Skip to content
All library documents

Controlling Long-Term Trend Bias in Intraday Strategy Evaluation

Article Quant Q&A · Author: user2330624
Read the original

Summary

The document asks how to evaluate intraday strategies whose long-only trades may benefit from an asset’s longer-term upward drift. Its examples enter an S&P 500 or gold position at the market open and exit around midday. The concern is that apparent strategy performance could reflect the asset’s historical rise rather than an intraday effect.

No control method or answer is provided, and the document offers no performance data or comparison. It frames an important backtesting question: separate returns associated with the chosen intraday entry and exit from the baseline return that comes with holding a long position in the asset. A sound evaluation would need to define an appropriate benchmark and account for the strategy’s exposure and sampling period; the text does not specify how to do this or establish that the cited historical trend will persist.

Key ideas

  • A long-only intraday strategy can appear profitable partly because its underlying asset has risen over time.
  • The examples use morning entries and midday exits in an equity index and gold.
  • Strategy evaluation should distinguish intraday timing effects from the return of a long market exposure.
  • The document asks for controls but provides no benchmark, method, or empirical results.

Tags

BacktestingEquitiesCommoditiesStatistics
Full text
# How to Control for longer term trends in an asset against an intraday strategy?


# How to Control for longer term trends in an asset against an intraday strategy?












Say I have a simple S&P500 intraday strategy that buys at open, close at noon.

What are some good methods for controlling for the fact that the S&P500 tends to move higher over the long term and this strategy has a long bias?

Another example: Buy Gold at 9:30AM and flatten at Noon.

How would you control for the fact that gold has had an upwards trend over the last 10 years?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

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