Intraday Limit Order Price Change Transition Dynamics Across Market Capitalizations Through Markov Analysis

Battle Count: 60

Highly relevant to quantitative trading, particularly for algorithmic execution strategies, market making, and intraday trading. The paper provides actionable insights: (1) capitalization-aware execution - patient strategies for HMC stocks vs dynamic strategies for LMC stocks; (2) time-adaptive algorithms - anticipating higher price persistence at open/close; (3) bid-ask asymmetry exploitation - sellers initiate closing dynamics earlier (T5 pre-close); (4) regime detection via clustering for adaptive parameter switching; (5) risk management - tighter controls for LMC bid-side during phase transitions. The Markov framework directly informs order placement timing, price shading logic, and inventory management decisions.

Moderate complexity. Core Markov chain estimation (TPM via MLE) is straightforward. However, the full pipeline involves: (1) processing multi-gigabyte tick data files, (2) state discretization into 9 categories, (3) G-test validation, (4) estimating 2,160 individual 9x9 TPMs, (5) computing spectral decomposition for metrics, (6) PCA + t-SNE dimensionality reduction, (7) two clustering algorithms (Hierarchical + DBSCAN), (8) stationary distribution computation, (9) Jensen-Shannon divergence calculations. The mathematical framework is well-defined but requires careful implementation of multiple statistical tests and machine learning components. Data preprocessing of 20-40 GB files per day adds engineering complexity.

The methodology is well-documented with clear mathematical formulations. MATLAB code for TPM estimation is referenced (Ref. 36). PyDTMC package is used for stationary distributions. scikit-learn is used for clustering. However, the raw tick-by-tick data from Algoseek is not redistributable due to licensing restrictions, limiting full reproducibility. The stock selection criteria, time intervals, and state discretization are clearly specified. Robustness analyses are provided in supplementary material.