{"paper_id":"2312.16190","title":"Hawkes-based cryptocurrency forecasting via Limit Order Book data","authors":["Raffaele Giuseppe Cestari","Filippo Barchi","Riccardo Busetto","Daniele Marazzina","Simone Formentin"],"paper_date":"2023-12-21","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2312.16190v1","page_url":"https://thequant.space/flowcharts/hawkes-based-cryptocurrency-forecasting-via-limit-order-book/","scores":{"empirical_rigor":8,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a Hawkes‑based return sign prediction model that estimates the next LOB order time and combines it with a Continuous Output Error (COE) model to forecast return signs.","quote":"For the ﬁrst time, as far as we are aware, we develop a return sign prediction model based on Hawkes processes to estimate the next LOB order time and we couple the predicted information with a COE model to predict the return sign.","summary":"Accurately forecasting the direction of financial returns poses a formidable challenge, given the inherent unpredictability of financial time series. The task becomes even more arduous when applied to cryptocurrency returns, given the chaotic and intricately complex nature of crypto markets. In…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Existence of a Continuous Output Error (COE) model linking base imbalance to future returns.","type":"model","stated":"explicit","quote":"We assume the existence of a Continuous Output Error (COE) model that connects BI and future returns as follows:"},{"assumption":"Zero‑return events are removed from the sample.","type":"data","stated":"explicit","quote":"In particular, we stick to remove zero return events, possibly losing information content."},{"assumption":"No transaction costs are incurred in the trading simulation.","type":"execution","stated":"explicit","quote":"No transaction costs are considered in this exercise."},{"assumption":"The Oracle benchmark assumes perfect knowledge of the next event time.","type":"model","stated":"explicit","quote":"Oracle algorithm, when combined with the COE model, replicates the process performed in Busetto and Formentin (2023), but with our dataset. In essence, it assumes knowledge of the timing of the next event."}]},"datasets":{"status":"available","items":[{"name":"Tether (USDT) LOB dataset","provider":null,"kind":"order-book","access":"unknown","period":null,"universe":"USDT/USD on a centralized exchange","quote":"The research draws on LOB measurements from a centralized cryptocurrency exchange where the stablecoin Tether is exchanged against the U.S. dollar."},{"name":"Real cryptocurrency LOB data","provider":null,"kind":"order-book","access":"unknown","period":null,"universe":null,"quote":"We apply this prediction strategies on real cryptocurrency data collected in a LOB dataset."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"l2-book"},"transaction_costs":{"status":"available","modeled":"no","headline_net_of_costs":null,"detail":null,"quote":"No transaction costs are considered in this exercise.","source":"front_matter","rule":"R3","candidates":{"front_matter":"no","evidence_card":"no","evidence_card_quoted":true},"card_detail":null,"card_quote":"No transaction costs are considered in this exercise.","meaning":"a trading or execution result is reported without costs"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"The Hawkes‑based strategy outperforms benchmark algorithms in return‑sign accuracy, approaching the performance of an ideal Oracle.","kind":"performance","quote":"Hawkes outperforms the Naive and MA due to its machine learning-based prediction method for the next event time."},{"finding":"In trading simulations, the Hawkes strategy yields higher total profit than Naive and Moving Average benchmarks, closely matching the Oracle’s profit.","kind":"performance","quote":"Hawkes shows satisfactory performances, approaching ideal performances—next, the MA and Naive benchmarking strategies."},{"finding":"Base imbalance (BI) exhibits a strong negative correlation with returns (ρ = −0.96), confirming its usefulness as a regressor for cryptocurrency LOB data.","kind":"statistical","quote":"ρ = −0.96. This result, shows the promising relation between BI and R, confirming the intuition of Busetto and Formentin (2023) of the relevance of base imbalance feature as meaningful regressor for return sign prediction."}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":35,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"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":"2026-11-14T18:17:26Z"},"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/"}