Paper: arXiv 2311.04599

Authors: Chunyang Huang, Shaoliang Zhang

Abstract

This study introduces an advanced machine learning method for predicting soccer players’ market values, combining ensemble models and the Shapley Additive Explanations (SHAP) for interpretability. Utilizing data from about 12,000 players from Sofifa, the Boruta algorithm streamlined feature selection. The Gradient Boosting Decision Tree (GBDT) model excelled in predictive accuracy, with an R-squared of 0.901 and a Root Mean Squared Error (RMSE) of 3,221,632.175. Player attributes in skills, fitness, and cognitive areas significantly influenced market value. These insights aid sports industry stakeholders in player valuation. However, the study has limitations, like underestimating superstar players’ values and needing larger datasets. Future research directions include enhancing the model’s applicability and exploring value prediction in various contexts.

Complexity vs Empirical Score

  • Math Complexity: 2.5/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Street Traders — empirical and applied, lower math complexity

Why this score: The paper primarily uses standard machine learning models (GBDT) with accessible explanations (SHAP) and relies on a large, real-world dataset with performance metrics like R-squared and RMSE, indicating high empirical readiness for sports analytics but limited advanced mathematical theory.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Predict & Explain Soccer Player Market Value"] --> B{"Data & Feature Engineering"};
  B --> C["Feature Selection<br>Boruta Algorithm"];
  B --> D["Dataset: ~12,000 players (Sofifa)"];
  C --> E["Model Training<br>GBDT Ensemble"];
  D --> E;
  E --> F{"Model Evaluation"};
  F --> G["Findings<br>High Accuracy: R²=0.901, RMSE=3.2M"];
  F --> H["Interpretability<br>SHAP Analysis: Skills & Fitness are Key Drivers"];

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes an advanced machine learning method that combines ensemble models with SHAP for interpretable prediction of soccer players' market values.

Source sentence
This study introduces an advanced machine learning method for predicting soccer players' market values, combining ensemble models and the Shapley Additive Explanations (SHAP) for interpretability.

Assumes 5

  • The data is split into 80% training/validation and 20% testing, with imputation and feature selection performed only on the training set to avoid data leakage. data
    Source sentence
    In the construction of the predictive model, the dataset underwent a randomized spliting process: 80% was allocated for training and model validation, while the remaining 20% was designated for testing purposes.
  • Ensemble learning improves predictive accuracy by reducing variance and bias compared to single models. model
    Source sentence
    The rationale behind selecting ensemble modeling lies in its multiple benefits, which include diminishing variance and bias while concurrently enhancing the overall predictive accuracy of the model.
  • Hyperparameters are tuned using 5‑fold cross‑validation and grid search to obtain unbiased performance estimates. model
    Source sentence
    To optimize the hyperparameters of each ensemble learning model, we employed a 5-fold cross-validation technique coupled with Grid Search.
  • The Boruta algorithm can effectively identify the most relevant features for the market‑value prediction task. model
    Source sentence
    Utilization of the Boruta algorithm, implemented through the BorutaShap package in Python, allowed for a reduction in the feature set to 22.
Show 1 more
  • Player attributes in skills, fitness, and cognitive areas are the primary drivers of market value. market
    Source sentence
    Player attributes in skills, fitness, and cognitive areas significantly influenced market value.

Data

  • Sofifa player dataset · Sofifa alternative public
    approximately 12,000 professional soccer players
    Source sentence
    Utilizing data from about 12,000 players from Sofifa, the Boruta algorithm streamlined feature selection.

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • The Gradient Boosting Decision Tree (GBDT) model achieved the highest predictive performance with an R‑squared of 0.901 and the lowest RMSE of 3,221,632.175 on the test set. predictive-accuracy
    Source sentence
    Notably, within the test set, the GBDT model sustained its advantage, achieving the highest R-squared value of 0.901 and the lowest RMSE of 3,221,632.175, indicative of its robustness in predicting player market values.
  • Across six ensemble algorithms, GBDT outperformed CatBoost, LightGBM, Random Forest, XGBoost, and AdaBoost in both R‑squared and RMSE metrics. performance
    Source sentence
    Within the ensemble of six learning algorithms evaluated, the Gradient Boosting Decision Tree (GBDT) model demonstrated superior performance, achieving the highest R‑Squared value of 0.889... the
  • Feature selection reduced the original 29 attributes to 22 key variables, including acceleration, heading accuracy, defensive awareness, and others. statistical
    Source sentence
    Utilization of the Boruta algorithm, implemented through the BorutaShap package in Python, allowed for a reduction in the feature set to 22, as indicated by the green bars in Figure 3. Acceleration, Heading accuracy, Defensive awareness, Vision, Volleys, Sprint speed, Long passing, Positioning, Standing tackle, Dribbli

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 20 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (94)

  1. DeepHoops : Evaluating Micro - Actions in Basketball Using Deep Feature Representations of Spatio - Temporal Data (2019) — link
  2. Evaluation of creating scoring opportunities for teammates in soccer via trajectory prediction (2022)
  3. Actions Speak Louder than Goals : Valuing Player Actions in Soccer (2019) — link
  4. Measuring Football Players ’ On -the- Ball Contributions from Passes During Games (2019) — link
  5. tHoops : A Multi - Aspect Analytical Framework for Spatio - Temporal Basketball Data (2018) — link
  6. Attention is All you Need (2017) — link
  7. Scene Transformer : A unified architecture for predicting future trajectories of multiple agents (2021) — link
  8. Offline Reinforcement Learning as One Big Sequence Modeling Problem (2021) — link
and 86 more
  1. A Decision - Theoretic Generalization of On - Line Learning and an Application to Boosting (1997) — link
  2. LightGBM : A Highly Efficient Gradient Boosting Decision Tree (2017) — link
  3. Greedy Function Approximation : A Gradient Boosting Machine (2001) — link
  4. CatBoost : gradient boosting with categorical features support (2018) — link
  5. XGBoost : A Scalable Tree Boosting System (2016) — link
  6. Scikit-learn: Machine Learning in Python (2011) — link
  7. microsoft/ LightGBM : A fast, distributed, high performance gradient boosting ( GBT , GBDT , GBRT , GBM or MART ) framework based on decision tree algorithms, used for ranking, classification and many other machine learning tasks. — link
  8. CatBoost : gradient boosting with categorical features support (2018) — link
  9. CatBoost - state-of-the-art open-source gradient boosting library with categorical features support — link
  10. Multistrategy ensemble learning: reducing error by combining ensemble learning techniques (2004) — link
  11. Ensemble learning: A survey (2018) — link
  12. Choosing prediction over explanation in psychology: Lessons from machine learning (2017)
  13. Model evaluation, model selection, and algorithm selection in machine learning (2018) — link
  14. Entropy \textbar Free Full - Text \textbar Explainable AI : A Review of Machine Learning Interpretability Methods — link
  15. Explainable AI : A Review of Machine Learning Interpretability Methods (2021) — link
  16. Continuous control with deep reinforcement learning (2019) — link
  17. Random decision forests (1995) — link
  18. Consistent Feature Selection for Pattern Recognition in Polynomial Time (2007) — link
  19. Wrappers for feature subset selection (1997) — link
  20. Ensemble Classification and Regression - Recent Developments , Applications and Future Directions [ Review Article ] (2016) — link
  21. Feature Selection with the Boruta Package (2010) — link
  22. Principles and practical application of the receiver-operating characteristic analysis for diagnostic tests (2000) — link
  23. Risk Identification of Bronchopulmonary Dysplasia in Premature Infants Based on Machine Learning (2021) — link
  24. Machine learning for the prediction of cognitive impairment in older adults (2023) — link
  25. Explaining dimensionality reduction results using Shapley values (2021) — link
  26. Analysis of regression in game theory approach (2001) — link
  27. Diagnosing and Dealing with Multicollinearity (1990) — link
  28. Bargaining in the transfer market: theory and evidence (1993) — link
  29. The economics of football (2001)
  30. Towards Data - Driven Football Player Assessment (2016) — link
  31. The use of the area under the ROC curve in the evaluation of machine learning algorithms (1997) — link
  32. Estimating the market value of collegiate football players from professional factor shares (2017) — link
  33. A Multiple Linear Regression Approach For Estimating the Market Value of Football Players in Forward Position (2018) — link
  34. Football Player ’s Performance and Market Value (2015)
  35. Prediction Markets (2004) — link
  36. The Economics of Superstars (1981) — link
  37. Season-long increases in perceived muscle soreness in professional rugby league players: role of player position, match characteristics and playing surface (2016) — link
  38. Talent and/or Popularity : What Does It Take to Be a Superstar ? (2012) — link
  39. Stardom and Talent (1985) — link
  40. A Unified Approach to Interpreting Model Predictions (2017) — link
  41. Analysis of the Football Transfer Market Network \textbar SpringerLink — link
  42. Professional Sports Firm Values : Bringing New Determinants to the Foreground ? A Study of European Soccer , 2005-2013 - CORE Reader — link
  43. Hierarchical Grouping to Optimize an Objective Function (1963) — link
  44. More Is Different \textbar Science — link
  45. Predictive analysis and modelling football results using machine learning approach for English Premier League (2019) — link
  46. Beyond crowd judgments: Data -driven estimation of market value in association football (2017) — link
  47. When the crowd evaluates soccer players’ market values: Accuracy and evaluation attributes of an online community (2014) — link
  48. Predicting transfer fees in professional European football before and during COVID -19 using machine learning (2022) — link
  49. A Machine Learning Ensembling Approach to Predicting Transfer Values (2022) — link
  50. A novel machine learning method for estimating football players’ value in the transfer market (2021) — link
  51. Assessing the market values of soccer players – a robust analysis of data from German 1. and 2. Bundesliga (2019) — link
  52. Estimating transfer fees of professional footballers using advanced performance metrics and machine learning (2023) — link
  53. " Improving your data transformations: Applying the Box - Cox transformat" by Jason Osborne — link
  54. Improving your data transformations: Applying the Box - Cox transformation (2019) — link
  55. An Analysis of Transformations on JSTOR — link
  56. SciPy 1.0: fundamental algorithms for scientific computing in Python (2020) — link
  57. An Analysis of Transformations (1964) — link
  58. SHAP for additively modeled features in a boosted trees model (2022) — link
  59. Unifying local and global model explanations by functional decomposition of low dimensional structures (2023) — link
  60. A $k$-additive Choquet integral-based approach to approximate the SHAP values for local interpretability in machine learning (2022) — link
  61. From local explanations to global understanding with explainable AI for trees (2020) — link
  62. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery (2018) — link
  63. Estimating transfer fees of professional footballers using advanced performance metrics and machine learning (2023) — link
  64. March Madness prediction: Different machine learning approaches with non‐box score statistics (2023) — link
  65. Evaluating the Quality of Machine Learning Explanations : A Survey on Methods and Metrics (2021) — link
  66. Estimating transfer fees of professional footballers using advanced performance metrics and machine learning (2023) — link
  67. Intelligent and Fuzzy Techniques in Big Data Analytics and Decision Making : Proceedings of the INFUS 2019 Conference , Istanbul , Turkey , July 23-25, 2019 (2020) — link
  68. A novel machine learning method for estimating football players’ value in the transfer market (2021) — link
  69. Predicting Future Transfer Fees : Will we see more £100m transfers? (2013) — link
  70. Transfer Pricing : An Analysis Of The Impact Of Player Brand Value On Transfer Fees In European Football
  71. Predicting transfer fees in professional European football before and during COVID -19 using machine learning (2022) — link
  72. Leveraging Machine Learning to Predict NBA Career Longevity - Stumbling Through The Dark (2023) — link
  73. Introducing an expert system for prediction of soccer player ranking using ensemble learning (2019) — link
  74. Mass spectrometry‐based proteomics approaches to interrogate skeletal muscle adaptations to exercise — link
  75. Beyond crowd judgments: Data -driven estimation of market value in association football (2017) — link
  76. Risk early warning safety model for sports events based on back propagation neural network machine learning (2019) — link
  77. A Machine Learning Ensembling Approach to Predicting Transfer Values (2022) — link
  78. Data‐driven exploratory approach on player valuation in football transfer market (2021) — link
  79. Recruiting talent in a global sports market: appraisals of soccer players’ transfer fees (2021) — link
  80. High- Level Machine Learning Framework for Sports Events Ticket Sales Prediction (2021) — link
  81. Assessing the market values of soccer players – a robust analysis of data from German 1. and 2. Bundesliga (2019) — link
  82. Prediction of Football Player Value using Bayesian Ensemble Approach (2022) — link
  83. An XGBoost -lasso ensemble modeling approach to football player value assessment (2020) — link
  84. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery (2018) — link
  85. Multicollinearity and Regression Analysis (2017) — link
  86. A Machine Learning Approach to Assess Injury Risk in Elite Youth Football Players (2020) — link

0 of 94 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.