Explainable Patterns in Cryptocurrency Microstructure

Paper: arXiv 2602.00776

Authors: Bartosz Bieganowski, Robert Ślepaczuk

Abstract

We document stable cross-asset patterns in cryptocurrency limit-order-book microstructure: the same engineered order book and trade features exhibit remarkably similar predictive importance and SHAP dependence shapes across assets spanning an order of magnitude in market capitalization (BTC, LTC, ETC, ENJ, ROSE). The data covers Binance Futures perpetual contract order books and trades on 1-second frequency starting from January 1st, 2022 up to October 12th, 2025. Using a unified CatBoost modeling pipeline with a direction-aware GMADL objective and time-series cross validation, we show that feature rankings and partial effects are stable across assets despite heterogeneous liquidity and volatility. We connect these SHAP structures to microstructure theory (order flow imbalance, spread, and adverse selection) and validate tradability via a conservative top-of-book taker backtest as well as fixed depth maker backtest. Our primary novelty is a robustness analysis of a major flash crash, where the divergent performance of our taker and maker strategies empirically validates classic microstructure theories of adverse selection and highlights the systemic risks of algorithmic trading. Our results suggest a portable microstructure representation of short-horizon returns and motivate universal feature libraries for crypto markets.

Complexity vs Empirical Score

• Math Complexity: 3.5/10
• Empirical Rigor: 8.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper provides a strong empirical evaluation of crypto market microstructure across different market caps using high-frequency LOB data, machine learning (CatBoost/SHAP), and real-world execution backtests. It relies on applied ML and microstructure intuition rather than heavy theoretical math derivations, but excels in rigorous out-of-sample stress testing (including flash crash analysis) and model explainability.

Research Flowchart

  flowchart TD
    A["<b>Research Goal</b><br/>Test cross-asset stability of LOB feature importance & SHAP patterns"] --> B["<b>Data Inputs</b><br/>Binance Futures 1s LOB & trade data (2022–2025)<br/><i>BTC, LTC, ETC, ENJ, ROSE</i>"]
    
    B --> C["<b>Modeling Pipeline</b><br/>CatBoost with direction-aware GMADL loss & Time-Series CV"]
    
    C --> D["<b>SHAP Interpretability Analysis</b><br/>Evaluate cross-asset feature rankings & dependence shapes"]
    
    D --> E["<b>Strategy Backtesting</b><br/>Top-of-book Taker vs. Fixed-depth Maker backtests"]
    
    E --> F["<b>Key Outcomes</b><br/>1. Universal, portable LOB feature representations<br/>2. Flash crash divergence validates adverse selection theory"]

Cited References (26)

1. Deep limit order book forecasting — → 2403.09267
2. Constructing time-series momentum portfolios with deep multi-task learning — → 2306.13661
3. Augmented bilinear network for financial time series prediction (2023)
4. Machine learning model for Bitcoin exchange rate prediction using economic and technology determinants (2021)
5. LOB-based deep learning models for stock price trend prediction: A benchmark study (2023)
6. Continuous Auctions and Insider Trading (1985)
7. Bid, Ask and Transaction Prices in a Specialist Market with Heterogeneously Informed Traders (1985)
8. Empirical Market Microstructure: The Institutions, Economics, and Econometrics of Securities Trading (2007)

4 of 26 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.