{"paper_id":"2406.19406","title":"Dissecting Multifractal detrended cross-correlation analysis","authors":["Borko Stosic","Tatijana Stosic"],"paper_date":"2024-06-09","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2406.19406v2","page_url":"https://thequant.space/flowcharts/dissecting-multifractal-detrended-cross-correlation-analysis/","scores":{"empirical_rigor":5.5,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes several new algorithmic options to handle negative cross‑covariance between two time series, aiming to provide a more robust assessment of the multifractal cross‑correlation spectrum.","quote":"To this end we propose several new options to deal with negative cross-covariance among two time series, that may serve to construct a more robust view of the multifractal spectrum among the series.","summary":"In this work we address the question of the Multifractal detrended cross-correlation analysis method that has been subject to some controversies since its inception almost two decades ago. To this end we propose several new options to deal with negative cross-covariance among two time series, that…","strategy":null},"assumptions":{"status":"none_found","items":[]},"datasets":{"status":"available","items":[{"name":"Synthetic dataset used in [11]","provider":null,"kind":"synthetic","access":"public","period":null,"universe":null,"quote":"We test different algorithms on synthetic series with an exact analytical solution, as well as on daily price series of ethanol and sugar in Brazil from 2010 to 2023. We substantiate these proposals through a comprehensive assessment of a synthetic dataset used in [11], uncorrelated white noise series, and daily sugar and ethanol price data in Brazil from 2010 to 2023.","support":"full"},{"name":"Two independent white‑noise series (size 2^17 = 131072)","provider":null,"kind":"synthetic","access":"public","period":null,"universe":null,"quote":"3.2 White noise We have applied all the algorithms (the ones in the existing literature, and the ones proposed in this work) on two independent white noise series of size 217=131072 with the results presented in Fig. 3.","support":"full"},{"name":"Daily sugar and ethanol price series in Brazil","provider":"CEPEA/ESALQ (Centro de Estudos Avançados em Economia Aplicad","kind":"market-prices","access":"public","period":"2010‑2023","universe":"Brazilian sugar and ethanol commodities","quote":"We test different algorithms on synthetic series with an exact analytical solution, as well as on daily price series of ethanol and sugar in Brazil from 2010 to 2023. We substantiate these proposals through a comprehensive assessment of a synthetic dataset used in [11], uncorrelated white noise series, and daily sugar and ethanol price data in Brazil from 2010 to 2023.","support":"full"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"none","findings":[{"finding":"All newly proposed algorithms produce wider and more asymmetric multifractal spectra than the ABS (modulus MF‑DXA) approach when applied to the Brazil sugar‑ethanol price data.","kind":"statistical","quote":"On the other hand, all the algorithms proposed in this work yield wider, more asymmetric spectra in comparison to the ABS (modulus MF_DXA) approach.","support":"full"},{"finding":"The percentage of point pairs captured by the new algorithms varies from 20.4% (PM) to 67.5% (PS), with consistent multifractal parameters across options.","kind":"statistical","quote":"The percentage of point pairs captured by different flavors of the algorithm varies between 20.4% (PM) and 67.5% (PS), and the resulting multifractal parameters are rather consistent across the set of options, as can be seen in Tab. 2. MF-DXA ABS MFCCA PS 67.5% MS 32.5% PB 57.4% MB 42.6% PP 28.4% PM 20.4% MP 22.1% MM 29.0% Figure 4.","support":"full"},{"finding":"On uncorrelated white‑noise series, the standard MF‑DXA algorithm yields no results, while MF‑CCA produces divergent H(q) near q≈0 and a scattered f(α) spectrum.","kind":"negative","quote":"As covariance between the two series has an equal chance of assuming positive and negative values the MF-DXA algorithm produces no results at all, and the MFCCA algorithm yields (for roughly half of the segments where the q-th order fluctuation function happens to be positive) the 𝐻(𝑞) function that diverges close to 𝑞~0, and a cloud of points for the 𝑓(𝛼) spectrum.","support":"full"},{"finding":"The MS algorithm captures 32.5% of point pairs but shows distorted scaling for large segment sizes, pushing h(q) downwards for positive q and shifting the f(α) spectrum leftward.","kind":"statistical","quote":"For example, the MS algorithm captures 32.5% of point pairs, but the scaling is distorted for very large segment sizes, pushing the ℎ(𝑞) curve downwards for positive 𝑞 values, and thus the 𝑓(𝛼) spectrum to the left.","support":"full"},{"finding":"The PP algorithm, with 28.4% point‑pair capture, provides the most linear log‑log plots across q values, indicating the most consistent scaling among detailed‑restriction methods.","kind":"statistical","quote":"The most consistent of more detailed restriction algorithms (the most “linear” of the log-log plots for different 𝑞 values) appears to be PP with 28.4% point pairs captured.","support":"full"}],"reported_performance":null,"out_of_sample":{"method":"none","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":18,"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":true,"code_url":"https://github.com/borkostosic/MFDCCALIB","note":null},"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":"2026-11-14T19:56:52Z"},"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/"}