Using Boxplots to Explore Seasonal Patterns in Financial Returns
Summary
The article describes a way to examine calendar effects in financial returns using boxplots. It converts EURUSD closing prices into percentage changes, groups returns by month, weekday, and hour, and compares their distributions across a historical sample. It explains how medians, quartiles, whiskers, and outliers can help distinguish typical returns from variation and unusual observations.
The examples suggest possible monthly and weekday tendencies, while the hourly distributions show wider movement around European and American session openings than during quieter Pacific hours. The article also discusses using these observations to inform trading rules and examining patterns at finer time scales. These are exploratory results from one instrument and period, not evidence that the patterns persist or will produce profits. The document offers limited statistical validation and warns that combining many cycles makes pattern discovery difficult; seasonal behavior can also change over time.
Key ideas
- Percentage returns are more suitable than raw prices for comparing distributions across time.
- Boxplots summarize the median, quartiles, spread, and unusually large observations for each calendar group.
- Grouping returns by month, weekday, or hour can reveal candidate seasonal effects for further study.
- Intraday variation may reflect differences in activity across trading sessions.
- Patterns found in historical data may change and require validation before use in a trading system.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.