{"paper_id":"2407.21791","title":"Deep Learning for Options Trading: An End-To-End Approach","authors":["Wee Ling Tan","Stephen Roberts","Stefan Zohren"],"paper_date":"2024-07-31","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2407.21791v1","page_url":"https://thequant.space/flowcharts/deep-learning-for-options-trading-an-end-to-end-approach/","scores":{"empirical_rigor":8.5,"math_complexity":8,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes an end‑to‑end deep learning framework for options trading that learns mappings from market data to optimal trading signals without requiring option pricing models.","quote":"We introduce a novel approach to options trading strategies using a highly scalable and data-driven machine learning algorithm.","summary":"We introduce a novel approach to options trading strategies using a highly scalable and data-driven machine learning algorithm. In contrast to traditional approaches that often require specifications of underlying market dynamics or assumptions on an option pricing model, our models depart…","strategy":"LSTM deep learning model"},"assumptions":{"status":"available","items":[{"assumption":"Only standard monthly option contracts that expire on the third Friday of the month are used, with special settlements excluded.","type":"data","stated":"explicit","quote":"We consider only standard monthly option contracts that expire on the third Friday of the month and exclude any special settlements due…"},{"assumption":"Delta hedging is performed at initiation and the early‑exercise premium of American‑style options is ignored.","type":"execution","stated":"explicit","quote":"we disregard the early exercise premium embedded in the American-style options and therefore ignore the possibility of early assignment of short options."},{"assumption":"A constant annualized volatility target of 15% is applied to scale realized returns.","type":"model","stated":"explicit","quote":"We set the annualized volatility target 𝜎tgt to be 15% and estimate the ex‑ante volatility 𝜎(𝑖,straddle)𝑡 with a 20‑day exponentially weighted moving standard deviation…"},{"assumption":"The market is assumed to be frictionless for the purpose of backtesting, aside from the explicitly modeled transaction‑cost component.","type":"market","stated":"implied","quote":"Some of the key challenges of rebalancing an options portfolio include market microstructure considerations arising from market liquidity and bid‑ask spreads, which can result in high transaction costs."}]},"datasets":{"status":"available","items":[{"name":"Option‑Metrics Ivy DB database","provider":"Option‑Metrics","kind":"options","access":"commercial","period":"2010‑2023","universe":"Equity options on S&P 100 constituents","quote":"Our dataset consists of option contracts sourced from the Option‑Metrics Ivy DB database, comprising end‑of‑day bid‑ask prices, implied volatility and Greeks of individual options…","known_id":"optionmetrics"},{"name":"CRSP daily stock file","provider":"CRSP","kind":"other","access":"commercial","period":"2010‑2023","universe":"Underlying stocks of S&P 100 constituents","quote":"We obtain underlying stock prices from the Center for Research in Security Prices (CRSP), which we use for computing option money‑ness and accounting for corporate actions such as stock splits.","known_id":"crsp"}],"known":[{"id":"crsp","name":"CRSP","provider":"Center for Research in Security Prices","kind":"market-prices","access":"commercial","quote":"We obtain underlying stock prices from the Center for Research in Security Prices (CRSP), which we use for computing option money-ness and accounting for corporate actions such as stock splits.","mentions":1,"used_in":1},{"id":"optionmetrics","name":"OptionMetrics IvyDB","provider":"OptionMetrics","kind":"options","access":"commercial","quote":"3 OVERVIEW OF DATASET Our dataset consists of option contracts sourced from the Option-Metrics Ivy DB database, comprising end-of-day bid-ask prices, implied volatility and Greeks of individual options.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":"Equities listed on the S&P 100","asset_class":"options","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"Transaction costs are incorporated via turnover‑adjusted returns with costs ranging from 0 to 50 basis points per trade.","quote":"compute Sharpe ratios adjusted for transaction costs by taking into account turnover‑adjusted returns","source":"evidence_card","rule":"R2","candidates":{"front_matter":"unclear","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"Transaction costs are incorporated via turnover‑adjusted returns with costs ranging from 0 to 50 basis points per trade.","card_quote":"compute Sharpe ratios adjusted for transaction costs by taking into account turnover‑adjusted returns","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The LSTM deep‑learning model achieves the highest risk‑adjusted performance, with a Sharpe ratio of 1.329 when rescaled to a 15% target volatility.","kind":"performance","quote":"(Table 2: Performance Metrics – Rescaled to Target Volatility)"},{"finding":"Turnover regularization improves LSTM performance under high transaction costs, allowing it to outperform benchmarks up to 50 bps costs.","kind":"performance","quote":"turnover regularization further enhances the performance of the LSTM for high transaction costs of 𝑐= 10 to 50 bps, allowing the regularized model to outperform other strategies at prohibitively high levels of transaction costs."}],"reported_performance":{"strategy":"LSTM deep learning model","sharpe":1.329,"sharpe_benchmark":0.697,"annual_return_pct":26.6,"total_return_pct":null,"max_drawdown_pct":27.4},"out_of_sample":{"method":"walk-forward","headline_out_of_sample":true}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":10,"citations":{"cited_by_in_corpus":5,"by_independent_authors":4,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":4,"unclassified":1},"independent_papers":[],"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":"2026-11-14T18:05:02Z"},"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/"}