A structured path from market fundamentals to advanced informed signal detection

Module One and Two: Market Microstructure and Order Flow Foundations

The first two modules establish the mechanical foundation. Module One covers how modern electronic markets match orders, how the limit-order book forms and evolves, and what the bid-ask spread actually measures. Module Two introduces order flow — how directional trades differ from liquidity-providing trades, and why that distinction is the starting point for any detection methodology. Both modules are anchored to documented real-market episodes from Australian and international equity venues, so the mechanics are always shown in context rather than described in isolation.

Module Three and Four: Measuring Informed Activity Statistically

Module Three introduces the Probability of Informed Trading model — its assumptions, its construction from trade and quote data, and its documented limitations in modern high-frequency environments. Module Four extends this to Volume-Synchronized Probability of Informed Trading, a methodology better suited to contemporary markets where trade arrival rates vary significantly across the session. Our team presents both methods with full worked examples, discussing where each measure performs well and where signal quality tends to degrade under specific liquidity conditions.

Module Five: Supervised Classification of Trade Direction and Sequence

This module applies machine learning classification methods to the problem of inferring trade direction and detecting informed sequences in order flow. Feature construction from tick data is covered in detail — including how to encode time, volume, and price interaction variables in a format suitable for classification algorithms. Our team documents the performance characteristics of several model families under realistic back-testing conditions, and discusses the practical trade-offs between model interpretability and predictive signal strength. All examples use historical data from documented market episodes.

Module Six and Seven: Anomaly Detection and Multi-Market Signal Synthesis

Module Six shifts from classification to anomaly detection — identifying volume, spread, and order-flow patterns that deviate from a statistically defined baseline in ways consistent with informed activity. Module Seven then addresses the more complex problem of synthesizing detection signals across multiple markets or instruments simultaneously, identifying cases where informed flow appears to migrate across venues or asset classes before a price event. Our team presents this as a pattern-recognition challenge with documented examples, not as a predictive exercise, and discusses the statistical caution required when interpreting cross-market signals.

Module Eight: Calibration, Regime Awareness, and Framework Application

The final module addresses a practical problem that the preceding modules raise: detection models calibrated on historical data behave differently when market regimes shift. This module covers signal recalibration methodology, regime identification using structural break tests, and the principles for determining when a previously reliable detection screen requires revision. Our team also presents a complete framework-application walkthrough using a documented real-market episode, demonstrating how all preceding modules connect into a coherent analytical process. The module concludes with guidance on contributing observations back to the community research loop.