DECISION SCIENCES PROJECT

DECISION ANALYSIS IN FOOTBALL TRANSFERS

GROUP-6

Ashik Ahmed – 190101032

A lot of money is involved with the transfers of top players in the big European football
leagues. For various reasons, obtaining a good economic valuation of football players
throughout the year is valuable, in other words, not just when a player has just transferred.
Furthermore, it is relevant to consider how the market value of a player relates to the
performance of that player. Both these factors again depend on the various parameters of the
player, which might be taken from various public sources on the web. In this project, we
demonstrate how a player and a club make a choice regarding a transfer while considering all
the possibilities by mapping the said options in the form of a decision tree.

Objective

The objective of the project is to understand the rationale behind a football transfer and to try
to predict the priorities of the players as well as the clubs. The payoffs for both parties in
almost all scenarios are available on the web for the decision tree to be complete.

Methodology

The methodology we are undertaking uses the concepts of Decision Analysis using
commonly available football data available on reliable sources on the internet (Opta sport,
Goal.com). These include hard values such as transfer values, wages, player statistics, etc,
probabilities of outcomes and alternative decisions and so on. Our main focus would be to
eventually use decision trees to find the best possible option for both parties considering our
chosen parameters and check whether the real-life scenario conforms to our findings. The
decision tree probabilities would be partly factual and partly intuitive in nature depending
upon the type of data that is being considered.

Literature review

Since the establishment of freedom of contract for Football League (FL)players during the
1977-78 league season, players have been able to negotiate a transfer to a new club on the
expiry of their existing contract. If a club currently holding a player’s league registration wish
to retain him, they are obliged to offer terms at least as favourable as those that applied in the
best year of his lapsing contract. If no offer is made, the player becomes eligible for a free
transfer. If such an offer is made and rejected, the player becomes eligible for transfer subject
to the payment of a transfer fee in respect of fair compensation, from the club taking over the
player’s league registration to the club relinquishing that registration, for the loss of the
player’s services.From an economic perspective, the compensation fee sought and offered by
clubs selling and buying a player’s registration are likely to be determined by the same basic
principles as apply in the sale and purchase of any (human) capital asset. That is, the present
discounted value of the expected future revenue now to a club arising from a player’s
contribution to on-the-field success and related financial returns (gate receipts, sponsorships,
merchandising and other revenues).In these terms a player’s ability and worth, over and
above any innate ability and other individual specific endowments, may be expected to
increase with age and experience up to a certain point and thereafter decline. Quoted transfer
fees might therefore be expected to lie between the capitalized value of the player to the
buying club and the capitalized value of the player to the selling club. However, such
(possibly implicit) assessments of a player’s worth are complicated by uncertainty over future
form and injury, and difficulties in appraising an individual player’s contribution to collective
team performance (especially overtime and across clubs). Inevitably, there is therefore
considerable scope for disagreement between clubs as to what constitutes fair compensation
most obviously due to differences in time-horizons, discount rates and expected revenues.
Additionally, fees sought and offered may deviate from their underlying capitalized values
due to perceived demand and supply considerations. Moreover, to the extent that all players
are heterogeneous, a selling club, as incumbent holder of a player’s registration, may seek to
extract monopoly rent from the sale, the size of this sought rent being determined by the
buyer’s ability to pay. Similarly, the buyer, where sole bidder, may seek to exploit
monopsony rents, dependent on the financial position and need to sell of the seller.

Research in the field of sports analytics has tried to use several methods such as regression
etc. to draw a relationship between the various parameters of a player ranging from
performance to financial viability etc. to the eventual market value of the said player. The
dataset includes 381 players from 4 of Europe’s top leagues. Our requirement for the project
requires data on players playing in Forward positions and therefore, we use the data from the
research study. The results are as follows.

Performance Indicators for Forward Players

There are significant differences in performance between positions. It has been suggested,
that there are key characteristics needed to play in certain positions within soccer.

They undertook a technical analysis of playing positions within elite level international
soccer at the European Championships 2004. Players were classified by position into
goalkeepers, defenders, midfielders or strikers. In addition, we have used the t-test to test
whether there are significant differences between positions when it comes to market value.
When considering the market value between the four categories, p = 0:001 which indicates a
very significant difference between the four positions. Furthermore, when comparing the
market value between specific sub-categories within each group (e.g. comparing various
types of defenders amongst each other), we get p-values above 0:8, which suggest it makes
sense to group such very similar sub-categories. Hence, also from a market value perspective,
the four categories are justified.

We attempted modelling the performance over the entire set of players, but failed to find
satisfactory results, due to the variance of performance per positions. Focusing on the
forward players specifically, it becomes possible to model their performance. Forward
players play an important role in the valuation of players. Although they only represent 30%
of all players, they appear as winners of the FIFA World Player of the Year in 17 out of 24
years. No goalkeeper has ever won the prize. The KPIs for forward players have been
selected as follows. A good player should have few Fouls (F). Shots and Goals in Penalty
area (SP &GP) are a big plus for player performance. Shots on Target (ST), Goals from out
of Box (GB), Dribble successfully (D), Assists total (A) also contribute to the final results.
The model for forward players now is:
Our Example

In the previous section, we explained how a football transfer system works and based on that
concept, we will look at one specific example and analyze it through the lens of decision
analysis. The example we will look at is of a player named Pierre Emerick Aubameyang, who
moved from Borussia Dortmund, a German club to Arsenal, a London based club that plays
in the English Premier League (EPL). The deal saw Dortmund earning 55m euros as transfer
fee, a club record spending for Arsenal.

For our example, we will consider the following parameters:

 Player performance
 Player market value
 Player wages
 Prospective buyers
 Buying power of clubs
 Player’s financial viability apart from football (merchandise, promotions)
 Long term vision (length of deal)
 Championship prospects

The payoffs and probabilities have been based on statistical data over a five-year period and
the decision tree analysis will involve the above-mentioned aspects and we will consider the
terms of the deal from the player and the club point of view.

Decision Tree description

The values within the triangular node show the EMV for that particular outcome while the
values to the right of the outcome node are the respective payoffs in monetary terms(euros)
over a five-year period based on previous statistics and predictions made through our
regression model.
DECISION ANALYSIS FOR THE PLAYER

The decision tree shows that the player choice was made based on the above-mentioned
information which means that the maximum EMV was 39, which means that the best choice
for the player was to leave his current club Borussia Dortmund and move to Arsenal.

DECISION ANALYSIS FOR THE CLUB

Arsenal had to choose between buying a forward, a defender, a midfielder or not buying at
all. Their decision tree is shown below
It is evident that the best-case scenario for Arsenal was also to buy Aubameyang and this too
was based on the expected return over a five year period.
REFERENCES

1. https://onlinelibrary.wiley.com/doi/epdf/10.1111/1467-9485.00053
2. https://link.springer.com/chapter/10.1007/978-3-030-23756-1_36
3. https://medium.com/@ar4198/football-transfers-predicting-using-machine-learning-
classification-techniques-66b489301d88
4. https://www.datasciencecentral.com/profiles/blogs/2018-world-cup-predictions-using-
decision-trees
5. http://web.tecnico.ulisboa.pt/mcasquilho/compute/_linpro/TaylorB_module_b.pdf
6. https://www.footballdatabase.eu/fr/
7. http://web.stanford.edu/group/msande135/cgi-bin/wp/wp-
content/uploads/2017/01/NPSM065-04-15.pdf
8. https://gclaw.wordpress.com/2014/05/13/player-contracts-football-transfers-v-european-
union-law-analysis/
9. https://www.transfermarkt.co.in/pierre-emerick-aubameyang/profil/spieler/58864
10. https://bleacherreport.com/articles/2644862-pierre-emerick-aubameyang