QuantumFeed Simulator

An interactive exploration of the high-fidelity, agent-based simulator for replicating and analyzing market microstructure dynamics.

Abstract: High-frequency trading (HFT) research is hampered by a "realism gap" in existing simulators. This paper introduces QuantumFeed, a novel, open-source, agent-based simulator designed to bridge this gap. It combines a high-performance C++ matching engine with a flexible Python agent framework, uses sophisticated stochastic models to generate realistic data, and is validated against the canonical stylized facts of financial time series.

The Core Challenge in HFT Research

Understanding the gap QuantumFeed was designed to fill.

The Simulation Gap

Existing simulators often fail to model the sub-second phenomena defining HFT. Many abstract away the core mechanics of price formation, like the Limit Order Book (LOB), which "significantly undermines the realism of the simulator" as HFT strategies exploit these very microstructures. This leads to a "microstructure fidelity gap," preventing accurate replication of emergent market events like flash crashes.

The QuantumFeed Contribution

QuantumFeed addresses this gap with a fourfold contribution: 1) A novel, open-source, distributed architecture for scalability and performance. 2) A high-fidelity microstructure model with a full-depth LOB. 3) A verifiable data generation process using sophisticated stochastic models. 4) A demonstration of scientific utility by replicating the 2010 Flash Crash.

Simulator in Action

A live, animated visualization of the QuantumFeed matching engine at work.

QFEED-SIM

BIDS

ASKS

Recent Trades

Interactive System Architecture

Click on components to learn more about the distributed, multi-tier system.

Select a component to see details.

Market Service (C++)

Matching Engine & LOB

Agent Service (Python)

Agent Personas & Logic

Orchestrator

Control & Configuration

Communication

Protobuf over TCP

High-Fidelity Data Generation Models

The stochastic processes that drive market dynamics.

Validation: Replicating Stylized Facts

QuantumFeed's output is validated against empirical properties of real financial markets.

Case Study: The 2010 Flash Crash

An interactive walkthrough of how QuantumFeed simulates a systemic crisis.

1. Stable Market
2. Initial Absorption
3. Liquidity Withdrawal
4. Feedback Cascade
5. Price Collapse