{"paper_id":"2409.07159","title":"Market information of the fractional stochastic regularity model","authors":["Daniele Angelini","Matthieu Garcin"],"paper_date":"2024-09-11","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2409.07159v3","page_url":"https://thequant.space/flowcharts/market-information-of-the-fractional-stochastic-regularity-m/","scores":{"empirical_rigor":3,"math_complexity":9,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a forecasting procedure for the regularity process H_t in the Fractional Stochastic Regularity Model and shows how it can be used to build statistical arbitrage strategies.","quote":"we determine theoretically the serial information of the regularity process H_t of the FSRM, giving some insight into one's ability to forecast future price increments and to build statistical arbitrages with this model.","summary":"The Fractional Stochastic Regularity Model (FSRM) is an extension of Black-Scholes model describing the multifractal nature of prices. It is based on a multifractional process with a random Hurst exponent H_t, driven by a fractional Ornstein-Uhlenbeck (fOU) process. When the regularity parameter…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"When the regularity parameter H_t equals 1/2 the efficient market hypothesis holds.","type":"market","stated":"explicit","quote":"When the regularity parameter H_t is equal to 1/2, the efficient market hypothesis holds"},{"assumption":"Past price returns contain information about future trends when H_t ≠ 1/2.","type":"model","stated":"explicit","quote":"1/2 past price returns contain some information on a future trend or mean-reversion of the log-price process"},{"assumption":"The time series of regularity H_t is not directly observable; only prices are observable.","type":"data","stated":"explicit","quote":"The time series of regularity Ht is never directly observable and the only observable data in our possession are prices"},{"assumption":"Hit variables must be independent for the binomial test to be valid.","type":"statistical","stated":"explicit","quote":"The use of a binomial test is legitimate only if the hit variables"},{"assumption":"Arbitrage opportunities disappear when specific transaction costs or a minimal time interval between trades are imposed.","type":"execution","stated":"explicit","quote":"arbitrage opportunities disappear when one imposes specific transaction costs or a minimal, and possibly extremely small, interval of time between two consecutive transactions"},{"assumption":"The forecast quality improves when H is large, η and λ are small, and the current regularity H_{mi} is far from 1/2.","type":"model","stated":"explicit","quote":"the quality of the forecast in the FSRM is improved when H is large, when η and λ are small, and when the current regularity Hmi is far from 1/2"}]},"datasets":{"status":"available","items":[{"name":"S&P 500 index (SP500)","provider":null,"kind":"market-prices","access":"unknown","period":"2010-03-29 09:30 to 2021-05-05 16:05","universe":"S&P 500 index","quote":"SP500 1,107,216 2010-03-29 09:30 2021-05-05 16:05 2,796"},{"name":"Dow Jones index (DJI)","provider":null,"kind":"market-prices","access":"unknown","period":"2010-03-29 09:30 to 2021-05-05 16:05","universe":"Dow Jones index","quote":"DJI 1,107,216 2010-03-29 09:30 2021-05-05 16:05 2,796"},{"name":"NYSE Composite index (NYSE)","provider":null,"kind":"market-prices","access":"unknown","period":"2006-07-20 09:30 to 2021-05-05 16:05","universe":"NYSE Composite index","quote":"NYSE 1,474,308 2006-07-20 09:30 2021-05-05 16:05 3,723"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"intraday-bars"},"transaction_costs":{"status":"not_extracted","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"none","rule":"R0","candidates":{"front_matter":"unclear","evidence_card":"unclear","evidence_card_quoted":false},"meaning":"the classifier was not confident"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"The predictor selects roughly 14% of S&P 500 days, 100% of Dow Jones days, and 38% of NYSE days, achieving hit rates close to 60% on those selected days.","kind":"performance","quote":"This happens roughly 14% of the days for S&P 500 index, 100% for Dow Jones index, and 38% for NYSE Composite index, with hit rates close to 60% for these selected days"},{"finding":"Serial information of the binarized fOU process shows a peak at high H in the stationary regime and shifts toward H=0.5 in the fractal regime; information grows when H ≪ 1/2 in the fractal regime.","kind":"theoretical","quote":"In the fractal regime, the lower λ, the more the peak moves towards the value of H = 1/2. Moreover, in this case, compared to the stationary regime, the serial information grows when the fOU is very rough, that is when H ≪1/2"},{"finding":"Forecast quality improves when the Hurst exponent H is large, and when the diffusion η and mean‑reversion λ parameters are small.","kind":"predictive-accuracy","quote":"the quality of the forecast in the FSRM is improved when H is large, when η and λ are small, and when the current regularity Hmi is far from 1/2"},{"finding":"BDS tests indicate independence of hit variables for large β values (β>0.69 for SP500, β>0.59 for τ=2, etc.), legitimizing the binomial test of hit rates.","kind":"statistical","quote":"We observe that the p-values rise above the 0.05 threshold, for large values of β in the case of SP500 and NYSE, and for all values of β in the case of DJI, implying that the null hypothesis of independence cannot be rejected"},{"finding":"The predictor for SP500 is significant only for τ=1 and high β, yielding about 60% hit rate on roughly 200 days; DJI predictor is efficient for τ=1 with hit rates above 57%; NYSE predictor significant for τ=1 and high β with hit rates above 60% on about 700 da","kind":"performance","quote":"β... hit rate of 60%... The predictor for DJI seems to be efficient for τ = 1 only, but whatever β, with a hit rate above 57%"}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":26,"citations":{"cited_by_in_corpus":1,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"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":"2026-11-14T18:02:44Z"},"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/"}