Free lesson · Technical analysis

Moon phases: encode a calendar hypothesis and try to falsify it

Open interactive lessonPractice calculationsExplore labs

• Study lunar timing as an empirical hypothesis, not a price law. Compare it with ordinary calendar and market controls.
• A cycle wraps around: the end of one cycle is near the beginning of the next. Sine and cosine encode that continuity.
• The mean synodic month is about 29.53059 days. A constant-period approximation is not a precise astronomical ephemeris.
• For an actual study, obtain validated phase-event timestamps, specify the reference timezone and map them to the tradable session calendar.
• Predeclare assets, phase windows, horizon and costs. Control weekday, month-end, trend and volatility effects where justified.
• Use blocked or circular-shift placebos that preserve relevant time dependence. Correct for testing many cycles, windows and markets.
• Report null and inconclusive results. A visually appealing phase overlay does not justify changing exposure.

Use the rule

• Fix the definition, units and information timestamp.
• Compute the example and inspect the graph.
• Compare with a simple baseline on untouched periods after costs.

Before moving on

Write the exact rule, its availability time, an invalidation condition and a fair out-of-sample test.

Research checkpoint · reviewed 12 September 2026

• Status: Economics and Environment 4(95), 2025. Full-text methods, results and limitations reviewed.
• Sample: eight Central/Eastern European equity indexes with complete data, January 2020–July 2024; 56 new-moon and 57 full-moon events. These correlated markets do not provide eight independent lunar calendars.
• The study compares phase-event log returns with zero. That is not the same test as incremental returns over ordinary trading days, calendar controls and a matched strategy after costs.
• Use a prospective, predeclared phase test with dependence-aware uncertainty and placebo timings. Findings from a selected high-volatility period do not establish a universal lunar edge.

Lisicki · Moon phases effect in the time of increased volatility (2025) ↗

Research sources, review dates and limitations

Extend the research question

Treat a named pattern as a hypothesis. Record every parameter tried and evaluate the full selection process after costs on a new time block.

Continue with the connected research module →

from math import sin,cos,pi

def lunar_features(age,period=29.53059):
    if period<=0: raise ValueError("Positive period required")
    angle=2*pi*(age%period)/period
    return sin(angle),cos(angle)

print(lunar_features(29.53059/4))