Simulation & Backtesting
Methodology
HFT simulation requires tick-level or order-book data replay rather than daily OHLCV bars. Historical tick data is replayed in chronological order, with configurable latency, queue position modeling, and fill assumptions (passive vs aggressive).
Key simulation parameters include round-trip latency, maker/taker fee structure, and queue position at the time of order submission. Survivorship bias and data quality are critical considerations when sourcing historical tick data.
HFT strategies require tick-level or order-book simulation to evaluate latency, fill assumptions, and market impact.
Tick-level replay
Replaying historical tick or order-book data in chronological order lets you test signal logic and execution logic against real market microstructure. Latency and queue position can be modeled.
Backtesting engine
Our High-Frequency Backtesting Engine is designed for ultra-fast simulation of HFT strategies. It supports tick data, configurable latency, and realistic fill models. See the Projects section on the homepage for the link.
Considerations
- Fill assumptions (e.g. passive vs aggressive) and queue position
- Latency and slippage modeling
- Survivorship bias and data quality for historical ticks