Reproduce “Early Detection of Latent Microstructure Regimes in Limit Order Books” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Prakul Sunil Hiremath, Vruksha Arun Hiremath
This research introduces a framework for the early detection of latent microstructure regimes in Limit Order Books, offering systematically positive lead-time for identifying impending liquidity stress. The proposed detector achieved a mean lead-time of +18.6 timesteps in simulations and +38 seconds in preliminary real-data analysis by identifying a 'latent build-up regime' characterized by subtle feature drifts.
Key insights
- Observable LOB stress is often preceded by a 'latent build-up regime' characterized by persistent but weak drift in LOB features that are below conventional detection thresholds.
- The detectability and lead-time of this latent regime are fundamentally governed by the signal-to-noise ratio (SNR) of the drift and the duration of the build-up phase.
- Depth erosion and HMM entropy emerge as primary leading indicators, signaling LOB instability well before reactive measures like order flow imbalance or volatility spikes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.