Lunar Phases & Market — ^DJI

SeasonAlpha's moon-phase analysis investigates an old market question: do returns differ by lunar phase — are the days around the full moon systematically weaker than those around the new moon? The tool computes the average return path within a freely chosen event window around each lunar date and displays it as a curve, KPI row and heatmaps. You can set any ticker, any period and the window width (days before and after the event) yourself, and you see immediately whether a measurable difference exists — or not.

Here is how to read the new-moon-versus-full-moon comparison: pick a phase in the sidebar and a window, for example t−5 to t+5. The main curve shows the normalized path around the lunar date, where t0 = 0 percent is the event day itself. The KPI row summarizes win rate, average return, median and the sample size (number of events, n). What matters is the comparison of both phases and the sample: if the win rate and average return of full and new moon sit close together, there is no meaningful effect. The heatmaps additionally reveal whether an apparent pattern holds across calendar months or only shows up in individual months.

On methodology: lunar dates are derived astronomically from the synodic cycle of 29.53 days (accuracy roughly one day), anchored to a known new moon; the full moon falls half a cycle later. For each date the tool measures normalized returns in the event window: t0 is set to 0 percent, and the cumulative log returns before and after form the average curve across all events. If a lunar date falls on a weekend or holiday, the next trading day applies. As an extra, the tool classifies supermoons (full moons near the closest point of the lunar orbit to Earth) and maps events onto a lunar calendar (lunar month 1–12) — here too the same caveat holds: small samples per cell, so interpret with caution.

An honest assessment: the moon-phase effect is one of the weakest and most disputed calendar effects. Individual studies found marginally lower returns around the full moon in the past, but these differences are small, frequently not statistically significant and unstable across markets and periods — fading once they become known is typical of such anomalies. There is also no plausible economic mechanism that would explain a reliable moon-market link. Treat the results as a statistical curiosity for context, not as a trading signal. This page is not investment advice; past patterns do not guarantee future results.

Frequently asked questions

Is there a real moon-phase effect in the stock market? The evidence is weak and disputed. Some academic studies found slightly lower average returns around the full moon than around the new moon in the past, but the differences are small, often not statistically significant and unstable across different markets and periods. The tool makes this weakness visible: the win rate and average return of both phases usually sit very close together. No reliable trading signal can be derived from it.

How do you read the new-moon-versus-full-moon comparison in the tool? Pick a phase in the sidebar (full moon, new moon or both) and an event window, for example five days before to five days after the event. The curve shows the normalized return path around the lunar date (t0 = 0 percent), and the KPI row summarizes win rate, average return, median and sample size. What matters is the number of events (n) and the comparison of both phases: if the values sit close together, there is no meaningful effect.

How are the moon phases and returns computed? Lunar dates are derived astronomically from the synodic cycle of 29.53 days (accuracy roughly one day), anchored to a known new moon. For each lunar date the normalized returns in the event window are measured: t0 is set to 0 percent, and the cumulative log returns before and after form the average curve across all events. If a lunar date falls on a weekend or holiday, the next trading day applies.