{"paper_id":"2511.05030","title":"The Shape of Markets: Machine learning modeling and Prediction Using 2-Manifold Geometries","authors":["Panagiotis G. Papaioannou","Athanassios N. Yannacopoulos"],"paper_date":"2025-11-07","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2511.05030v2","page_url":"https://thequant.space/flowcharts/the-shape-of-markets-machine-learning-modeling-and-predicti/","scores":{"empirical_rigor":4,"math_complexity":8.5,"quadrant":"Lab Rats"},"proposes":{"status":"available","statement":"The paper proposes a Geometry-Informed Model that embeds high‑dimensional market data onto constant‑curvature 2‑manifolds for financial forecasting.","quote":"We introduce a Geometry-Informed Model for financial forecasting by embedding high-dimensional market data onto constant-curvature 2-manifolds.","summary":"We introduce a Geometry Informed Model for financial forecasting by embedding high dimensional market data onto constant curvature 2manifolds. Guided by the uniformization theorem, we model market dynamics as Brownian motion on spherical S2, Euclidean R2, and hyperbolic H2 geometries. We further…","strategy":"Geometry-Informed Model (GIM) with toroidal embedding"},"assumptions":{"status":"available","items":[{"assumption":"Market dynamics follow Brownian motion on the chosen manifolds.","type":"model","stated":"explicit","quote":"we model market dynamics as Brownian motion on spherical (S2), Euclidean (R2), and hyperbolic (H2) geometries."},{"assumption":"The torus is a compact, flat manifold admissible as a quotient space of the Euclidean plane for capturing cyclical dynamics.","type":"model","stated":"explicit","quote":"We further include the torus (T), a compact, flat manifold admissible as a quotient space of the Euclidean plane‑anticipating its relevance for capturing cyclical dynamics."},{"assumption":"Raw daily prices are taken from Yahoo Finance and transformed into log‑returns.","type":"data","stated":"explicit","quote":"Raw daily prices (Yahoo Finance) ; log-returns are formed"},{"assumption":"Long‑only (LO) and risk‑parity (RP) benchmark series are constructed for reference.","type":"execution","stated":"explicit","quote":"basic long-only (LO) and risk-parity (RP) benchmark series are computed for reference."},{"assumption":"Eigenportfolios are built using expanding‑window PCA with lagged loadings to avoid look‑ahead bias.","type":"data","stated":"explicit","quote":"so pi,t is the (out-of-sample) return of the ith eigenportfolio."},{"assumption":"Simulated regimes are generated as correlated Brownian motions projected to 3D.","type":"other","stated":"explicit","quote":"Simulated Regimes: CrossManifold Performance"}]},"datasets":{"status":"available","items":[{"name":"Yahoo Finance daily price data","provider":"Yahoo Finance","kind":"market-prices","access":"public","period":null,"universe":"broad multi‑asset universe (equities, sectors, rates, credit, commodities, volatility indi","quote":"Raw daily prices (Yahoo Finance) ; log-returns are formed and basic long-only (LO) and risk-parity (RP) benchmark series are computed for reference.","known_id":"yahoo-finance"},{"name":"Simulated correlated Brownian motions (CBMs)","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"To check that geometry performance is not a PCA artifact, we evaluate on equicorrelated CBMs projected to 3D ."},{"name":"Simulated regime data for cross‑manifold performance","provider":null,"kind":"synthetic","access":"unknown","period":null,"universe":null,"quote":"Table 1 Crossmanifold performance on simulated data"}],"known":[],"sample":{"start":"2005","end":"2025","market":null,"asset_class":"multi-asset","frequency":"daily"},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":false,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"no","evidence_card":"n/a","evidence_card_quoted":false},"conflict_kind":"applicability","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"The toroidal geometry yields the highest Sharpe ratio (0.721) on the z‑coordinate, outperforming Euclidean and spherical benchmarks.","kind":"performance","quote":"The standout is the toroidal z coordinate (Sharpe 0.721), consistent with a latent cyclic component amplified by a T2 embedding."},{"finding":"Integrated Geometry‑Informed Model (GIM) outperforms all baselines, delivering a Sharpe of 0.64 with lower drawdown and turnover.","kind":"performance","quote":"Integrated GIM (eigenvalue + curvature) outperforms all baselines."},{"finding":"On simulated data, the torus (T2 GIM) achieves the lowest MAE (0.02) and RMSE (0.04) among tested manifolds.","kind":"predictive-accuracy","quote":"T2 GIM 0.02 0.04 49.59 0.407, 0.022, 0.020"},{"finding":"For flat correlated Brownian motions with high correlation (ρ=0.9), Euclidean geometry gives the highest Sharpe (0.211), while toroidal performs worse.","kind":"performance","quote":"High correlation ρ = 0.9: – Euclidean: Total Sharpe 0.211 – Spherical: −0.119 Toroidal: −0.172 Hyperbolic: −0.109"},{"finding":"With moderate correlation (ρ=0.6), toroidal geometry achieves the highest Sharpe (0.273) among the tested manifolds.","kind":"performance","quote":"Moderate correlation ρ = 0.6: – Toroidal: Total Sharpe 0.273 – Euclidean: 0.097 Spherical: −0.337 Hyperbolic: −0.310"}],"reported_performance":{"strategy":"Geometry-Informed Model (GIM) with toroidal embedding","sharpe":0.721,"sharpe_benchmark":0.39,"annual_return_pct":null,"total_return_pct":null,"max_drawdown_pct":null},"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":24,"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":"2026-11-14T17:22:43Z"},"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/"}