There's a ~2.1-year cycle in market volatility. We found it with astrology.

We didn’t find an astrological signal. We found an economic cycle. The astrology was just the detector.

Let us explain.

The story so far

Over the past two days, we’ve run an astrological enrichment experiment across increasingly rigorous tests. The short version: we computed natal charts for 40,000 events across four datasets — S&P 500 extreme trading days, VIX spikes, M6+ earthquakes, and major wars — and tested whether any astrological feature is overrepresented among these events compared to stratified null distributions.

The initial results were dramatic. Dozens of features survived the strictest multiple-comparison corrections. Then we built two controls and peeled apart three layers of signal. The first layer was slow-planet aliasing — Saturn and Jupiter barely move during a financial crisis, so their positions are confounded with temporal clustering. Dead. The second layer was calendar seasonality — the Sun is in Libra during October, the most volatile month. Trivially explained.

The third layer refused to die. Mars in Capricorn is enriched among extreme market days in every market we tested: S&P 500, VIX, Nikkei 225, and FTSE 100. It survives the fast-planet restriction (Mars’s ~2-year cycle is decorrelated from multi-year crisis clustering). It survives block-shuffle permutation testing (which controls for temporal clustering at all timescales). The combined replication p-value across Nikkei and FTSE — two markets we hadn’t previously tested — is 0.000042.

Mars in Capricorn is not an astrological claim. It’s a phase measurement. Mars takes about 780 days to orbit the Sun. “Mars in Capricorn” means “Mars is currently in the 270°–300° segment of its orbit.” That segment corresponds to roughly 6 weeks out of every 2.1 years. The question is not “does Mars cause volatility” — it’s “does market volatility have a ~2.1-year cycle, and if so, is its phase aligned with Mars’s orbit?”

The spectral analysis

We ran a Lomb-Scargle periodogram on realized volatility (21-day rolling standard deviation of daily returns, annualized) for the S&P 500 and VIX.

S&P 500 (76 years of data, 1950–2026):

The third-strongest spectral peak sits at 848 days (2.32 years), with a power-to-median ratio of 26.6×. The only stronger peaks are at 2,410 days (~6.6 years, consistent with the business cycle) and 1,500 days (~4.1 years, consistent with the presidential/election cycle).

VIX (36 years, 1990–2026):

A peak at 803 days (2.20 years), power-to-median ratio of 10.1×. Also a top-6 peak in the periodogram.

Mars’s synodic period is 780 days. The spectral peaks at 803–848 days are close but not exact matches. This is actually more consistent with a real economic cycle than with a direct planetary influence — a planetary mechanism would hit 780 days precisely, while an economic cycle with its own dynamics would have an approximate period that drifts.

The phase structure

If there’s a real ~2.1-year volatility cycle, it should show up as a smooth pattern across Mars’s orbital phase. Mars spends roughly equal time in each zodiac sign (~6 weeks per sign), so the 12 signs act as 12 evenly-spaced phase bins around the orbit.

We computed mean absolute daily return for each Mars sign, across all trading days:

S&P 500 mean |return| by Mars sign (relative to overall mean):

VIX mean daily |change| by Mars sign (relative to overall mean):

Capricorn is the highest-volatility phase in both markets. A sinusoidal fit to the 12-sign pattern peaks near Sagittarius-Capricorn and troughs near Taurus-Virgo — consistent with a smooth cycle, not a spike at one sign.

The Kruskal-Wallis test confirms the 12 Mars-sign groups have significantly different volatility distributions: p = 0.000015 for S&P 500, p = 0.002 for VIX. This is not chance variation.

The international replication

The critical test was whether the Mars-in-Capricorn signal appears in markets we had never analyzed. We downloaded daily data for the Nikkei 225 and FTSE 100, computed Mars signs for every trading day, and ran the same block-shuffle permutation test.

Mars in Capricorn is enriched among extreme days in all four markets. The combined replication p-value across the two independent markets (Nikkei and FTSE) using Fisher’s method is 0.000042. Mars in Capricorn is the only sign consistently elevated across all four markets.

The out-of-sample caveat

An out-of-sample temporal split on S&P 500 told a more complicated story. The Mars-in-Capricorn effect is absent in 1950–1999 (OR = 0.99, completely null) and only appears in 2000–2026. The VIX shows signal in both halves but with a decaying effect size.

This is actually consistent with the hidden-cycle interpretation. If the ~2.1-year cycle strengthened or shifted phase in the post-2000 era of highly correlated global markets, it would appear absent in pre-2000 data but present across all modern international markets simultaneously. The fact that it replicates across the Nikkei and FTSE — both predominantly post-2000 datasets — supports this interpretation.

It could also mean the signal is a post-2000 artifact that will eventually fade. Only time will tell.

What this is not

We want to be very explicit about what we are not claiming.

We are not claiming Mars influences financial markets. There is no known physical mechanism by which a planet 100+ million kilometers away affects stock prices. The prior probability is effectively zero, and nothing in this analysis changes that.

We are not claiming astrology works. The zodiac is a coordinate system. “Mars in Capricorn” is a phase label for a ~780-day orbital cycle. The astrological encoding happened to discretize this cycle into testable bins. A Fourier decomposition would find the same thing.

What we are claiming

There is a spectral peak in market volatility near 800–850 days (~2.1–2.3 years). This is visible in the Lomb-Scargle periodogram of both S&P 500 and VIX realized volatility. It’s the third-strongest cycle in 76 years of S&P data, after the business cycle and the presidential cycle.

This cycle’s phase, as measured by Mars’s orbital position, correlates with extreme market days across four independent markets. Volatility peaks when Mars is in the Sagittarius-Capricorn-Aquarius segment of its orbit and troughs when it’s in the Taurus-Virgo segment.

The ~2.1-year period is not a well-studied cycle in financial econometrics. It sits between the annual cycle and the Kitchin inventory cycle (~3.3 years). It may correspond to:

• A sub-harmonic of the Kitchin cycle
• A capital expenditure or corporate earnings cycle
• A monetary policy lag effect (central bank actions take ~18–24 months to affect the real economy, which is close to this period)
• Something we haven’t thought of

The astrological encoding was the wrong tool that found the right thing. If we had started with spectral analysis of volatility, we would have found the ~2.1-year peak directly and never mentioned Mars. Instead, we started with astrology, spent days eliminating artifacts, and arrived at the same place through a much stranger path.

The methodology built for this experiment — stratified null distributions, multiple-comparison correction, permutation testing, fast-planet restriction, block-shuffle null models — was designed to be honest. It killed the Saturn-in-Taurus birth-date hypothesis. It killed the slow-planet aliasing artifact. It correctly identified the September-October seasonal effect. And it surfaced a spectral feature in market volatility that, as far as we can tell, hasn’t been widely reported.

That last part is worth investigating with proper tools. The zodiac was the metal detector. Now it’s time to dig.

The methodology

Spectral analysis. Lomb-Scargle periodogram (scipy.signal.lombscargle) on 21-day rolling realized volatility. Frequency grid: 5,000 points from period 60 days to 3,000 days. Peaks identified via scipy.signal.argrelmax with order=50 (minimum 50 frequency bins between peaks).

Mars phase analysis. Mean absolute daily return grouped by Mars zodiac sign (12 bins of 30° ecliptic longitude each). Mars sign computed via Swiss Ephemeris (Kerykeion) for each trading day. Kruskal-Wallis H-test for group differences.

Cross-market comparison. Volatility ratios (mean |return| in sign / overall mean |return|) computed independently per market, then compared. Capricorn ranked #1 across both S&P 500 and VIX. Nikkei and FTSE confirmed via Fisher’s exact test and block-shuffle permutation on extreme days.

Out-of-sample split. S&P 500 split at year 2000 (training: 1950–1999, test: 2000–2026). VIX split at 2005 (training: 1990–2004, test: 2005–2026). Mars-in-Capricorn Fisher’s exact test and block-shuffle permutation run independently on each half.