Backtesting High-Frequency Grid Market Making with Binance Futures Data

Summary

This tutorial presents a workflow for evaluating a high-frequency grid market-making approach on Binance Futures. It covers selecting trading pairs, obtaining historical depth and trade data, converting that data into the backtester’s format, modeling order-entry and response latency, and running a Rust-based simulation. Its analysis calculates equity from balance, position, mid-price, and fees, plots equity and position over time, and selects pairs with positive ending equity for further comparisons. The example assumes a maker rebate of 0.005%, described as the highest available in the referenced program.

The tutorial highlights that realistic latency assumptions matter and that data conversion can require substantial memory. It also notes rapid drawdowns during extreme price moves and points to multi-market combinations and price protection as possible areas to explore. The presented selection rule uses in-sample positive ending equity, so it does not by itself establish robust out-of-sample performance. Results depend on the rebate, latency model, data quality, and market period; the article provides no general guarantee of profitability.

Key ideas

• The workflow uses historical futures order-book and trade data to simulate grid market making.
• Data must be downloaded, converted, and paired with an explicit latency model before backtesting.
• The example evaluates equity and position paths, then retains pairs with positive ending equity.
• The simulations assume a 0.005% maker rebate, which materially affects the results.
• The article reports sharp drawdowns during extreme price moves and raises price protection as a possible mitigation.
• Positive ending equity in the selection period does not establish out-of-sample robustness.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.