Researching Trading Edges Before Building Strategies
Summary
This essay argues that systematic traders should investigate a market effect before building elaborate backtests, optimized rules, or machine learning systems. It recommends forming small, falsifiable hypotheses and using direct, data-efficient analysis such as scatter plots and factor plots to explore possible causes, consistency, strength, noise, and conditions under which an effect may fail. The aim is to understand the edge before engineering its implementation.
The author also emphasizes the probabilistic nature of trading: randomness can dominate an individual strategy’s short-term results, so a smooth equity curve should not be the expectation. Instead, traders are encouraged to tolerate strategy-level variation, manage exposures, and combine multiple opportunities at the portfolio level. The discussion is methodological rather than empirical; it offers no tested signal or measured performance. Its advice is framed around retail systematic trading and does not prescribe specific analysis procedures or portfolio targets.
Key ideas
- Investigate the observed market effect before building a trading implementation.
- Use simple analyses and small hypotheses to understand possible causes and failure conditions.
- Try to disprove an apparent edge rather than treating an initial observation as established.
- Expect randomness to affect individual strategy results, especially over short periods.
- Develop and manage multiple strategies with attention to their combined portfolio exposures.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.