Tick-Level Backtesting with Queue-Aware Order Matching

Summary

This article compares bar-based and tick-based backtests, focusing on how coarse bars can hide the path prices took within a period. That ambiguity can produce unrealistic fills or conceal whether a stop would have triggered before a favorable exit. It also discusses price gaps and limit-price fills as sources of mismatch between backtest assumptions and actual market behavior. Tick data provides a finer record, but the article argues that matching orders only when the displayed price reaches a limit is still insufficient for passive orders.

The proposed matching approach accounts for both price priority and queue volume. Its example tracks changes in displayed quantity at an unchanged price to infer when an order behind existing liquidity might fill. The article uses a sequence of ticks and a brief futures trade example to illustrate the method, then recommends choosing bar or tick resolution according to strategy frequency and holding period. These examples explain mechanics rather than prove predictive value. Queue changes can reflect cancellations as well as trades, and the article’s claimed realism is bounded by its data and simulation assumptions; it also notes that tick data support is limited to certain markets.

Key ideas

• Bar data can hide the sequence of intraperiod price moves, making entries, exits, and stops appear in the wrong order.
• Tick data improves temporal detail, but price-only matching can still overstate passive limit-order fills.
• A queue-aware simulator uses displayed volume changes at a price to estimate whether an order could have executed.
• Backtest resolution should reflect the strategy’s trading frequency and typical holding period.
• The matching method is a simulation whose accuracy depends on available tick data and assumptions about queue changes.

Tags

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