{"paper_id":"2306.13371","title":"Fractal properties, information theory, and market efficiency","authors":["Xavier Brouty","Matthieu Garcin"],"paper_date":"2023-06-23","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2306.13371v1","page_url":"https://thequant.space/flowcharts/fractal-properties-information-theory-and-market-efficienc/","scores":{"empirical_rigor":6.5,"math_complexity":9,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper provides a theoretical expression for market information based on entropy when log‑prices follow a fractional Brownian motion or its stationary delampertized version, and introduces a multiscale method to analyse entropy and market information at different information‑set sizes.","quote":"We thus provide a theoretical expression for the market information when log-prices follow either a fractional Brownian motion or its stationary extension using the Lamperti transform. In addition, we introduce a multiscale method to get a deeper interpretation of the entropy and of the market information, depending on the size of the information set.","summary":"Considering that both the entropy-based market information and the Hurst exponent are useful tools for determining whether the efficient market hypothesis holds for a given asset, we study the link between the two approaches. We thus provide a theoretical expression for the market information when…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Log‑prices are assumed to follow a fractional Brownian motion (fBm) with Hurst exponent H.","type":"model","stated":"explicit","quote":"We thus provide a theoretical expression for the market information when log-prices follow either a fractional Brownian motion or its stationary extension using the Lamperti transform. In what follows, we are interested in the entropy analysis for dynamics of log-prices generated by an fBm or by the stationary delampertized fBm.","support":"full"},{"assumption":"Log‑prices are also considered to follow a stationary delampertized fBm obtained via the Lamperti transform.","type":"model","stated":"explicit","quote":"The delampertized fBm relies on the Lamperti transform, which makes it possible to transform a stationary process Yt in a H-selfsimilar process, (LHY )t = tHYln(t), and reciprocally a H-selfsimilar 6 process Xt in a stationary process, (L−1 H X)t = e−HtXexp(t) [26]. This inverse Lamperti transform has a stationarising effect like the mean-reverting Ornstein-Uhlenbeck mechanism [29].","support":"full"},{"assumption":"When H = 1/2 the market information is zero; deviations from 1/2 increase market information.","type":"model","stated":"explicit","quote":"Consistently with the above remark, the case where H = 1/2, that is 5 where the fBm is a standard (Markovian) Brownian motion, is the only case for which the market information is equal to zero. When H moves away from 1/2, from above or from below, the market information increases.","support":"full"}]},"datasets":{"status":"available","items":[{"name":"Bitcoin (BTC‑USD)","provider":null,"kind":"crypto-onchain","access":"public","period":"January 2015 to January 2023","universe":null,"quote":"We focus on four datasets corresponding to various asset classes: a cryptocurrency, the Bitcoin (BTC-USD), between January 2015 and January 2023, two stock indices, the French CAC 40 index and the Japanese Nikkei 225 index, respectively between January 2012 and January 2023 and between January 2022 and June 2022, and an FX rate, the EUR/USD, between December 2006 and January 2016.","support":"full"},{"name":"CAC 40 index","provider":null,"kind":"market-prices","access":"public","period":"January 2012 to January 2023","universe":null,"quote":"We focus on four datasets corresponding to various asset classes: a cryptocurrency, the Bitcoin (BTC-USD), between January 2015 and January 2023, two stock indices, the French CAC 40 index and the Japanese Nikkei 225 index, respectively between January 2012 and January 2023 and between January 2022 and June 2022, and an FX rate, the EUR/USD, between December 2006 and January 2016.","support":"full"},{"name":"Nikkei 225 index","provider":null,"kind":"market-prices","access":"public","period":"January 2022 to June 2022","universe":null,"quote":"We focus on four datasets corresponding to various asset classes: a cryptocurrency, the Bitcoin (BTC-USD), between January 2015 and January 2023, two stock indices, the French CAC 40 index and the Japanese Nikkei 225 index, respectively between January 2012 and January 2023 and between January 2022 and June 2022, and an FX rate, the EUR/USD, between December 2006 and January 2016.","support":"full"},{"name":"EUR/USD FX rate","provider":null,"kind":"market-prices","access":"public","period":"December 2006 to January 2016","universe":null,"quote":"We focus on four datasets corresponding to various asset classes: a cryptocurrency, the Bitcoin (BTC-USD), between January 2015 and January 2023, two stock indices, the French CAC 40 index and the Japanese Nikkei 225 index, respectively between January 2012 and January 2023 and between January 2022 and June 2022, and an FX rate, the EUR/USD, between December 2006 and January 2016.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"intraday-bars"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"n/a","evidence_card":"no","evidence_card_quoted":true},"conflict_kind":"applicability","card_detail":null,"card_quote":"18 and do not take into account relevant liquidity constraints that a trader would face. If a trading strategy based on the information analysis leads the trader to issue a bid order, it cannot be at the close price if this one is an ask price.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"unclear","findings":[{"finding":"Theoretical market information is zero when H = 1/2 and increases as H moves away from 1/2, with a stronger increase for H > 1/2.","kind":"theoretical","quote":"Consistently with the above remark, the case where H = 1/2, that is 5 where the fBm is a standard (Markovian) Brownian motion, is the only case for which the market information is equal to zero. When H moves away from 1/2, from above or from below, the market information increases.","support":"full"},{"finding":"Empirically, for Bitcoin the market information is dominated by a single lag (L = 1) across all horizons, indicating short‑term memory in price‑increase indicators.","kind":"statistical","quote":"They illustrate that mainly one lag, L = 1, contributes to the market information, whatever the horizon m considered. The really relevant quantity to be considered is instead the difference between the market information and the confidence bound of the zero information introduced in Section 4.2.","support":"full"},{"finding":"For the EUR/USD minute‑frequency series, market inefficiency is observed at all time scales and lag lengths, with the most recent lagged observations being the most informative.","kind":"statistical","quote":"The results are gathered in Figure 9 and indicate the presence of market inefficiency at all the time scales and all the numbers of lags considered, even though the most recent lagged observations has a predominant informativeness.","support":"full"},{"finding":"The short‑term dynamics of the CAC 40 index exhibit a Hurst exponent of 0.60, suggesting forecasting opportunities for horizons shorter than about 50 minutes.","kind":"statistical","quote":"Indeed, the Hurst exponent for the short-term dynamic is 0.60, indicating forecasting opportunities for time horizons less than about 50 minutes.","support":"partial"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":21,"citations":{"cited_by_in_corpus":8,"by_independent_authors":6,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":5,"unclassified":3},"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":2,"extracted_at":"2026-11-14T20:02:30Z"},"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/"}