!

Secondary and Primary Assist-locations

in the UEFA Champions League

An investigation of essential passes in football

Douglas Jakobsen

GYMNASTIK- OCH IDROTTSHÖGSKOLAN

T2A [Tränarprogrammet] 2017-2020
Dan Wiorek
Table of Contents

Introduction ................................................................................................................................1
1 Background..............................................................................................................................1
2 Research Question ...................................................................................................................2
3 Methodology............................................................................................................................2
4 Results .....................................................................................................................................4
4.1 Secondary Assist Locations ..............................................................................................4
4.2 Primary Assist Locations ..................................................................................................6
4.3 Secondary & Primary Assist Locations ............................................................................8
5 Discussion................................................................................................................................9
References................................................................................................................................. 11
Appendix
Introduction
This study is an examination component in the course Theory and Methodology of Training 2
(7.5 credits) of the Sport Coaching Program at The Swedish School of Sports & Health
Sciences (GIH) in Stockholm. The assignment is to produce a tactical analysis within the
students sport, which in this case will be an investigation to pinpoint the locations of the
secondary assist pass, and its correlation to the primary assist in Champions League football.

1 Background
Over the years, football has seen a constant tactical development on the pitch. The realm of
football specific statistics has also seen a radical evolution; the array of measurable statistics
has broadened extensively, with virtually every occurrence in a top level match being
documented. As Anderson and Sally outline in their book The Numbers Game, “it is numbers
that will challenge convention and invert norms (in football).” (2013, p.1)
The availability of in-depth statistics to top teams, for example those competing in the UEFA
Champions League, has made a more objective analysis of football’s seemingly archaic nature
possible. Today, objective analysis based upon these statistics are used by elite clubs to
identify collective team and individual player qualities. Objectively-based statistical values
have also been used as variables in various academic studies, to determine a players quality
by valuing player actions for example; the paper “Actions Speak Louder Than Goals: Valuing
Player Actions in Soccer”, introduces the “HATTRICS-OTB (Honest Attribution of Credit in
Soccer for On-the-Ball Actions) algorithm for valuing on-the-ball player actions” (Decroos,
Bransen, Van Haaren & Davis 2018) which uses data sets of documented player involvements
as variables with the goal of producing a more objective analysis of a players quality. This is
significant as it provides coaches and managers with an objective analytical option when
attempting to determine a players quality, as opposed to the very subjective individual
interpretations they would previously have relied on. Other studies have targeted team
performance as opposed to individual players, and used statistics to investigate the role of one
specific football action, for example passing, and its correlation to overall performance.
Already in 1997, Pollard and Reep examined passing behaviours in elite level football in
“Measuring the effectiveness of playing strategies at soccer.” However, as Rein, Dominik
Raabe, Memmert discuss in their 2017 work “Which pass is better? Novel approaches to

!1
assess passing effectiveness in elite soccer”, “The mere statement that the winning team has
made XY number of passes, whereas the losing team made YX passes does not help to attach
any specific relevance to individual game events.” In that sense, pinpointing the pitch-
locations of certain football actions may be a better approach; as Muhamad, Norasrudin, and
Rahmat did in their 2013 study “Differences in Goal Scoring and Passing Sequences between
Winning and Losing Team in UEFA-EURO Championship 2012”, by determining the
percentage of goals scored in and outside the penalty area. This investigation will look at two
specific types of passes in a specific competition; the secondary and primary assist passes in
the 2017/2018 UEFA Champions League. The UEFA Champions League is considered one of
the most prestigious club competitions in male professional football, with the top teams from
Europe’s top leagues competing against each other. The secondary assist is defined as the pass
that leads to the primary assist pass. The primary assist pass is defined as the pass that leads to
a scored goal. The respective secondary and primary assist locations will be analysed and
explored in search of existing correlations. Knowing the most prominent specific locations of
both secondary and primary assist passes, and the correlation there between, may be useful
information for a football coach in regard to both offensive and defensive tactical approaches.

2 Research Question
What were the pitch-locations of the secondary assist passes (the pass leading to the primary
assist) in the 2017/2018 men’s Champions League knockout stage? How do these secondary
assist pitch-locations correlate with the location of the primary assists?

3 Methodology
Scholarly studies within the field of football-related statistics and tactics were searched for
using the Discovery database, using both references to the game; “soccer” and “football”. The
data collection process, was as follows: The 2017/2018 Champions League knockout stage
featured a total of 29 matches. 3 matches ended in a 0-0 result, and were omitted. The 26
remaining matches were analysed using statistics and data sets from Opta Sports via
www.whoscored.com. A data set containing more than 45 000 documented occurrences from

!2
these 26 matches was extracted and downloaded into Excel 2016 using the Python 3.7,
Requests 2.19.1 and Beautiful Soul 4.6.3 softwares. The match occurrences were then
manually analysed to identify goals scored; a total of 91 goals were scored and were analysed
further. Goals that were scored from a set piece, for example from a free-kick, were omitted.
Goals that did not have any assist, for example if a player acquired the ball and scored
himself, were also omitted. After this process, a total of 64 goals scored with a primary assist
remained, 58 of which had both primary and secondary assists. The x and y location of the 64
primary assists, and the 58 secondary assist passes were extracted from the data set and
plotted onto a respective football pitch for visual representation, using the Matplotlib 2.2.3
software. Finally, both the primary and secondary assists were plotted onto the same football
pitch coordinated in their respective colours to allow for a clearer representation.
In order to specify the zones on the football pitch, the pitch was divided into 5 corridors of
play in the depth direction of play, a widely accepted and used method among elite coaches.
These corridors are defined as they are in the Swedish Football Association “UEFA A”
coaching diplomacy; starting from the outer right perimeter, the corridors are labelled right
outer, right inner, central, left inner and left outer. In the width direction, each penalty area
was extended to the byeline. At half the distance from this extended line and the halfway line,
another line was drawn, effectively making one half of a football pitch into three thirds. These
thirds were labelled, starting from the central area and moving towards each end, as follows;
primary third, middle third, and final third both for the defensive and offensive half. The
offensive penalty area was divided into 3 zones; the “right assist zone”, “left assist zone”, and
the “golden zone” in order to establish precise and unique terminology to discuss the
relationship between essential football actions in the offensive penalty area, where the
majority of goals are scored (Muhamad, Norasrudin, & Rahmat 2013). The pitch geography
and terminology is visually represented below in figure 1.

!3
Figure 1 illustrates the football pitch and its compartmentalised geography, which will be essential for determining
the precise locations of the secondary and primary assist locations. The inner corridors in the offensive penalty area
are labelled “assist-zone”, and the central corridor as “golden zone”.

4 Results

4.1 Secondary Assist Locations

The locations of the secondary assist passes indicate a fairly even distribution throughout the
offensive half of the pitch, with all areas except one featuring at least one secondary assist
pass. In reference to the corridors of play, there is also a fairly even distribution of pass
locations, with the central corridor showcasing a slightly higher value of 14 (24%). The left
outer and inner corridor both featured 11 (19%) secondary assists. The right inner corridor
had a slight advantage over the right outer corridor, with 12 (21%) and 10 (17%) respectively.
The results in reference to the corridors of play are represented below in table 1.

!4
Table 1 shows the number of secondary assist passes in each corridor of play.
Secondary Assist Locations: Corridors of Play

Left outer corridor 11

Left inner corridor 11
Central corridor 14
Right inner corridor 12
Right outer corridor 10
0 4 7 11 14

In reference to the half and offensive/defensive thirds, the defensive half of the pitch is
significantly less represented. In fact, if the data from the defensive primary third is omitted,
only 3 (5%) passes coming from the defensive middle and final third resulted in a secondary
assist pass. Furthermore, the offensive middle third is the most represented with 21 (35%),
followed by the offensive primary third with 16 (27%), and followed by the defensive
primary and offensive final thirds, each with 10 (17%) secondary assist passes. The results in
reference to the offensive/defensive thirds are represented below in table 2.

Table 2 shows the number of secondary assist passes in relation to the width of the pitch; the defensive and offensive
thirds of the pitch.

Second Assist Locations: Offensive/Defensive Thirds

30

23

21

15 16

10 10
8

1 2
0
DEF final third DEF middle third DEF primary third OFF primary third OFF middle third OFF final third

!5
 4.2 Primary Assist Locations
The primary assist locations are predominantly centred near the attacking goal, with a
substantial amount coming from within the penalty area. In reference to the corridors of play,
the central corridor, followed by the left inner and right inner corridors, accounted for the
most amount of primary assists, with 21 (33%), 14 (22%) and 13 (21%) respectively. The
outer corridors accounted for the least amount of primary assists, with 10 (16%) recorded in
the right outer corridor, and 5 (8%) in the left outer corridor. Below in table 3, the results in
relation to corridors of play, are displayed.

Table 3 shows the primary assist locations in reference to the corridors of play.

Primary Assist Locations: Corridors of Play

Left outer corridor 5

Left inner corridor 14

Central corridor 21

Right inner corridor 13

Right outer corridor 10

0 6 11 17 22

In reference to the offensive and defensive thirds, the offensive middle and final thirds
account for a significant number of primary assists, with 18 (28%) and 34 (54%) respectively.
Specifically looking at the offensive final third, the value of the “assist zones” as opposed to
the corresponding outer corridor area is established, with both the right and left assist zones
accounting for more primary assists than the right and left outer corridor in the offensive final
third. In the central corridor, the distribution of assists was less correspondent with proximity
to the goal than other corridors, especially the outer corridors, where a vast majority of the

!6
 recorded assists came from the offensive final third. Furthermore, and particularly the case
with the right inner corridor, a substantial amount of assists in the inner corridors came from
inside the penalty area, in the area known as the “assist zone”. The right assist zone,
accounted for 9 of the 13 assists in the entire right inner corridor. The left assist zone,
accounted for 6 of the 14 assists in the left inner corridor. The entire defensive half of the
pitch recorded very few primary assists, a total of 5 (8%). The offensive primary third
accounted for a few primary assists with 7 (11%), although significantly less than the
offensive middle and final thirds. The primary assist locations in relation to the offensive/
defensive thirds are represented below in table 4.

Table 4 shows the distribution of primary assist locations in reference to the width of the pitch, divided into offensive
and defensive half thirds.
Primary Assist Locations: Offensive/Defensive Thirds
40

34
30

20

18

10

7
1 2 2
0
DEF final third DEF middle third DEF primary third OFF primary third OFF middle third OFF final third

!7
 4.3 Secondary & Primary Assist Locations
Figure 2 below, provides a visual representation of the pitch-locations for both the secondary
and primary assist passes. The 58 secondary assist passes are plotted in black dots, whereas
the 64 primary assist passes are plotted in red dots. The individual representations showing
only the secondary and primary assist-locations can be found in the appendix.

Figure 2 shows the plotted secondary assists (in black), as well as the primary assists (in red). Figure 2 also shows
the “assist-V” which encompasses a substantial amount of secondary and primary assists passes, and is examined
further in the discussion.

!8
5 Discussion
The aim of this study was to provide an understanding of how goals are scored in elite male
football, by pinpointing the locations of both the secondary and primary assist pass. The
method used was effective in answering the research question, but due to the quantity of the
data set (over 45 000 occurrences) that were manually evaluated, it is prone to human error.
Knowledge of the most prominent assist locations, as well as recognising reemerging patterns
in assist locations, may help coaches adjust their tactics to exploit certain areas from an
offensive perspective, and protect certain areas from a defensive perspective.
The results indicate a few interesting trends in terms of secondary and primary assist
locations, and the correlation between assist locations and aspects such as proximity to the
goal, and the relevance of occupying the central and inner corridors. Furthermore, the value of
the “assist zones” and the inner corridors in the offensive middle third is quantified. In their
investigation that compared successful and unsuccessful football teams based on ball
possession, Dixon and Protheroe suggested that further research “using a different grid
template focusing on narrower wide areas and a specific area designated to the penalty box
can also produce more accurate data with regard to pass location” (Dixon & Protheroe 2015),
which this work to some extent accomplished. Thanks to the specific grid template made up
of the 5 corridors of play and the 6 “half thirds”, we are able to get a detailed representation
based on objective values, of what areas are more attractive and useful than others in male
elite football. Although, a problem with the grid template used in this study is that it creates
zones that vary in area, which may obscure the results and their interpretation. However, the
corridors used are fairly normalised within the domain of football coaches, which is important
as it can allow those familiar with this method of dividing the pitch to better interpret the
results. From an offensive standpoint, there seems to be a correlation between the right inner
corridor in the offensive middle third, and the right assist zone. A significant number of
secondary assists from the right inner corridor (particularly in the offensive middle third) and
a significant amount of primary assists from the right assist zone suggests that playing the ball
from the inner corridor into the assist zone, and thereafter playing to into the “golden zone” is
an effective tactic. Additionally, the overrepresentation of primary assists in comparison to
secondary assists in the offensive middle third, central corridor area suggests that accessing
this space could also be an effective tactic. As well as providing coaches with information on
how what statistically are good areas to attack or defend collectively, the results of this study
!9
can also be used to show individual players in certain positions the importance of specific
areas on the pitch, which may help them improve their performance. For example, an
offensive midfielder may get a quantified picture of the opportunities available by accessing
the inner corridor in the offensive middle third, and a right wide midfielder may be
encouraged to access the right assist-zone rather than crossing from the outer corridor.
As seen above in Figure 2, when generalising the plotted data of secondary and primary assist
passes, a “V” shape can be applied to the offensive half of the pitch, which will encompass a
substantial amount of passes. This “V of assists”, or simply the “assists-V” can be regarded as
a major finding of this study, as it exhibits a new way of dividing the football pitch to aid
strategic understanding and discussions, and can be a foundation for tactical systems. The
assists-V is a product of similarities in the styles of play used by some of the worlds best
teams in the Champions League, but it can act as a substructure for coaches of all ages groups
and levels when designing their own footballing strategy and tactical schemes; how can a
team collectively access the assists-V with possession of the ball, what offensive formation
should players position in when possession is established in the offensive half, what player
qualities are most effective in reference to position within the assists-V, and how should
players act collectively to minimise time and space of opponents when defending their own
assists-v area, are examples of questions which answers will be instrumental when
constituting a coach's strategy.
This study has also laid a foundation for further research, which can provide great insight into
goal-scoring sequences with the help of more in-depth analysis such as goal-locations, the
number of touches a player has before playing the secondary and primary assist pass and
linking each secondary and primary assist to the goal-location of the specific sequence.
Additionally, taking factors such as a players tactical position and their dominant foot into
consideration when plotting secondary and primary assists can help coaches foster a more
objective understanding of how to use certain player qualities from a tactical perspective.
Football has long been prone to simple truths, where certain concepts in play have become
cemented norms and are regarded as the only best or only way to do things. New ways of
approaching the game are essential to its continued development along all dimensions of
analysis; tactically, technically, subjectively, objectively and so on.

!10
References

Anderson, C., Sally, D. (2013). The Numbers Game. Penguin Group.

Decroos, T., Bransen, L., Van Haaren, J., & Davies, J. (2018). Actions Speak Louder Than
Goals: Valuing Player Actions in Soccer.

Pollard, R., & Reep, C. (1997). Measuring the Effectiveness of Playing Strategies at
Soccer. The Statistician. Vol 40, 541-550.

Rein, R., Raabe, D., & Memmert, D. (2017). Which pass is better? Novel approaches to
assess passing effectiveness in elite soccer. Human Movement Science. Vol 55, 172-181.

Muhamad, S., Norasrudin, S., & Rahmat, A. (2013). Differences in Goal Scoring and Passing
Sequences between Winning and Losing Team in UEFA-EURO Championship
2012. International Journal of Social, Behavioral, Educational, Economic, Business and
Industrial Engineering. Vol 7, 332-337.

Dixon, L., & Protheroe, L. (2015). A comparison of chances created between successful and
unsuccessful Premier League teams based on possession. Hartpury Student Research Journal.
Issue 1. Retrieved from https://studentjournal.hartpury.ac.uk/issue1/project-articles/
chancesbasedonpossession/.

!11
Appendix

Figure 3 shows the plotted secondary assists.

Figure 4 shows the plotted primary assists.

!1
Afterword.

Douglas Jakobsen, known online under the pseudonym “Coachdogge” is a 23 year old
football player and football coach from Sweden. Currently, along with the practical education
of playing in the third tier of the Swedish football pyramid, formal football education is
provided through studies in the Coaching Program at The Swedish School of Sport & Health
Sciences in Stockholm. Additional formal education has come through previous university
studies and the UEFA A and Goalkeeper A licensing/coaching certification. Douglas is in a
unique position as one of very few who completed the UEFA A license before his 20th
birthday. In terms of practical coaching education, Douglas began coaching officially at 15,
within multiple domains including the high school football team, goalkeeper training,
individual football training and physical training. With 2 parents both heavily invested in
sport at the international level both as active athletes and coaches, the art of coaching and the
charisma of sports was an early and pivotal part of childhood. Additionally, active
investments into “informal” education as a coach have been consciously made for many
years, including “self-teaching” within several football/leadership/coaching related concepts.
Furthermore, much knowledge has been acquired through study-trips across the globe, and
the gleaning of many highly competent individuals through mentor-protege relationships. The
usefulness of these types of relationships cannot be overstated.

This study serves as a new type of approach (a more academic and objective one) to studying
football. I believe that the results of this study have added a new dimension in my football
understanding, and football philosophy. It is my hope that publishing this work will help
spark discussion and foster ideas with other coaches and football thinkers! Thank you for
reading and for your tremendous support, please do not hesitate to contact me with questions
and feedback via Twitter (@coachdogge) or via mail (douglas.jakobsen@student.gih.se).

Special thanks to Luke Redman of Redmanalytics (twitter @redmanalytics) for his

instrumental help in the production of this work! An extremely competent individual who’s
work I strongly recommend!

Yours in football,

Douglas

!2