Contributed Factors in Predicting Market Values of Loaned Out Players of English Premier League Clubs

Muhammad Daffa Arviano Putra, Deshinta Arrova Dewi, Wahyuningdiah Trisari Harsanti Putri, Retno Hendrowati, Tri Basuki KurniawanInternational Journal of Advanced Computer Science and Applications 2023 · DOI

Abstract

The English Premier League (EPL) represents the pinnacle of professional football, where player "loans"—temporary transfers between clubs—are a common strategic tool. Each player carries a market value, an estimate of the cost to acquire their contract. This study investigates the factors influencing a player's market value at the conclusion of a loan period. Using a dataset of Transfermarkt records from 2004 to 2020, we applied linear regression analysis. Our findings indicate that post-loan market value is primarily driven by the player's initial market value, goals scored, total appearances, and the volume of loans managed by the parent club.

Introduction: The Economics of the Loan System

In the high-stakes environment of the EPL, the loan system serves multiple purposes: developing young talent, managing squad depth, and optimizing financial assets. A player's market value fluctuates during these periods based on performance and exposure. Despite the financial implications, academic research on the specific drivers of value at the end of a loan has been sparse. This paper aims to fill that gap, providing insights for club owners and investors to better manage player portfolios.

Methodology

The research followed a structured data science pipeline to model and predict market values.

Research methodology flowchart illustrating the progression from data collection to model selection.
  1. Data Acquisition: Data was scraped from Transfermarkt, covering 339 records from 40 unique EPL clubs between 2004 and 2020.
  2. Exploratory Data Analysis (EDA): Initial analysis revealed right-skewed distributions in financial and performance metrics. Log transformations were applied to normalize these variables.
  3. Feature Selection: Least Absolute Shrinkage and Selection Operator (LASSO) regression was utilized to identify the most significant predictors and prevent model overfitting by penalizing less impactful coefficients.
  4. Model Development and Validation: Three linear regression models were tested, ranging from a complex 9-predictor model to a streamlined 5-predictor model. Robustness was ensured through k-fold cross-validation (k=3, 5, and 10).

Key Insights and Discussion

The analysis yielded several critical observations regarding the valuation of loaned players:

  • The Dominance of Initial Valuation: The strongest predictor of a player's value at the end of a loan is their value at the start. As seen in the correlation matrix, there is a near-perfect positive relationship (0.96) between these two variables.
  • Performance "Nudges": While the starting value is the anchor, performance metrics like goals (0.66 correlation) and appearances (0.74 correlation) provide the necessary adjustments to the final valuation.
  • The "Loan Army" Effect: Interestingly, a slight negative correlation exists between the total number of players a club loans out and the individual value growth of those players. This suggests that a "quality over quantity" approach to loan placements may be more effective for value appreciation.
Correlation heatmap showing the relationships between performance metrics and market values.

Results

The researchers evaluated model performance using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared ($R^2$).

Surprisingly, Model 3—the simplest model with only five variables (Total Loans, Appearances, Starting Formation, Goals, and Starting Market Value)—outperformed more complex iterations. It achieved an $R^2$ between 90.1% and 90.5%, explaining the vast majority of variance in end-of-loan valuations. Furthermore, Model 3 demonstrated a significantly lower RMSE (~0.32) compared to its more complex counterparts (~7.0), indicating superior predictive precision.

The resulting predictive equation is: $Value_{End} = -2.282 - (0.046 \times TotalLoans) + (0.523 \times Appearances) - (0.00007 \times StartingFormation) + (0.105 \times Goals) + (0.871 \times Value_{Start})$

Conclusion and Future Directions

This study demonstrates that a few core performance and baseline metrics can accurately predict the market value of EPL players following a loan spell. While linear models proved highly effective, future research could expand this work by incorporating larger datasets, exploring non-linear machine learning architectures, and including "soft" variables such as injury history or social media influence.

Summary updated

Abstract

The top tier of the English football league division is occupied by the English Premier League (EPL). It has become a global phenomenon with exhilarating skills and has been one of the most-watched professional football leagues on the planet. The possibility of a player temporarily playing for a club other than the one to whom they are now contracted is known as a "loan player" in the English Premier League (EPL) hence, each player has a market value. Market value is an estimate of how much a player costs when a club wants to buy his contract from another club. The purpose of this study is to determine the factors that influence a player's market value at the conclusion of a loan period. With the Transfermarkt player transfer record dataset for the years 2004 through 2020, we use linear regression analysis. Our study found that a football player's market worth at the end of a loan period is influenced by several aspects, including market value at the beginning, goals, appearances, and total loan.
BibTeX
@article{putra2023contributed,
  author = {Muhammad Daffa Arviano Putra and Deshinta Arrova Dewi and Wahyuningdiah Trisari Harsanti Putri and Retno Hendrowati and Tri Basuki Kurniawan},
  title = {{Contributed Factors in Predicting Market Values of Loaned Out Players of English Premier League Clubs}},
  journal = {International Journal of Advanced Computer Science and Applications},
  year = {2023},
  doi = {10.14569/ijacsa.2023.0140939},
}
Figure