Why Continuous Futures Backtests Change as Price History Is Adjusted

Summary

This forum exchange explains why a futures backtest can differ from a course example even when the strategy and settings are the same: the data series may be revised over time. It describes the platform’s 888 series as a continuously smoothed main-contract series, comparable to a forward-adjusted series. As time passes, historical prices are recalibrated for contract roll gaps, with older observations potentially changing more than recent ones. A backtest run against a later version of that series may therefore produce different results.

The discussion contrasts this with the 99 index series, which a participant says does not change, and the 889 series, described as backward-adjusted and also subject to changes, though generally smaller. This is a brief explanation rather than a controlled comparison: it gives no strategy details, data snapshots, or measured performance differences. Its practical lesson is to identify the exact futures series and adjustment convention used in a backtest, and to account for revisions when trying to reproduce results.

Key ideas

• A continuously smoothed main-contract series can have its past prices recalibrated as roll adjustments evolve.
• Older observations in the 888 series may change more than recent observations.
• The discussion describes the 99 index series as stable and the 889 backward-adjusted series as changeable to a lesser degree.
• Different versions or adjustment methods for futures data can lead to different backtest results.
• The exchange offers explanations but no quantitative test of the size of these effects.

Tags

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