Paper: arXiv 2304.02472

Authors: Artem Lensky, Mingyu Hao

Abstract

Introduction: The paper addresses the challenging problem of predicting the short-term realized volatility of the Bitcoin price using order flow information. The inherent stochastic nature and anti-persistence of price pose difficulties in accurate prediction. Methods: To address this, we propose a method that transforms order flow data over a fixed time interval (snapshots) into images. The order flow includes trade sizes, trade directions, and limit order book, and is mapped into image colour channels. These images are then used to train both a simple 3-layer Convolutional Neural Network (CNN) and more advanced ResNet-18 and ConvMixer, with additionally supplementing them with hand-crafted features. The models are evaluated against classical GARCH, Multilayer Perceptron trained on raw data, and a naive guess method that considers current volatility as a prediction. Results: The experiments are conducted using price data from January 2021 and evaluate model performance in terms of root mean square error (RMSPE). The results show that our order flow representation with a CNN as a predictive model achieves the best performance, with an RMSPE of 0.85+/-1.1 for the model with aggregated features and 1.0+/-1.4 for the model without feature supplementation. ConvMixer with feature supplementation follows closely. In comparison, the RMSPE for the naive guess method was 1.4+/-3.0.

Complexity vs Empirical Score

  • Math Complexity: 3.5/10
  • Empirical Rigor: 6.5/10
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper introduces an innovative spatial representation of limit order book data, converting order flow into multi-channel images for computer vision models like CNNs, ResNet, and ConvMixer. While mathematically lightweight, focusing primarily on standard deep learning architectures and basic error metrics, the empirical evaluation demonstrates clear performance improvements over traditional GARCH and naive baselines for short-term Bitcoin volatility.

Research Flowchart

  flowchart TD
    A[Goal: Predict Short-Term Bitcoin Realized Volatility] --> B[Input: Order Flow Data & LOB Snapshots<br/>Jan 2021 Data]
    B --> C[Transform Order Flow into Image Channels & Extract Features]
    C --> D[Train Models:<br/>CNN, ResNet-18, ConvMixer vs. GARCH & Baselines]
    D --> E[Evaluate Model Performance via RMSPE]
    E --> F[Outcome: CNN with Aggregated Features Achieves Best Performance<br/>RMSPE: 0.85 vs. Naive Baseline: 1.4]