In-Depth Statistical Analysis Of LA Galaxy’s Goalkeepers In The 2016 Season

18th Apr 2017 by TSZ

In the next part of our series analysing LA Galaxy in the 2016 MLS campaign, we look at another aspect of our bespoke data collection to study their performance.

In February, we published two articles that explored the data around LA Galaxy’s chance creation and their designated players. Both pieces helped us understand the team’s performance during the 2016 MLS regular season in 2016.

At the end of the designated player piece we briefly touched upon our next area of study and the topic of this piece - goalkeepers. To create an easily comprehensible structure to this analysis, we have decided to divide it into two separate articles. Below you will find the first section, which focuses on LA Galaxy’s goalkeepers, and in a future piece we will explore their opponents’ goalkeepers.

In recent years, the role of goalkeepers has changed significantly. Whilst their main job remains as a shot-stopper, managers such as Pep Guardiola have revolutionised their positioning. In his style of play, he requires the goalkeeper to have excellent technical footwork, participate more in the game and work as a sweeper who is able to set up play from the back.

Different managers have different ideologies around how to best manipulate the goalkeeper, so where are LA Galaxy on this scale? Are they looking for certain types of distribution? Are they successful in their distribution? And do the answers to these questions perhaps differ depending on the current score in the game? If managers are clear about their ideology and how they want to play, in theory they should be able to implement preferred tactics right from the back. This all leads us to the ultimate question; was there a trend in how LA Galaxy conceded shots and goals during the 2016 MLS regular season? If this can be understood, it provides huge benefit for managers and goalkeeping coaches in how to set up defensively to avoid awkward positioning. 

Before we get into the numbers it is important to note a few restrictions in the analysis. Firstly, we have only looked at LA Galaxy’s regular season (2016) games, so the data does not include the three play-off matches they participated in (or the start of the 2017 season). Secondly, when we refer to attempted distribution throughout the study, it signifies when the goalkeeper has had a clear option in making a decision around where to pass the ball, i.e. last minute clearances are not included in this. A successful distribution consists of a throw, kick or a header from the goalkeeper that reached a player in the own team. Thirdly, when we refer to distribution type we have omitted all goal kicks, as the only way to kick these are from the ground, which would inflate this measure if it were taken into account. Lastly, in the graphs below, whilst all three goalkeepers who appeared for LA Galaxy in the 2016 regular season have been grouped together (as their measures did not differ too much between each other), it should be stressed that any graph below can be split into player specific preference if that would be of interest.

This report is divided into five sections. The first sections look at total tables for LA Galaxy’s three goalkeepers to give an overview of performance. The second part consists of three stacked bar graphs showing distribution success rate for LA Galaxy’s goalkeepers depending on the level of pressure they are under and the current score in the game (winning/drawing/losing). The pressure is divided into four categories where high pressure suggests an opposing player was in a position to charge the kick down, medium pressure means the opposition forced the goalkeeper to play the ball early, low pressure would be where the players were not able to block the kick but still decided to put pressure on the goalkeeper, and lastly when the goalkeeper was not put under any pressure at all. This will enable managers to identify at what parts of the game and under what level of pressure specific goalkeepers seem to perform/struggle. Combined with the aforementioned follow-up section to this analysis looking at their opponents’ goalkeepers, managers will be able to understand how (and why) they should distribute the ball to fit the team and maintain possession, which ultimately should be seen as the goal. 

The third section of the analysis is made up of three bar graphs showing the distribution target depending on the current score in the game. Within each graph, there is a split to see the number of successful and non-successful distributions and where they happened. The fourth section builds on this and looks at the distribution type, again visualised through stacked bar graphs depending on the score. Comparing the second to fourth part of the analysis enables managers to understand the overall distribution from goalkeepers that LA Galaxy aim for at certain parts of the game. For example, if the aim is to keep possession, they will be able to identify that they should use certain distributions more than others.

In the final part of the analysis, we look at shots on goal and goals conceded, taking into account where the shots took place in relation to the goal. This is visualised through a heat map layered over a goal to see if there is a propensity for LA Galaxy to concede shots and goals in certain areas of the goal. This could help the manager and goalkeeping coaches to understand the positioning of their goalkeepers. It would also help them in the positioning of their defence to give extra support to areas that show concession of the majority of goals. With this in mind, let’s get into the analysis.  

In the table below, we can see that Brian Rowe was LA Galaxy's chief goalkeeper during the 2016 MLS regular season:


Rowe made 31 appearances and played a total of 2,795 minutes, outperforming the other two goalkeepers on all measures aside from goals conceded per game. His score for this measure sits at 1.1, which is 0.1 more than Dan Kennedy at 1.0. However, as the other two goalkeepers played so few games it is hard to get comparable measures. This is another reason as to why we have decided to group all goalkeepers for the rest of this analysis.

The graph below shows LA Galaxy’s success rate for distribution while winning:


A notable observation from this graph is that the success rate improved when more pressure was applied on the goalkeeper, against one’s natural expectation. However, a closer look into each of the graphs shows that there were only a small number of distributions where the pressure was low or more, suggesting that it is hard to draw strong conclusions from these observations. Overall, the grand total measure to the far right displays that roughly half of the distributions on average were successful. 

Now let’s move on to the success rate while LA Galaxy were drawing the game:


This graph is a little more aligned with what one would expect. The success rate gets worse as more pressure was put on the goalkeeper, again with exception of high pressure (with a total of 11 observations). The grand total of 68.5% success rate is significantly higher than the previous graph (53.1%) when LA Galaxy were winning the game. One possible explanation for this would be that perhaps LA Galaxy had the option to play the ball short to defenders when drawing, as the opponents were not chasing the game and hence play with a lower defensive line. This is something that the next few graphs investigate further. 

To complete this section of the analysis, we’ll now look at the distribution success rate when LA Galaxy were losing:


The graph above does not show a strong correlation between distribution success and pressure put on the goalkeeper. However, the grand total succession rate at 71.2% is the highest when compared to the other game states. With the previous two graphs in mind, this is potentially down to the pressure put on the goalkeeper, which supposedly is less from an opposing side when LA Galaxy were losing. 

Our next focus is the distribution target; essentially, who the goalkeeper chose to pass the ball to. Again, we’ll start with when Galaxy were leading a match:


The distribution target when LA Galaxy were winning is heavily skewed towards strikers, with 40.8% of the total number of distributions heading to the forward line. Wingers were the second most common distribution target, which suggests that when LA Galaxy were winning, they preferred to play the ball long rather than roll it out to one of their defenders. Before we move on, it is worth noting that the record for passing the ball to centre and full backs is perfect, with all aimed passes reaching their target.   

Looking at the game state when LA Galaxy were drawing, the bars have moved around a little:


In this game state, passing to centre-backs was the most common distribution target with a total of 40.2% of all distributions going this way (of which only three out of 264 occasions were unsuccessful).

So how does this compare when we look at the times when LA Galaxy were losing? 


This graph illustrates a very similar distribution, in that passing to centre-backs was by far the most common way of distributing the ball at 34.6% (this time with a perfect success rate). Therefore, distribution success is correlated to the score-line in the game. When LA Galaxy were drawing or losing, they were given more opportunities to pass the ball short to their centre-backs. This comes as a result of a lower sitting defensive line from the opponents, enabling this type of distribution to a higher degree. 

So, how does the type of distribution differ between the varying game states? 


We can see that a kick from the ground was the preferred way of distributing the ball regardless of the score-line in the game. However, the preference to kick the ball from the ground seems to be slightly higher when drawing or losing the game. The goalkeepers did not often kick the balls from their hands, and when doing so we can see that the red bar is significantly higher when winning or drawing, suggesting that this was an ineffective distribution type if the aim was to remain in possession. Clearly, managers may want to discourage this type of distribution if they prefer a possession-based game. 

Let’s now look at shot distribution:


More than half of the shots (56.1%) were taken in the lower parts of the goal, and the middle of the goal follows as the second most common part to shoot with a total of 23 shots or 13.9% of the total. When looking at the top of the goal we can see that the distribution is quite even across all three parts of the goal, but what does this look like if we move on and look to where the goals were conceded? 

The graphic shows that the vast majority of goals were conceded in the lower corners, with the bottom left yielding the highest rate at 23.1%. If we compare this to the previous graph, it makes sense that most goals were scored in this part of the goal as most shots were aimed at this area. Another observation is that 12.8% and 7.7% of all goals were scored in either of the top corners. This is actually higher than the proportion of shots taken in these corners. It is generally more difficult to aim a shot in either of the top corners, but it is also more difficult to save the successful attempts. This naturally translates into a better conversion rate for shots places in the top corners. 

The study showed that, on average, the success rate for goalkeepers distributing the ball to their teammates is higher when drawing or losing the game. When looking at the distribution target, we could see that when LA Galaxy were winning, they were forced to hit the ball long (pass to strikers) as a result of the opposition being likely to play a higher defensive line. The distribution type showed that LA Galaxy mostly kicked the ball from ground even though the accuracy when throwing was significantly better and close to perfect. In the final part of the analysis, we could see that most goals were scored in the lower corners of the goal, even though the conversation rate was highest in the top corners. This should be used in conjunction with video analysis to see whether there is a trend in terms of the positions where goals have been scored. So now that we know how goalkeepers best retain possession and where most goals were being scored, how can LA Galaxy improve their pressure on opposing goalkeepers and improve their chances to win games? To get the answer, keep your eyes open for the next article in this series!

The Stats Zone provides bespoke services for some of the leading teams, associations and confederations in world football. This analysis on the LA Galaxy was part of an internal study and not used in a commercial sense. Please do not hesitate to get in touch with us through our ‘Contact’ page - http://www.thestatszone.com/contact if you would like to discuss any bespoke research projects for your organisation.

Leading the way in the fast-moving world of sports analytics, The Stats Zone provides data services to brands, teams, associations and federations in sport.
Newsletter Signup

Scribble your email address here and click subscribe to receive our newsletter on all things sport, data and analysis!

Sports Data & Analysis Jobs
Related Content