Paper: arXiv 2610.05740

Authors: Anjali Thawait

Abstract

Market-regime models typically assume a finite set of discrete latent states. We examine whether high frequency limit-order-book dynamics exhibit distinct regime separation or apparent regimes result from discretising an underlying continuum, analysing deep limit-order book data for EURO STOXX 50 index futures across 987 clean trading days from 2022 to 2025 using 30, 45 and 60-second aggregation windows. Market states are represented by symmetric positive definite covariance matrices and analysed under Log-Euclidean and affine-invariant geometries; the resulting state cloud has an effective dimension of approximately 1.4. A causal anomaly layer filters statistical, calendar and rollover contamination, while label-free economic validation evaluates market-state separation independently of volatility-based proxy labels. A 17-method zoo spanning conventional and Riemannian representations, reinforced by full-cloud geometric certificates, consistently favours a continuum over a discrete regime structure. A dominant latent coordinate captures between 83.66% and 85.27% of variation in the covariance-state geometry and tracks VSTOXX without regime labels, remaining stable across years although its average level shifts with market conditions, indicating that conventional regimes are coarse quantisations. A divisive hierarchy spanning macro, macro-subregime, micro and micro-subregime tiers is evaluated using expanding-year walk-forward tests, with out-of-sample performance strongest at the macro tier while finer tiers fail to generalise; continuous fine-scale information nonetheless remains predictive, and its discretisation causes the performance loss. Forecasting performance peaks near ten minutes, reaching a maximum pooled R2_OOS of 0.7483. Overall, the latent stress coordinate defines a stable, low-dimensional geometric continuum that captures market stress and its predictive dynamics.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 9.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: This paper presents a highly rigorous empirical analysis of LOB dynamics using advanced mathematical concepts like Riemannian geometry. Its central finding of a continuous stress continuum rather than discrete regimes is novel and well-supported. The methodology is thoroughly detailed, and the provision of a GitHub repository enhances reproducibility.

Research Flowchart

  flowchart TD
    A["Research Goal: Discrete Regimes vs. Latent Continuum in LOB Dynamics?"] --> B{"Data: EURO STOXX 50 Futures LOB (2022-2025)"}
    B --> C["Methodology: Covariance Matrices (Log-Euclidean/Affine-Invariant Geometry) & Anomaly Filtering"]
    C --> D{"Computational Process: 17-Method Zoo (Conventional & Riemannian) & Geometric Certificates"}
    D --> E["Outcome 1: Effective Dimension ~1.4; Consistent Favoring of Continuum Over Discrete Regimes"]
    E --> F["Outcome 2: Dominant Latent Coordinate (83.66%-85.27% Variance) Tracks VSTOXX, Stable Across Years"]
    F --> G["Outcome 3: Divisive Hierarchy Fails at Finer Tiers; Continuous Information Predictive; Forecasting R2_OOS=0.7483"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a latent stress coordinate that defines a stable, low‑dimensional geometric continuum capturing market stress and its predictive dynamics.

Source sentence
Overall, the latent stress coordinate defines a stable, low-dimensional geometric continuum that captures market stress and its predictive dynamics.

Assumes 6

  • Scaling, standardisation, reference points and metric choice are explicitly specified, using affine‑invariant geometry as primary and Log‑Euclidean as diagnostic. model
    Source sentence
    We therefore specify scaling, standardisation, reference points, and metric choice explicitly, using AI as the primary geometry and LE as the diagnostic comparison.
  • A causal anomaly layer filters statistical, calendar and rollover contamination from the high‑frequency data. data
    Source sentence
    A causal anomaly layer filters statistical, calendar and rollover contamination
  • Covariance matrices are represented as symmetric positive definite and analysed under Log‑Euclidean and affine‑invariant geometries. model
    Source sentence
    Market states are represented by symmetric positive definite covariance matrices and analysed under Log-Euclidean and affine-invariant geometries
  • The analysis assumes that the latent stress coordinate varies smoothly over time and can be decomposed into slow and fast components. statistical
    Source sentence
    Predictive information shifts smoothly between slow and fast components as the decomposition timescale changes
Show 2 more
  • The 30‑, 45‑ and 60‑second aggregation windows provide sufficient resolution for the continuum analysis. data
    Source sentence
    using 30, 45 and 60-second aggregation windows
  • Economic validation uses VSTOXX tertiles as a proxy for market stress without relying on volatility‑based labels. other
    Source sentence
    Economic NMI measures agreement between the resulting state labels and VSTOXX tertiles

Data

  • Deep limit‑order‑book data for EURO STOXX 50 index futures order-book unknown
    2022-2025 · EURO STOXX 50 futures
    Source sentence
    analysing deep limit-order book data for EURO STOXX 50 index futures across 987 clean trading days from 2022 to 2025

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • A dominant latent coordinate captures between 83.66% and 85.27% of variation in the covariance‑state geometry and tracks VSTOXX without regime labels. statistical
    Source sentence
    A dominant latent coordinate captures between 83.66% and 85.27% of variation in the covariance-state geometry and tracks VSTOXX without regime labels
  • Forecasting performance peaks near ten minutes, reaching a maximum pooled out‑of‑sample R² of 0.7483. predictive-accuracy
    Source sentence
    Forecasting performance peaks near ten minutes, reaching a maximum pooled R2_OOS of 0.7483
  • Finer partitions improve in‑sample fit but transfer poorly out‑of‑sample, indicating over‑partitioning rather than data resolution causes the loss. negative
    Source sentence
    Although finer partitions improve in-sample fit, they transfer poorly out of sample
  • A 17‑method zoo consistently favours a continuum over a discrete regime structure across representations. statistical
    Source sentence
    A 17-method zoo spanning conventional and Riemannian representations, reinforced by full-cloud geometric certificates, consistently favours a continuum over a discrete regime structure
  • The micro‑subspace adds significant predictive value, but discrete micro‑regimes fail to recover this information. performance
    Source sentence
    The micro-subspace adds significant predictive value, with the largely linear forecasting relationship indicating that this information originates from the covariance state, not model complexity

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 28 related papers). Coverage is partial, so this is neither replication nor refutation.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.