{"paper_id":"2307.13807","title":"Sports Betting: an application of neural networks and modern portfolio theory to the English Premier League","authors":["Vélez Jiménez","Román Alberto","Lecuanda Ontiveros","José Manuel","Edgar Possani"],"paper_date":"2023-07-11","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2307.13807v1","page_url":"https://thequant.space/flowcharts/sports-betting-an-application-of-neural-networks-and-modern/","scores":{"empirical_rigor":7,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel approach that combines expected utility theory, deep learning, and advanced Kelly Criterion formulations to optimize betting strategies.","quote":"This paper presents a novel approach for optimizing betting strategies in sports gambling by integrating Von Neumann-Morgenstern Expected Utility Theory, deep learning techniques, and advanced formulations of the Kelly Criterion.","summary":"This paper presents a novel approach for optimizing betting strategies in sports gambling by integrating Von Neumann-Morgenstern Expected Utility Theory, deep learning techniques, and advanced formulations of the Kelly Criterion. By combining neural network models with portfolio optimization, our…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"All wagers result in a positive return (W1 > 0).","type":"market","stated":"explicit","quote":"It is assumed that all wagers result in a positive return, that is, W1 > 0."},{"assumption":"Odds and probabilities of events are fixed over time.","type":"market","stated":"explicit","quote":"It is assumed that the odds and probabilities of the events are fixed over time."},{"assumption":"No short selling or borrowing; money is infinitely divisible with no minimum bet requirement.","type":"execution","stated":"explicit","quote":"The analysis assumes certain conditions regarding the nature of the betting system. Specifically, it assumes the absence of short selling and borrowing, and assumes that money is infinitely divisible with no minimum bet requirement."},{"assumption":"Each wager amount ℓi is non‑negative (ℓi ≥ 0).","type":"execution","stated":"explicit","quote":"First, it is implied that the wager amount ℓi for each event i is non‑negative (ℓi ≥0) for all i."},{"assumption":"Total wager across all events does not exceed total wealth (Σi ℓi ≤ 1).","type":"execution","stated":"explicit","quote":"Second, the assumption is made that the total wager across all events satisfies the constraint"},{"assumption":"Data from the first matchweek of each season are excluded due to high uncertainty.","type":"data","stated":"explicit","quote":"It is important to note that observations for teams on the first matchweek of each season were removed from the dataset."}]},"datasets":{"status":"available","items":[{"name":"EA Sports ratings","provider":"EA Sports","kind":"alternative","access":"public","period":"2014/2015 to 2020/2021","universe":"English Premier League matches","quote":"were utilized to assess the overall quality, offense, midfield, and defense of teams on a weekly basis."},{"name":"Understats database","provider":"Understats","kind":"fundamentals","access":"public","period":"2014/2015 to 2020/2021","universe":"English Premier League match‑level team statistics","quote":"The Understats database provides match‑level data for each team in the English Premier League."},{"name":"Match odds from multiple bookmakers","provider":"Football Data U.K.","kind":"other","access":"public","period":"2014/2015 to 2020/2021","universe":"English Premier League matches","quote":"match odds from multiple bookmakers, obtained through Football Data U.K.."}],"known":[],"sample":{"start":null,"end":null,"market":"English Premier League betting","asset_class":"other","frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Commission (track take) tt is included; tt may be negative, affecting gross return R(ℓA)=1/(1+tt).","quote":"where the tt is known as the commission (commonly known as track take) that a casino charges. Therefore tt > 0. However, since the betting market is a non‑efficient market","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Commission (track take) tt is included; tt may be negative, affecting gross return R(ℓA)=1/(1+tt).","card_quote":"where the tt is known as the commission (commonly known as track take) that a casino charges. Therefore tt > 0. However, since the betting market is a non‑efficient market","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The proposed method achieved a profit of 135.8% relative to initial wealth during the latter half of the 2020/21 EPL season.","kind":"performance","quote":"our method achieved remarkable profits of 135.8% relative to the initial wealth during the latter half of the 20/21 season of the English Premier League."},{"finding":"Multinomial regression on the reduced variable set attained 41.76% accuracy (cross‑entropy 1.34), while using all variables gave 51.76% accuracy (cross‑entropy 1.22).","kind":"predictive-accuracy","quote":"the multinomial regression model trained on this reduced set of variables achieved a predictive accuracy of 41.76% and a cross‑entropy value of 1.34. In contrast, when trained on the complete set of variables, the model achieved an accuracy of 51.76% and a cross‑entropy value of 1.22."},{"finding":"When the market commission tt is negative, the gross return under the arbitrage strategy satisfies R(ℓA)=1/(1+tt) > 1, and the Sharpe ratio becomes positively infinite.","kind":"theoretical","quote":"... the Sharpe Ratio is positively infinite, as the numerator is positive and the"}],"reported_performance":{"strategy":null,"sharpe":null,"sharpe_benchmark":null,"annual_return_pct":null,"total_return_pct":135.8,"max_drawdown_pct":null},"out_of_sample":{"method":"holdout","headline_out_of_sample":true}},"independent_evidence":{"status":"available","supports":[],"contradicts":[],"qualifies":[],"compared_with":14,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":1,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[{"paper":{"paper_id":"2402.15588","title":"Sizing the bets in a focused portfolio","paper_date":"2024-02-23","authors":["Vuko Vukcevic","Robert Keser"],"empirical_rigor":2,"page_url":"https://thequant.space/flowcharts/sizing-the-bets-in-a-focused-portfolio/"},"how":"uses_method_or_data","quote":null}],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:24:55Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}