Paper: arXiv 2403.09267

Abstract

We exploit cutting-edge deep learning methodologies to explore the predictability of high-frequency Limit Order Book mid-price changes for a heterogeneous set of stocks traded on the NASDAQ exchange. In so doing, we release `LOBFrame’, an open-source code base to efficiently process large-scale Limit Order Book data and quantitatively assess state-of-the-art deep learning models’ forecasting capabilities. Our results are twofold. We demonstrate that the stocks’ microstructural characteristics influence the efficacy of deep learning methods and that their high forecasting power does not necessarily correspond to actionable trading signals. We argue that traditional machine learning metrics fail to adequately assess the quality of forecasts in the Limit Order Book context. As an alternative, we propose an innovative operational framework that evaluates predictions’ practicality by focusing on the probability of accurately forecasting complete transactions. This work offers academics and practitioners an avenue to make informed and robust decisions on the application of deep learning techniques, their scope and limitations, effectively exploiting emergent statistical properties of the Limit Order Book.

Complexity vs Empirical Score

  • Math Complexity: 6.0/10
  • Empirical Rigor: 7.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced deep learning models (DeepLOB) and discusses microstructural theory, indicating moderate-to-high math complexity, while also releasing an open-source codebase (LOBFrame) for processing large-scale NASDAQ data and proposing a novel operational evaluation framework, demonstrating high empirical rigor.

Research Flowchart

  flowchart TD
  A["Research Goal: Forecast Mid-Price<br>Changes in High-Frequency LOB"] --> B["Methodology: <br>Deep Learning Models"]
  A --> C["Data Input: <br>NASDAQ Limit Order Book Data"]
  B --> D{"Computational Process"}
  C --> D
  D --> E["Analyze Forecasting Power<br>vs. Microstructural Characteristics"]
  D --> F["Evaluate Model Utility<br>via Transaction Accuracy Framework"]
  E --> G["Key Findings/Outcomes"]
  F --> G
  subgraph G ["Outcomes"]
      H["Deep Learning shows<br>high prediction accuracy"]
      I["Accuracy does not equal<br>actionable trading signals"]
      J["Proposed operational<br>evaluation framework"]
      K["Released Open-Source<br>LOBFrame library"]
  end

Cited References (126)

  1. NASDAQ Official Website
  2. What is LOBSTER?
  3. NASDAQ Stock Screener
  4. Companies Market Cap
  5. Deep reinforcement learning for active high frequency trading (2021) — link
  6. Trades, quotes and prices: financial markets under the microscope (2018)
  7. LOB-Based Deep Learning Models for Stock Price Trend Prediction: A Benchmark Study — → 2308.01915
  8. Towards Robust Representation of Limit Orders Books for Deep Learning Models (2021) — link
and 118 more
  1. The short-term predictability of returns in order book markets: a deep learning perspective (2022) — link
  2. A survey of data leakage detection and prevention solutions (2012)
  3. Transaction costs (1989)
  4. Deeplob: Deep convolutional neural networks for limit order books (2019)
  5. Market microstructure: A survey (2000)
  6. Market microstructure: A survey of microfoundations, empirical results, and policy implications (2005)
  7. Econophysics review: I. Empirical facts (2011)
  8. Econophysics review: II. Agent-based models (2011)
  9. Statistical modeling of high-frequency financial data (2011)
  10. Limit order books (2013)
  11. Limit order books (2016)
  12. Statistical properties of stock order books: empirical results and models (2002)
  13. Empirical analysis of limit order markets (2004)
  14. The price impact of order book events: market orders, limit orders and cancellations (2012)
  15. Order flow and price formation (2021) — link
  16. A continuous and efficient fundamental price on the discrete order book grid (2018)
  17. An empirical behavioral model of price formation (2005)
  18. The price impact of order book events (2014)
  19. High-frequency trading in a limit order book (2008)
  20. Enhancing trading strategies with order book signals (2018)
  21. Optimal execution with limit and market orders (2015)
  22. Limit order strategic placement with adverse selection risk and the role of latency (2017)
  23. Exogenous and endogenous price jumps belong to different dynamical classes (2022)
  24. Studies of the limit order book around large price changes (2009)
  25. Price jump prediction in limit order book (2012) — link
  26. Universal features of price formation in financial markets: perspectives from deep learning (2021)
  27. Daily volume forecasting using high frequency predictors (2010)
  28. Identifying expensive trades by monitoring the limit order book (2017)
  29. Effects of the limit order book on price dynamics (2014)
  30. Designating market maker behaviour in limit order book markets (2018)
  31. Stock market analysis: A review and taxonomy of prediction techniques (2019)
  32. Using deep learning to detect price change indications in financial markets (2017)
  33. Forecasting stock prices from the limit order book using convolutional neural networks (2017)
  34. Time-series classification using neural bag-of-features (2017)
  35. Machine learning for forecasting mid-price movements using limit order book data (2019)
  36. Using deep learning for price prediction by exploiting stationary limit order book features (2020)
  37. Deep learning modeling of limit order book: A comparative perspective (2020) — link
  38. Self-organised criticality in high frequency finance: the case of flash crashes (2021) — link
  39. Heterogeneous criticality in high frequency finance: a phase transition in flash crashes (2022)
  40. Temporal attention-augmented bilinear network for financial time-series data analysis (2018)
  41. Deep adaptive input normalization for time series forecasting (2019)
  42. Data normalization for bilinear structures in high-frequency financial time-series (2021)
  43. Transformers for limit order books (2020) — link
  44. Temporal logistic neural bag-of-features for financial time series forecasting leveraging limit order book data (2020)
  45. Multi-horizon forecasting for limit order books: Novel deep learning approaches and hardware acceleration using intelligent processing units (2021) — link
  46. Forecasting the Mid-price Movements with High-Frequency LOB: A Dual-Stage Temporal Attention-Based Deep Learning Architecture (2023)
  47. Axial-LOB: High-Frequency Trading with Axial Attention (2022)
  48. Price dynamics in a Markovian limit order market (2013)
  49. Market microstructure design and flash crashes: A simulation approach (2013)
  50. Effects of limit order book information level on market stability metrics (2017)
  51. Market making with signals through deep reinforcement learning (2021)
  52. Deep reinforcement learning for market making (2020)
  53. Reinforcement learning approaches to optimal market making (2021)
  54. Deep Reinforcement Learning for High-Frequency Market Making (2023)
  55. JAX-LOB: A GPU-Accelerated limit order book simulator to unlock large scale reinforcement learning for trading (2023)
  56. Asynchronous Deep Double Duelling Q-Learning for Trading-Signal Execution in Limit Order Book Markets — → 2301.08688
  57. Reinforcement learning for optimized trade execution (2006)
  58. Machine learning for market microstructure and high frequency trading (2013)
  59. Modelling limit order trading with a continuous action policy for deep reinforcement learning (2023)
  60. Diversity-driven knowledge distillation for financial trading using deep reinforcement learning (2021)
  61. Deep reinforcement learning for financial trading using price trailing (2019)
  62. Benchmark dataset for mid-price forecasting of limit order book data with machine learning methods (2018)
  63. FinRL-Meta: Market environments and benchmarks for data-driven financial reinforcement learning (2022)
  64. Predicting stock price changes based on the limit order book: a survey (2022)
  65. Deep order flow imbalance: Extracting alpha at multiple horizons from the limit order book (2023)
  66. How and When are High-Frequency Stock Returns Predictable? (2022)
  67. Dependency structures in cryptocurrency market from high to low frequency (2022)
  68. Gradient-based learning applied to document recognition (1998)
  69. An introduction to convolutional neural networks (2015) — link
  70. Understanding of a convolutional neural network (2017)
  71. Long short-term memory (1997)
  72. A review on the long short-term memory model (2020)
  73. Decoupled weight decay regularization (2017) — link
  74. Adam: A method for stochastic optimization (2014) — link
  75. nanoGPT
  76. Language models are few-shot learners (2020)
  77. UCL HPC CLuster Specs
  78. Pytorch: An imperative style, high-performance deep learning library (2019)
  79. Dark trading and price discovery (2015)
  80. Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation (2020) — link
  81. Ten quick tips for machine learning in computational biology (2017)
  82. An ecological perspective on the future of computer trading (2013)
  83. How markets slowly digest changes in supply and demand (2009)
  84. How market ecology explains market malfunction (2021)
  85. The Impact of High Frequency Trading on the Canadian Market (2009)
  86. High-frequency trading, stock volatility, and price discovery (2010)
  87. The impact of high-frequency trading on markets (2011)
  88. A dysfunctional role of high frequency trading in electronic markets (2012)
  89. Where is the value in high frequency trading? (2012)
  90. Deep learning for limit order books (2019)
  91. Sequence classification of the limit order book using recurrent neural networks (2018)
  92. Comparing two K-category assignments by a K-category correlation coefficient (2004)
  93. Elementary principles in statistical mechanics (1902)
  94. Activation functions: Comparison of trends in practice and research for deep learning (2018) — link
  95. Deep LOB trading: Half a second please! (2023)
  96. Trading with the momentum transformer: an intelligent and interpretable architecture (2021) — link
  97. Attention is all you need (2017)
  98. Deep unsupervised learning using nonequilibrium thermodynamics (2015)
  99. Denoising diffusion probabilistic models (2020)
  100. Generative modeling by estimating gradients of the data distribution (2019)
  101. Improved denoising diffusion probabilistic models (2021)
  102. Are transformers effective for time series forecasting? (2023)
  103. Informer: Beyond efficient transformer for long sequence time-series forecasting (2021)
  104. Transformers in time series: A survey (2022) — link
  105. Homological Convolutional Neural Networks (2023) — link
  106. Homological Neural Networks: A Sparse Architecture for Multivariate Complexity (2023)
  107. Sparsification and Filtering for Spatial-temporal GNN in Multivariate Time-series (2022) — link
  108. What is the optimal tick size? a cross-sectional analysis of execution costs on NASDAQ (2017)
  109. Market microstructure theory (1998)
  110. Market microstructure in practice (2018)
  111. Continuous auctions and insider trading (1985)
  112. Bid, ask and transaction prices in a specialist market with heterogeneously informed traders (1985)
  113. The flash crash: High-frequency trading in an electronic market (2017)
  114. Agent-based models for latent liquidity and concave price impact (2014)
  115. Applications of deep learning in stock market prediction: recent progress (2021)
  116. Limit Order Book Dynamics and Order Size Modelling Using Compound Hawkes Process — → 2312.08927
  117. Limit Order Book Simulations: A Review — → 2402.17359
  118. Generative AI for End-to-End Limit Order Book Modelling: A Token-Level Autoregressive Generative Model of Message Flow Using a Deep State Space Network (2023)

4 of 126 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.