Game Theory Report 2 1

Player Acquisition Analysis: How do European clubs buy soccer players?

Parthiv Mathur, Arjun Narendra

Honors 221C: Game Theory

Instructor: Dr Jake Cooper

5 March 2023
Game Theory Report 2 2

Introduction:
Imagine, for a moment, that you are in charge of player acquisitions for a European
football club, and a dilemma has just arisen. Your club has sanctioned 50 million euros in
spending for a new striker, and you have whittled the pool of potential players down to two
key contenders - Player A and Player B. Player A is an experienced forward at the peak of
his powers, and will give you an almost guaranteed 15 goals a season for the next few years.
Player A’s stellar reputation and long success streak give him a market valuation of 60
million euros. Other clubs are unwilling to pay such a lofty sum of money for just one player,
so you know that you can play it safe and snag him for 50 million euros. Player B is an
up-and-coming forward who has just recently made his professional debut. Without much
name-recognition and trophies under his belt, the market values him at 20 million euros.
Although he cannot assure a return of 15 (or even any) goals a season yet, he has huge
potential, and could become a far greater player than Player A some years down the line.
Other clubs also see the burgeoning talent of Player B, so you know that you need to take
the risk and pay the 50 million euros for him. What do you do?

This difficult decision-making process is one that many soccer clubs have faced in
the past and will certainly face in the future. There are some notorious examples
throughout history where clubs certainly did not get what they paid for. In 2018, Alexis
Sanchez was signed from Arsenal in a bargain deal worth 34 million euros. Although 34
million euros is objectively a large sum of money to pay any player, Sanchez was a
well-known forward with years of experience under his belt, so the market valued him at
an astronomical 70 million euros. Unfortunately, Sanchez was only able to net three goals in
the one-and-a-half seasons he spent at Manchester United due to an unfortunate series of
injuries. Here is a player who, by all objective measures, was expected to make a significant
positive impact for United but fell short. While United’s cautious strategy may not have paid
dividends, taking a risk on an inexperienced player does not guarantee higher rewards by
any means. Take the Spanish striker Bojan Krkić. In 2013, Barcelona paid 13 million euros
to acquire the 23-year old Krkić. Clearly, Barcelona saw a lot of potential in Krkić, especially
considering the fact that the market valued him at 8 million euros at the time. In what
seemed like a hit-or-miss scenario, this was surely a miss: Krkić ended up scoring only 4
goals in his 2 seasons. Clearly, there is no wrong or right approach when it comes to signing
these soccer players, so how do clubs decide where to draw the line?

From a purely financial perspective, clubs need to spend wisely. Although it is

unintuitive to think of it this way, each player acts as a sort of investment. If a player is
bought for 5 million euros and sold for 10 million euros, then the club makes a profit of 5
million euros from this player. On the other hand, if a player is bought for 10 million euros
and sold for 8 million euros, then the club loses 2 million euros. Therefore, if and when a
club decides to make a bet on a player, they must factor in the long-term repercussions of
their investment.

These days, name recognition also means a lot for players, fans, and clubs alike. If a
club has a good reputation, good players will tend to gravitate towards this club, and the
club will rack up more fans. The cycle continues. A risky investment that goes wrong not
Game Theory Report 2 3

only sinks the quality of the team, but at the end of the day it stains their reputation. On the
flip side, if a club’s risky investment ends up reaping rewards and defying the pundits’ odds,
the club could gain a much-needed recognition and success boost.

Game Concepts:
Central to the Player A Player B dilemma is the concept of risk management, defined
as “the practice of systematically selecting cost effective approaches for minimizing the
effect of threat realization” (Farnaz). Simply put, risk management loosely defines the
strategies actors use to maximize their utility. Risk management appears in games where
an actor or set of actors must choose between multiple options, where at least one option
has a value which is unknown. In regards to our proposed dilemma, Player A is a relatively
low-risk buy, since he is worth more than you would pay for him. Even if he performs
somewhat under his 60 million expectation, you have 10 million euros worth of protection.
Player B on the other hand would be a very risky buy. You need them to hit the ground
running and add value to your team equivalent to 30 million euros just to break even.
Though low-risk buys may always seem like the best option, this isn’t always the case.
Imagine you unearth a player you believe to be a future superstar, worth hundreds of
millions of euros. Then, in the short term, it is certainly worth overpaying for them to reap
the long term benefits.

More generally, clubs must align their risk-adversity towards acquiring players with
their projected total value added to the club. How does this translate into actions? Namely,
should clubs take large risks frequently? Should they be risk-averse? Somewhere in the
middle? Understanding the ideal level of acquisition risk allows clubs to generate maximum
value from the players they acquire, which will help propel them to high levels of
performance on the field.

One key aspect of risk management is perceived risk, which refers to the level of risk
an actor perceives a given choice to be. In order for clubs to develop an optimal
risk-management strategy, they must be cognizant of how their perceived risk compares to
the true risk of acquiring certain players.

Description:
In this report, we will be analyzing 5 player acquisitions from 3 different European
soccer clubs- in particular, German club Borussia Dortmund (Dortmund) English club
Manchester United (United), and English club Brighton & Hove Albion F.C. (Brighton). For
simplicity’s sake, the chosen players all play in a forward position1. We will classify each
acquisition as a “risky” or “non-risky buy”, and evaluate the amount of risk each club took
across their 5 acquisitions. Additionally, we will analyze the total value each player
provided for their club, and determine if there is indeed any correlation between the
proportion of risky acquisitions and their success. Key terms, such as “risky” and
“non-risky”, along with key metrics, like player value, will be discussed in further detail
within their respective contexts.

1
In this paper, forwards are defined as wingers (either left or right) and strikers. Attacking midfielders are
not included as forwards in this paper.
Game Theory Report 2 4

Analysis & Discussion:

To begin, we will analyze each club’s perceived risk in regards to their player
acquisitions. Before going further, it is important to mention how we are going to quantify
the level of risk-aversion a club has. Each player that a club buys has a perceived risk value
(how much money the club spends on the player) versus the true risk value (how much
money the player is actually worth). In order to calculate the true risk value, we will use a
popular metric known as market value. Market value is a running statistic about a player
which estimates their value in monetary terms. In other words, if a club wanted to buy a
player, most likely they would need to meet that player’s market value to have a real chance
of getting an agreement done. According to transfermarkt.us, the source which provided us
with all of the data shown below, market value is defined as “[a prediction of] the the
expected value of a player in a free market”, and incorporates factors such as “future
prospects, age, performance at club and national level”, and much more. If a club spends
more than the market value on a player, they are classified as “risky”, and if they spend less,
they are “non-risky”. The magnitude of difference between purchase cost and market value
informally denotes how risky a transfer is.

The following three tables contain the risk information for each of the five players
selected from the three chosen clubs. For each player, we keep track of their true cost at the
time of purchase (Market Value), how much the club paid for them (Cost of Acquisition),
and the amount of money the club overpaid for the player based on their true value
(Overpay). Additionally, we keep track of whether each player is risky or not (Risk
Category). Note that units of each value in the tables regarding money are in millions of
euros.

Borussia Dortmund:
Player Market Value Cost of Overpay Risk Category
(millions of €) Acquisition (millions of €)
(millions of €)

Jadon Sancho 5 20.59 15.59 Risky

Pierre-Emerick 15 13 -2 Non-Risky
Aubameyang

Ousmane 12 15 3 Risky
Dembele

Alexander Isak 8.60 9 .4 Risky

Robert 4.5 4.75 .25 Risky

Lewandowski
Game Theory Report 2 5

Brighton:
Player Market Value Cost of Overpay Risk Category
(millions of €) Acquisition (millions of €)
(millions of €)

Kaoru Mitoma 1.8 3 1.2 Risky

Deniz Undav 5 7 2 Risky

Leandro 16 15.56 -.44 Non-Risky

Trossard

Neal Maupay 12 15.56 3.56 Risky

Facundo 9 6 -3 Non-Risky
Buonanotte

Manchester United:
Player Market Value Cost of Overpay Risk Category
(millions of €) Acquisition (millions of €)
(millions of €)

Alexis Sanchez 70 34 -36 Non-Risky

Robin Van 47.5 30.7 -16.8 Non-Risky

Persie

Anthony 8 60 52 Risky
Martial

Cristiano 45 17 -28 Non-Risky

Ronaldo

Zlatan 15 02 -15 Non-Risky

Ibrahimovic

Below are four graphs synthesized from the data collected above.The first three
graphs portray the relationship between market value and cost of player acquisition for all
three clubs respectively, and the fourth graph compares the percentage overpay for each
player across all clubs.

2
Ibrahimovic joined Manchester United on a free transfer, meaning he didn’t renew his contract at his
previous club, PSG, and joined Manchester United as a free agent. Since United didn’t have to pay PSG
anything to secure his services, Ibrahimovic’s cost is 0.
Game Theory Report 2 6

Borussia Dortmund:

Figure 1. Depicts the relationship between Market Value and Cost for Dortmund
acquisitions. Compares this relationship to a direct mapping of market value = cost of
acquisition.
Game Theory Report 2 7

Brighton:

Figure 2. Depicts the relationship between Market Value and Cost for Brighton acquisitions.
Compares this relationship to a direct mapping of market value = cost of acquisition.
Game Theory Report 2 8

Manchester United:

Figure 3. Depicts the relationship between Market Value and Cost for Manchester United
acquisitions. Compares this relationship to a direct mapping of market value = cost of
acquisition.
Game Theory Report 2 9

All teams:

Figure 4. Depicts relative overpay for each player, which is calculated for each player as
𝑐𝑜𝑠𝑡/𝑚𝑎𝑟𝑘𝑒𝑡 𝑣𝑎𝑙𝑢𝑒. All Dortmund, Brighton, and Manchester United players are
color-coded Yellow, Blue, and Red respectively, in alignment with each clubs’ colors.

There are a number of interesting trends depicted in each of the first three graphs. A
quick note on the orange graph lines - they are all standard 𝑦 = 𝑥 lines. If a player’s
acquisition point lies on that line, then they were purchased for exactly their market value.
The presence of this line (hopefully) makes comparing the cost of acquisition to market
value for a player much easier.

In regards to Dortmund, they pretty consistently pay around (or slightly higher
than) market value for players - with a notable exception for Jadon Sancho. They paid
roughly 4 times over his market value for him, in the hopes that he would provide massive
returns on their investment. If we are including Sancho overall, then Dortmund would be
classified as a risky club. They overpay for their players a substantial amount, hoping for
long-term benefits. Without Sancho, it is fair to say that they are perhaps less risky, but still
take a fair few risks in the market. As the market value of a player increases, Dortmund’s
cost of acquisition tends to fall closer to / below the orange line, indicating that they were
less interested in taking risks on players who cost more money.

Brighton, like Dortmund, shows signs of becoming more risk-averse as the market
value of a player increases. As market value tips over 14 million, Brighton stop their trend
of linearly increasing their acquisition cost and begin to pay under market value for players.
However, since they consistently pay above the market value for players, their
risk-management strategy seems to involve taking more risk on players with lower
acquisition fees. Note that Facundo Buonanotte is a slight exception, since Brighton paid
Game Theory Report 2 10

(perhaps surprisingly) less than his market value for him. Interestingly, Brighton are the
only club without any significant outliers.

Manchester United certainly has the most interesting data out of the three involved
clubs. For one, they shop consistently in higher price ranges than Dortmund and Brighton -
not surprising given the size and stature of the club. What is surprising, or at the very least
somewhat counterintuitive, is that they wildly underpay for most of their players. With the
notable exception of Anthony Martial, whom they paid almost 8(!) times his market value
for, every player was purchased for under their market value. Overall, this would suggest
that United are operating as quite a risk-averse club - the most risk-averse of the three
clubs being analyzed. There is no clear trend for how their risk-averse behavior changes as
a player’s market value changes, unlike the other two clubs.

Figure 4 is an amalgamation of the data from the previous three figures, and it
depicts the proportion of overpay paid for a player based on their cost (overpay is denoted
by the y-axis value being greater than 1). From the figure, it is clear that Dortmund paid
what each player was worth quite consistently, with Jadon Sancho being the only notable
exception. For him, Dortmund paid roughly 4 times what he was actually worth. Brighton
were similarly consistent, but instead of matching market values, they consistently paid
around 1.25 - 1.5 times more than a player’s market value. This figure showcases United in
a whole new light - though they were largely risk-averse, they overpaid for Anthony Martial
to such an extreme that they appear to be extremely risky. Generally speaking, though, it
seems that Brighton and Dortmund were the most risky clubs overall, with Dortmund just
about being the most risky club (taking into account the Sancho transfer). In the following
section, we will analyze how these risk-management strategies paid off in total value added
to the club for each player.

Next, we mapped how a player’s risk category determines the total value a player
provided for the club. The following three tables contain the value information for each of
the five players selected from the three chosen clubs. For each player, we recorded their
cost of acquisition (Cost of Acquisition) and sale value (Sold For)3. We also included a tag
for each player, denoting if they were a risky or non-risky buy (Risk Category). Note that
units of each monetary value in the tables are in millions of euros.

3
If a player has not been sold yet, we use the current market value of the player as an approximation of
what the player would sell for. This applies to Kaoru Mitoma, Deniz Undav, and Facundo Buonanotte, all
of Brighton.
Game Theory Report 2 11

Borussia Dortmund:
Player Risk Category Sold For (millions of Cost of Acquisition
€) (millions of €)

Jadon Sancho Risky 85 20.59

Pierre-Emerick Non-Risky 63.75 13

Aubameyang

Ousmane Dembele Risky 140 15

Alexander Isak Risky 6.5 9

Robert Risky 50 4.75

Lewandowski

Brighton:
Player Risk Category Sold For (millions of Cost of Acquisition
€) (millions of €)

Kaoru Mitoma Risky 35 3

Deniz Undav Risky 6 7

Leandro Trossard Non-Risky 30 15.56

Neal Maupay Risky 11.8 15.56

Facundo Buonanotte Non-Risky 9 6

Game Theory Report 2 12

Manchester United:
Player Risk Category Sold For (millions of Cost of Acquisition
€) (millions of €)

Alexis Sanchez Non-Risky 04 34

Robin Van Persie Non-Risky 6.5 30.7

Anthony Martial Risky 16 60

Cristiano Ronaldo 5 Non-Risky 0 17

Zlatan Ibrahimovic Non-Risky 0 0

Below are five graphs synthesized from the data collected above. The first three
graphs portray the value each player brought to their respective clubs. The fourth graph
compares the relative percentage loss/gain in value each player brought to their club. The
fifth graph compares the average percentage loss/gain for risky and non risky players.

4
Sanchez left Manchester United on a free transfer, meaning he didn’t renew his contract and joined his
next club, Inter Milan, without Milan having to pay up. Since Manchester United received no fee from Inter
Milan, we consider Sanchez’s ‘sold for’ cost to be 0. This is true also for Cristiano Ronaldo and Zlatan
Ibrahimovic.
5
Cristiano Ronaldo has transferred to Manchester United twice in his career - this data concerns only the
second transfer.
Game Theory Report 2 13

Figure 5. Depicts the change in player value between time of acquisition and time of selling
for Dortmund.

Figure 6. Depicts the change in player value between time of acquisition and time of selling
for Brighton.
Game Theory Report 2 14

Figure 7. Depicts the change in player value between time of acquisition and time of selling
for Manchester United.

Figure 8. Depicts the percentage change in player value between time of acquisition and
time of selling for all clubs’ players. Yellow bars correspond to Dortmund, blue bars to
Brighton, and red bars to United.
Game Theory Report 2 15

Figure 9. Depicts the average percentage change in player value between time of acquisition
and time of selling for risky and non-risky players.

In Figures 5, 6, and 7, we can see the value Dortmund, Brighton, and United are
getting out of their players, respectively.

As shown in Figure 5, Dortmund is making a huge return on investment for its

players, indicating that their strategy of choosing risky players is successful. Out of the five
players analyzed for Dortmund, four out of five of them are what we classify as “risky” bets.
On average, Dortmund made 56.6 million euros on each of its player acquisitions. It is
worth noting that one of the best “risky” bets that Dortmund profited from was Ousmane
Dembele. They purchased him for 15 million euros but sold him for 140 million euros, an
over 800% increase in his value at his time in Dortmund.

As shown in Figure 6, Brighton is making an average return on investment for its

players, indicating that their strategy of choosing a mix of risky and non-risky players is
moderately successful. Out of the five players analyzed for Brighton, three out of five of
them are what we classify as risky bets, while the other two are non-risky. On average,
Brighton made 9 million euros on each of its player acquisitions. Most of the players
Brighton bought did not change their value substantially in either direction with the
notable exception of Kaoru Mitoma. They purchased him for 3 million euros but sold him
for 35 million euros, an over 1000% increase in his value at his time in Brighton.

As shown in Figure 7, United is losing money on the players it invests in, a rather
troubling indication that their strategy of playing it safe by choosing big-name, non-risky
players is backfiring. Out of the five players analyzed for United, four out of five of them are
Game Theory Report 2 16

what we classify as “non-risky” bets. On average, United lost 20 million euros on each of its
player acquisitions. All of the players that United bought either lost value or stayed static.

Figures 8 and 9 synthesize the data from Figures 5, 6, and 7. In Figure 8, we see that,
on a relative scale, many of Dortmund’s biggest bets paid off big time; growth rates are off
the charts, with some over 500%. Again, Brighton was more of a mixed bag, with most of
their players failing to experience significant growth or decay. United fell flat on their face,
so to speak, with all of their “safe” bets actually losing value over time. Luckily, these losses
were limited to 100% or less. In Figure 9, it becomes clear that there is more to what we are
seeing than any one club - the real determining factor (or at least one of the major ones) in
how a player fares comes down to whether they were a risky or non-risky bet. While, on
average, both risky and non-risky players yielded some return on investment, risky players
yielded an almost 400% increase in value during their time playing at the club. For
non-risky players, that number was only 40%. To put it in perspective, a 400% increase in a
player’s value would mean a player is worth five times more than they were when they first
joined the club; if they were worth 5 million euros at time of acquisition, then at time of
selling they would be worth 25 million euros.

Limitations:
While this report was a thorough analysis of the data collected from the 3 involved
clubs, it is important to note that there are many limitations to our analysis which must be
kept in mind. For one, we only considered 5 players from each club. This is quite a small
sample size, meaning one outlier could skew the results of our analysis quite consistently.
As discussed, for two of our clubs in particular, an outlier certainly shifted the perception of
the club more than one might have expected.
Additionally, we are only considering the ‘cost of acquisition’ and ‘sold for cost’ of
each player when determining the value added to their club. This is simplifying a number of
variables down into two numbers. One such example is “Goals and Assists”. For simplicity’s
sake, we weren’t able to find a heuristic function which reasonably estimated the monetary
value of our player’s goals and assists during their time at the club. Had we been able to
pursue this path, the ‘value added to club’ data would be more representative of how much
a player benefitted their team.
Another limitation was our use of the ‘market value’ metric from transfermarkt.us.
Though they are thorough in their analysis when determining market value, transfermarkt
isn’t 100% accurate, and oftentimes market values are updated after the player in question
is sold to more accurately reflect what the player’s market value was at the time of being
purchased.

Discussion:
Soccer clubs have a lot at stake when choosing which players will be the face of their
team for the years to come: fan base, reputation, and bottomline. However, by making wise
investment decisions from day one, they have a real shot at success. From our analysis, we
found that the proportion of players a club picks that are risky versus non-risky serves as a
strong indicator of success. Of course, there are other extraneous factors, but
risk-management strategy is a good place to start for any introspective club seeking to
revamp their team. In particular, by adopting a high-risk, high-reward strategy, clubs have
Game Theory Report 2 17

seen the greatest return on investment, and it stands to reason that such a strategy will
continue to pay dividends in the foreseeable future.
Game Theory Report 2 18

References

“Alexis Sánchez.” Wikipedia, Wikimedia Foundation, 2 Mar. 2023,

https://en.wikipedia.org/wiki/Alexis_S%C3%A1nchez.

Barnwell, B. (2020, May 5). Ranking the 50 Worst Premier League transfers of all time.
ESPN. Retrieved March 5, 2023, from
https://www.espn.com/soccer/soccer-transfers/story/4087601/ranking-the-50-w
orst-premier-league-transfers-of-all-time

Football transfers, rumours, market values and news. Transfermarkt. (n.d.). Retrieved
March 5, 2023, from https://www.transfermarkt.us/

Ghavamifar, Farnaz. “Game Theory as a Tool for Converting Risk to Opportunity .”

Researchgate.net, Jan. 2007,
https://www.researchgate.net/publication/228464214_Game_Theory_as_a_Tool_for
_Converting_Risk_to_Opportunity.

WhoScored. (n.d.). Retrieved March 6, 2023, from https://www.whoscored.com/

Wilson, R. C., Geana, A., White, J. M., Ludvig, E. A., & Cohen, J. D. (2014). Humans use directed
and random exploration to solve the explore–exploit dilemma. Journal of
Experimental Psychology: General, 143(6), 2074–2081.
https://doi.org/10.1037/a0038199