Choosing Tick or Minute Bars for CTA Backtesting

Summary

This forum exchange weighs tick-level data against minute-bar data for backtesting CTA strategies. One participant notes that a year of tick observations can exceed a computer’s practical capacity, limiting the period available for testing and parameter tuning. They also ask whether the greater detail of tick data can encourage overfitting. Another participant replies that minute bars are generally sufficient for CTA work.

The original poster questions whether minute bars represent execution realistically, since many ticks occur inside a longer bar and could affect order matching and fills. The discussion raises a real tradeoff between computational scale, detail, and execution fidelity, but it does not resolve it or provide comparative tests. It gives no strategy-specific evidence about when tick data changes conclusions. Data granularity should therefore be chosen in light of the strategy’s holding period and execution assumptions, while recognizing that more detailed inputs can also make parameter fitting more vulnerable to overfitting.

Key ideas

• Tick data can create a large computational burden and restrict the historical period available for testing.
• Minute bars are presented as a practical default for many CTA strategies.
• The poster questions whether minute bars capture intrabar order matching and fills accurately enough.
• The discussion raises overfitting as a possible concern with highly granular data but presents no test results.
• The exchange does not establish a universal best data frequency.

Tags

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