How does Salah create space to score easier goals ?

How do the best strikers score goals ? If you watch a highlight reel of premier league golden boot winners (player having scored the most goals in the league each season), you would notice that most of the goals are not spectacular. You would see a lot of tap-ins, one on ones or scruffy pee-rollers. There are a couple of reasons for this. Traditional "beautiful" goals (long range pile-drivers or mazzy solo runs) are hard to score and they come from positions where players are less likely to score. Hence, the best scorers score so many goals because they find themselves in the position to convert easy opportunities. This is where the difficulty of their trade lies. No matter how lethal a player is, if the ball does not come to him, he will not score. In many cases, strikers make goal-scoring opportunities by creating space for the ball to come into. This is what we will be looking at today. We will discuss a way to evaluate the space players create through their movement in order to analyse goals scored by Liverpool in recent seasons.

It may look easy to create space. However, the defender's main job is to close down the space, get close to you, make sure you find yourself in an uncomfortable position when receiving the ball. The best strikers work tirelessly all game, making run after run and often receive very little service. Only a few of their runs will be rewarded by a pass and even then, the pass may not be completely accurate. This is why it is important to find a method to incorporate a player's movement in an analysis of performance. Sometimes a player's run is essential to the build up of a goal, yet may go unnoticed. Take a look at Neymar's goal against Real Madrid back in 2014 (you may have to open it up on youtube directly):

It may not be obvious at first glance but an essential part in the build up to that goal is Messi's run that drags Modric out of position, which in turn allows Neymar to get into space at the edge of the penalty box and score. Without this run, Neymar would have run into Modric and perhaps not been able to take the shot on. Yet statistically, Messi will not be awarded any credit.

We will look at a model that evaluates how much of the pitch a team controls. By control, we mean if the ball goes to a given a location on the pitch, how likely it is that either team takes possession of the ball. Through this model, we can analyse when players create space by looking at the control of the pitch that arrises around players from runs or movement. There are many variables that affect which team might take over possession. If you pick out a specific spot on the pitch and want to calculate the pitch control, you need to consider how long it might take for the ball to arrive there if the player in possession passes the ball to that spot. You need to consider who are the players closest to that spot and how long will they also take to run there given their current speed and direction they are running in relative to the considered spot. From this and other parameters, the model outputs a value representing how likely it is that either team will control the ball. This is a broad overview of the functioning of the model, which in itself is relatively involved. If you are interested in the detail, a version of the model was developed in this paper by William Spearman although there are many other papers out there on the topic - https://www.researchgate.net/publication/327139841_Beyond_Expected_Goals. There is also a video of him explaining how it works, which may be more accessible - https://www.youtube.com/watch?v=X9PrwPyolyU&feature=emb_logo.

For us to be able to apply this model in practise, we need tracking data. This is data which tracks the movement of the ball as well as all the players on the pitch at a certain frequency. Usually the frequency is 25 or 20 recordings per second. This type of data is not widely available online. However, the Friends of Tracking initiative have made available tracking data on 19 recent Liverpool goals available. This is what we use.

Let us first consider a common scenario: a player makes a run towards goal to create space between the defensive line and the goalkeeper. The goal in question is Liverpool's second goal in the return leg of the Champion's League quarter final against Porto in 2019. You can watch it by jumping to 3:30 in the video below.

In this situation, you have Alexander-Arnold leading a counter-attack with both Salah and Mané making runs ahead of him, offering passing solutions. If you look closely, Salah makes a first burst of speed from behind the defender who is tracking back and does not immediately see him. The defender eventually notices and responds to the acceleration. Salah then slows down for a fraction of a second giving the impression that he is stopping his run before accelerating again. This time he goes slightly to the left and distances himself from the defender who hovers to the right in the hope of intercepting the pass. However, Salah having picked up some speed allows Alexander-Arnold the option to pass the ball deeper as to avoid the defender cutting it out. This is what he does and Salah finishes clinically. It looks like a fairly simple goal at first glance but there is actually a lot going on. Salah makes two separate runs, the dummy first one, then indicates to his teammate where he wants the ball (this is subtle but at 3:32 of the video you can see him pointing to the area ahead of him) followed by the decisive acceleration.

So this is what we can gather by looking at the video. Let's see what our model can tell us in terms of space control. In the first instance below, Alexander-Arnold (number 66) is running with the ball. The red sections are the areas controlled by Liverpool, the blue section are the areas controlled by Porto and the white areas are more uncertain. Here, Porto have a good control of the area in behind their defence. If Alexander-Arnold were to make his pass now, it would most likely be unsuccessful. He recognises this and decides to continue dribbling forward with the ball.

Fast-forward one second (image below), Salah (number 11) makes his first burst. You can see that he has opened up space ahead of him and to the left. For him to get a good goalscoring opportunity from a pass at this instant would be difficult. The space is not quite central enough, so the pass would have to be accurately placed to the top part of the zone ahead of him. This would make the pass hard to execute, especially given that the angle is not in Alexander-Arnold's favour without playing a high ball. Again, Alexander-Arnold waits for the right moment.

This moment eventually comes. The frame when he makes the pass is the one below. The space Salah has created with his second acceleration is a lot more central and prone to a good goal-scoring opportunity. The angle for Alexander-Arnold is also much better for him to make the pass. These factors lead to the pass being successful and then the goal being scored. More importantly than being successful, the pass is dangerous. It arrives in a dangerous area. That is because Salah managed to peel off to distance himself from the defender and accelerate just at the right time in order to create this area for the ball to come into that is well positioned relative to the opposition's goal.

Salah's goal against Porto was a scenario where a player made a run, created space for himself, received the ball and scored. However, as we have seen, there are often situations where the player making the run opens up space for one of his team-mates to receive the ball. We will briefly look at a situation of this type through Firmino's goal against Leicester in December of 2019. You can find it below at time 2:03.

This is somewhat subtle but as Alexander-Arnold readies his pass, both Firmino and Mane are coming into the same area of the box. Mane then makes a burst towards the near post, dragging the defender with him allowing Firmino plenty of time to control the ball, compose himself and score. As before, let's take a look at what our model says in terms of pitch control.

As Alexander-Arnold is about to pass, the area ball-side of Firmino (number 9) is mostly white or blue. Hence, even if the ball does get to him, he will quickly be put under pressure. There is however some space to his left, which can be exploited.

As the pass is coming towards Firmino, you can see by the arrows that Mane (number 10) is going towards the near post and dragging the defender with him. At this point already, the space behind the defender is a lot more free.

When the ball finally reaches Firmino, the defender is not close to him and momentum is pushing him the other way, leaving him with plenty of time to place a composed finish in the opposite corner.

Scenarios such as this one with players creating space for team-mates are harder to analyse than Salah's simple run in behind. There are more elements and factors involved. It also makes it difficult for defenders, they have to make a choice. Get close to one player but leave another open. This is why collective movement as team can be very effective.

Our model has done its best to encapsulate the creation of space induced by the movement of players on the pitch but as always, it can be better. Improvements can be made on several fronts. It assumes that all the players have the same acceleration and maximum speed and that the ball always travels at a constant speed. It does not take into account fatigue or technical qualities of a player to get the ball under control. These are all hard to quantify but are interesting ways amongst many others to develop this model further. Nevertheless, this model is great at conceptually introducing the topic of pitch control, which adds a whole new dimension to football analytics.