close to each other,
  • 1 central midfielder that roams around in the space in front of the defence,
  • 2 highly mobile and versatile wingbacks that take up any role depending on the phase of the attack and
  • the attacking unit, enganche y tres puntas comprising of one creative attacking midfielder behind one central striker and two wide wingers
Bielsa uses the 3–3–1–3 in his first ever league game with Lille in France
Slide from Bielsa’s presentation on which formations counter which ones according to his theory of the game

From Bielsa’s presentations, it is clear that he chooses the 3–3–1–3 formation against two opposition formations, the 4–3–1–2 and the 4–2–2–2. The mathematical reasoning behind this choice is known only to him, but the decision to have one more defender than the number of opposition strikers makes sense. Both the formations by the opposition use 2 strikers, so having 3 defenders gives immediate numerical superiority while building up play from the defensive third.

So far this season, he has used the 3–3–1–3 twice before Southampton, against Burnley and Sheffield United. Both games were played with close margins, with Leeds winning 1–0 and 0–1 respectively. In the game against Southampton Leeds secured a triumphant 3–0 win, although Bielsa claimed in his press conference post-game, “The margins were closer than what the result suggests.”

Attacking organization

The centre forward is the focus of every attack in this system due to the lack of density of players in the middle. The first phase of buildup coming out of the keeper is straightforward and based on numerical superiority. The defensive flat back-3 provides sufficient width to deal with the press from the 2 strikers. The following three phases of buildup — construction, creation and finishing — however, are quite direct and follow rapidly. The reason is that the outlet out of the first phase to bypass the first line opposition pressure is only via the wingbacks.

The central midfielder in front of the defence is often either tightly marked or finds himself in a position of heavy inferiority. Hence, progression to the construction phase is initiated by the wingbacks either directly by a pass, or indirectly by starting a third-man run. As the play moves outwards from deep, the opposition finds it easy to apply high pressure forcing the play wide since there are no options back into the middle. The wingback is forced to make a quick decision.

An experienced utility player like Ayling or Dallas is able to play the pass quickly forward in the same lane to the winger ahead and cut inside into the half space. By beating the wide player pressing with a quick one-two pass, the wingback now creates an overload in the half space along with the centre forward and the attacking midfielder. The central midfielder now has time to cover defensively in case possession is lost.

Simultaneously, on the opposite flank, the other wingback can drift inside forward with the winger staying wide and create a 2v1 against the opposing fullback. This movement pins the defensive back line and forces them back, but it needs to be played quickly and the switch immediately. The players need to mobilize rapidly and attack the spaces in front. A miss pass or interception by the opponent midfielders risks a 4v4 turnover with the 3 centrebacks left isolated to manage the defensive line themselves.

Superiority in the middle can also be created by the forward dropping deep. An extra player in midfield provides support to the central midfielders. The winger can now attack the space left by the centre forward and the wingback can overlap to pin back the fullback and overload the same side. Once again the ball needs to be played directly without spending too much time in the middle with short passes. Most often we see direct balls and quick plays into the final third in the 3–3–1–3 system. In the game against Southampton, Leeds had 55 long balls and 277 short passes versus Southampton’s 38 long balls and 337 short passes. Leeds also had lesser percentage of possession.

It is important for the attacking midfielder to be positioned between the lines up ahead and not play too deep. This is probably why the 3–3–1–3 can be differentiated from a regular 3–4–3 (like the one used by Conte at Chelsea) because the latter makes use of a double pivot in the middle with the two midfielders playing more box-to-box and covering greater combined area in the centre.

When the attacking midfielder plays between the lines he can always function as a third-man off a direct ball to the centre forward. When he receives the pass, the defence is forced to step back to secure the spaces behind them. This buys him time to turn and look to switch play. If the wingback on the opposite side pushes up and inside to support, along with the winger they create a 3v2.

The centre forward and the winger on the other flank can pin back their markers with runs in behind and this gives space for the wingback on the other side to make an overlap giving the player on the ball options to attack from both sides. At the end of the finalizing phase, the team has width as well as players in the box to finish from a cross. We see the dynamic versatility demanded from the wingbacks who are required to function as either midfielders or wingers depending on the situation. This explains how Dallas created the opportunity to score the second goal for Leeds against Southampton in this game.

Defensive organization

The moments of transition are quite intense in this system. The priority of defensive organization is to delay and win the ball back as soon as possible. The more time the opposition is allowed on the ball in the centre, the greater the risk of conceding shots on goal. Illan Meslier needed to have a good game to keep a clean sheet against Southampton. He made 5 crucial saves which is greater than the average of 4 saves he makes per game.

Leeds’ man-to-man marking is the hallmark of the defensive organization. They close down all options in the middle, and the centre forward and wingers cut passing angles using their cover shadows. In this system, if the opposition creates a free man in the middle, one of the 3 centrebacks needs to step out of his line to close down the player.

Timing plays a big role here as to make this decision, the centreback needs his wingbacks to be in place to cover the backline. With the wingbacks always in transition, there is always a risk of running a scenario of inferiority against the backline as the opponents counter attack.

If you notice in this instance, Leeds have all men marked in the middle and in order to create this man-to-man marking, Struijk, the centreback has to also step up and pick up a free man. Then, the left winger and centre forward are able to cut passing lanes and force the play back to the goalkeeper. The vulnerable points of the 3–3–1–3 structure are always the spaces right in front of the defence and the spaces left behind by the wingbacks pushing forward.

Lack of central stability

An understanding of the system from the previous sections provides a hint as to why the 3–3–1–3 lacks central stability. If we look at the heatmap against Southampton we see a U-shaped occupation of the area around the 18-yard box in Leeds’ own half.

Leeds’ Heatmap against Southampton (source: WhoScored)

But the crucial areas in the middle around the centre circle are empty. For a possession based team, it is crucial to control the centre of the pitch and this explains why Leeds have lesser possession than usual in the 3–3–1–3 system. It is a formation better designed for direct play.

Leeds’ action zones against Southampton (source: WhoScored)

Leeds also spent a lot of time trying to build out of from wide areas in their own half using their wingbacks to progress into the next phase of attacking organisation. A central pivot who is primarily assigned as a defensive midfielder like Kalvin Phillips holds responsibility of constructing the attack after surpassing the first line of pressure.

Leeds’ touchmap against Southampton with halfspaces highlighted along with low control around the centre (source: WhoScored)

We can clearly notice Phillips’ absence especially looking at the areas of occupation. In the touchmap above, notice how the 3–3–1–3 tends to dominate play along the wide channels and in deeper areas. We see a lack of density of touches along the half spaces or inside channels. The 4–1–4–1, with a second row of four, also helps control the centre better and it is impossible for the single enganche to dominate the space around the centre circle in the 3–3–1–3 system.

The two central players, Klich and Roberts had in fact, the least number of touches in the team during the game (Harrison was substituted at half time). Klich usually averages a pass accuracy of 82% but in this game he only managed 13 accurate passes out of 19 with a pass accuracy of 68%. The central players have very little time to make the right decision, always having to play in numerical inferiority and need to play the right pass in two touches at the most.

Roberts (in midfield) provides the assist for Bamford’s goal against Southampton with his second touch

Roberts’ assist to Bamford for the first goal was played with the second touch, almost falling off balance attempting it. The central midfielder in front of the defence is simply overwhelmed by the area to cover as a box-to-box player. Without constant support from the wingbacks, it’s nearly an impossible position to play, and the wingbacks always need a certain cue from the opponents to decide their move. It is no surprise now why my pivot was left helpless at half time when she played this position, despite the attacking opportunities created.

3–3–1–3: conceptual or practical?

The success of this system relies on the ability to be brave and dominate the wide lane despite being pressed to the touchline. The linear stack of three players in a narrow lane need to engage in dynamic rotations to overcome opponents one by one. Bielsa’s four core principles: concentración, permanente movilidad, rotación y repenitización (concentration, permanent focus, rotation and improvisation) echoes louder than ever in the 3–3–1–3.

Leeds against Hull City in 2018 using rotations of players in a single narrow channel to progress forward

The risk-reward ratio is also pretty even, especially in transitions as you can have a 4v4 in attack than can soon turn into a 4v4 in defence. The absence of central players to provide defensive stability is the reason. It makes sense why Bielsa attested to the margins between the two sides being fine despite the result. The disconnect between attack and defence only gets bigger once the central players and wingbacks experience fatigue. In terms of periodization, playing a 3–3–1–3 in a competitive league every week is sure to result in burnout eventually.

Nevertheless for tactical inquisitiveness, the 3–3–1–3 is a beautiful system to explore for the understanding of superiorities and moments that lead to transition. There is very less room for error, but if every position dominates its zone individually, the outcome is fast-paced and exciting to watch. With every strength comes a weakness, and both appear quite evident in the 3–3–1–3.

Expansive football: A simplistic categorization

Expansive football has drawn a lot of attention in the last two decades for its aesthetic expression of the game, tactical prowess and technical superiority. The reason is simple. Teams that dominate possession and are able to successfully maintain an attacking phase, look dominant, create more chances to score, and win more often than teams that don’t. Although there are numerous ways in which teams achieve an expansive football structure in their attacking phase, I attempt to simplify them in this article by categorizing them as low expansive or high expansive.

Fig. 1 Categorization of expansive football models — low expansive versus high expansive

Most football formations depict how a team defends without the ball. However, when a team has possession of the ball, the players need to position themselves to make the field as big as possible. They need to exploit certain areas of the pitch that gives them more control over the opposition, such as the centre, or half spaces. The attacking team also needs to move relative to the structure presented by their opponents who are defending them. Here comes the importance of positioning advanced players between the lines and pinning back defenders with forwards. Finally, the buildup structure provides a platform to recycle possession and continue maintaining the attacking organization. This buildup structure depends on how the opposition presses from the front — with one, two or three strikers.

The basic elements of a positional formation remain constant — width, superiority, defensive cover. Thus we would expect any kind of expansive positional attack to have players positioned on the wings to stretch the width, between the lines to create superiority behind lines of pressure, and close support range that can immediately offer defensive cover if the ball is lost. So how does categorizing expansive structures into low or high expansive help us?

Although the static occupation of zones on the pitch look similar in both systems, the dynamic movement of the players affects opposition structures differently. At a glance, we can see that in high expansive structures, players move outward from the top and tuck inside from the bottom. In low expansive structures, players move outward from the bottom and tuck inside from the top. In most game scenarios, opposition markers will track down players in their zones. Thereby, the movement of players create spaces than can be exploited by teammates. As a consequence, different movement of players are bound to create different spaces to be exploited. Let’s look at their nuances in detail with each type of structure:

Low expansive model

This type of positional attack is typical to a lot of teams from Spain. The fullbacks strive to push higher up as wingers and stretch the width. This draws the centrebacks out wider, and the central pivot drops deeper as the third centreback. The wingers tuck inside and either pin back the last line of defenders or play between the lines.

Fig. 2 Low expansive model

A lot of room opens up in the lower half spaces that needs to be occupied by central midfielders. Quite often we see the fullback pass on their defensive responsibility to a central midfielder who covers for them as they push higher. We often see creative playmakers building up the play from deeper where there is more space.

A consequence of facing low expansive structures is that the opposition tends to get more compact at the back line. The overlapping fullback adds superiority on the wide lane and, along with the inverting winger, creates a 2v1 overload for the opposition fullback. To balance this overload, the opposition winger is forced to track back and collapses onto the backline if he gets pinned positionally.

Fig. 3 Overload on the wide lane in a low expansive model

The backline gets compressed and the midfielders or wingers get spread out trying to defend the overloads on the wings. A 4–5–1 easily turns into a 5–4–1. This creates space in the middle which is why a creative player can thrive from deep. Sevilla during the 2019 season were a perfect example of this system with Jesus Navas, the RB pushing high up as an attacking player and Ever Banega having a lot of freedom to create plays from deep.

As the fullbacks push higher up in the wide channels, the centrebacks are drawn out and forced to defend wide spaces. This requires atleast one central midfielder to balance the distances between the centrebacks either by splitting them, or playing alongside them. This central midfielder is usually the pivot. Bringing the central midfielder out of the centre and onto the backline tends to sacrifice his qualities as a midfielder to beat markers with his body orientation, break lines of press by receiving on the farther foot and creating superiority in the middle. Instead the team relies more on his distribution in low expansive formations. Pay close attention to Busquets as he adapts to different systems Barcelona uses against different opponents.

Fig. 4 Barcelona vs Getafe from 2019–20 season demonstrating a low expansive model

Low expansive structures also affect the movement of the striker and the spaces that open up for a striker to exploit. As the opposition defensive structure gets compact, there is less room for a striker to make behind the defensive line of the opposition. A striker finds other players pinning back defenders and is often an excessive addition against the backline. A striker who can play as a false nine, however, can find a lot of opportunities to drop deeper and receive the ball as there are more spaces in the centre with the dispersion of the opposition to the wings. We frequently see Messi playing this role in the current Barcelona system (2020–21).

High expansive model

This system is more recent than the former, famously reinvented by Pep Guardiola at Bayern Munich. The wingers tend to play a more traditional role of hugging the touchline. The central midfielders push up high and play between the lines, while the fullbacks invert and defend central spaces left by the midfielders.

Fig. 5 High expansive model

As the fullbacks invert and play more centrally, very often the centreback, in possession of the ball, looking to play around an opposition, finds the winger directly in the wide channel. It becomes a tempting prospect for the opposition fullback on the same side to track the winger dropping deep to receive the ball and apply pressure, because if the press is successful, the centreback could get easily isolated against a counter attack from the wide.

With no threat behind the fullback, he is free to neutralize the winger by closing down the space immediately. Thus, contrary to how low expansive structures tend to make the back line compact, high expansive structures tend to draw out the backline and disperse it.

Fig. 6 Spaces created between the defenders in a high expansive model

In such situations, an advanced midfielder looks to exploit the dispersed backline by making runs along the half spaces or inside channels. This makes the opposition vulnerable to through balls as the spaces open up at the back. How often have we seen these runs made by De Bruyne at Manchester City? If low expansive structures open up spaces in lower half spaces, high expansive structures create room in higher half spaces.

Another difference to low expansive structures is the room created for the striker. The striker has a lot more freedom to attack the spaces created by a shifting backline and can position himself anywhere along the line. However, for the central midfielders to successfully occupy spaces between the lines, he most often needs to pin back specific defenders in certain instances.

Fig. 7 Manchester City vs Schalke from the 2018–19 season demonstrating a high expansive model

The fullbacks need to be technically skilled to play in the centre as overloading the middle will tend to attract opposition midfielders to mark players. As the superiority is created in the middle, the spaces tend to open up on the wide areas, especially on the opposite flank. There needs to be a winger positioned on the opposite touchline ready for the switch as an overload is created on one side.

“Low expansive structures tend to make the back line compact, high expansive structures tend to draw out the backline and disperse it.”


This categorization is conceptual and based on logical interpretation of how spaces are created. Actual game scenarios tend to be a lot more complex. Team structures nowadays tend to be asymmetric even, having an overlapping fullback on one wing, and an inverted fullback on the other. The decision making involves the qualities and characteristics of players more than merely tactical reasoning.

Fig. 8 A summary of both models

If you notice, both these systems create a 3-at-the-back framework in two different ways. I deliberately chose the 4-man defence systems to explain these models, not to infer that 3-man defence systems are inflexible, but rather the oversimplified assumption that a 3–5–2 is inherently expansive with the dynamic role of wingbacks. Tactical intricacies to a 3-man-defence is not only possible based on concepts like superiorities and overloads, but commonly utilized by many teams like those of Tuchel, Conte and Nagelsmann.

Although conceptual, this categorization hopes to provide clues on opponent behavior based on which model we chose to attack with. The spaces that open up by virtue of a team’s formation is a consequence of small-sided games that play out in different zones of the pitch, a small example of which I demonstrated earlier in both models. The infinite possibilities and combinations are what makes the tactical framework of the game interesting.