Who this is for — Calendar-linked statistical biases (e.g. «sell in May», weak Mondays, December rally). Useful as context, not a standalone strategy.
Market seasonality describes recurring tendencies tied to month, weekday, holidays, tax deadlines or predictable institutional flows. It is slow cyclicity — annual or intraday period — distinct from Hurst cycles measured bar by bar.
In plain terms — «At this time of year the market, on average, behaves like this.» Historical mean, not a guarantee for any single year.
Common types
| Scale | Examples |
|---|---|
| Annual | Year-end, January, summer low liquidity |
| Monthly | Turn of the month, op-ex week |
| Weekly | Monday/Friday, mid-week |
| Intraday | US open, FX fix, crypto funding |
Seasonality overlaps market regime, macro and sessions — a calendar pattern can vanish for years when context shifts.
Operational use
- Filter: avoid setups against strong seasonal bias on a matching horizon
- Size: reduce exposure in historically weak windows
- Research: test on long samples; separate real effect from data mining
Do not confuse with Hurst cyclic period: seasonality is anchored to solar/work calendar, not troughs measured on chart.
Common mistake — Trading only because «it's November and it usually rises» — a bear year can override the average.
Example — S&P: last five December days often positive (window dressing); in 2008 macro context dominated seasonality.
Card
- Data: historical means per calendar bucket.
- Role: context, not trigger.
- Validate: large sample + current regime.