The Saturn signal survived. Now we need to explain why.

Yesterday we published the result: Saturn in Taurus is statistically overrepresented among TIME’s 100 Most Influential People of the 20th Century. Bonferroni-corrected p = 0.021 across 393 tests. Nineteen of 100 subjects have it, versus 6.9% in a stratified null distribution.

We also flagged the obvious problem. Saturn takes 29.5 years to orbit the Sun. It passed through Taurus three times in the relevant century: 1881–83, 1910–12, and 1940–42. Maybe those just happened to be productive eras. Maybe the signal is an artifact of history, not astronomy.

Today we ran the tests that would kill it. They didn’t.

The permutation test

The permutation test is the gold standard for this kind of analysis. It doesn’t assume anything about distributions or test statistics. It just asks: if we randomly shuffle the “subject” and “null” labels across all 10,100 charts, keeping the feature matrix fixed, how often does Saturn in Taurus score as well or better than what we actually observed?

We ran 10,000 shuffles.

In zero of them did the randomly-labeled “subjects” have 19 or more charts with Saturn in Taurus. The maximum across all 10,000 permutations was 17. The mean was 7.0 with a standard deviation of 2.5.

Permutation p-value: less than 0.0001.

This is not a marginal result. The observed count is nearly 5 standard deviations from the permutation mean. Whatever is happening, it’s not random noise surviving multiple comparison correction by luck.

But is it Saturn, or is it 1882?

The permutation test confirms the signal is real. It doesn’t tell us what’s causing it. The 19 subjects are born in three clusters that correspond to Saturn’s orbital windows — but those same windows might correspond to historically productive periods for entirely mundane reasons: economic conditions, immigration patterns, demographic booms.

To disentangle Saturn from era, we need to look inside the windows.

Here’s the logic. Saturn takes about 2.5 years to transit Taurus. Our era windows are roughly 5 years wide (1881–85, 1910–14, 1939–42). Within each window, some months have Saturn in Taurus and some don’t. If the signal is just “born in the right era,” subjects should be spread evenly across both types of months. If the signal is specifically about Saturn’s position, subjects should cluster in the Saturn-in-Taurus months.

The month-level data

We classified each month in each window as a “Saturn-in-Taurus month” (more than 50% of null dates that month have Saturn in Taurus) or not.

Window 1881–85. Twenty-four of 54 months are Saturn-in-Taurus months. Eight of 12 subjects born in this window (67%) were born in those months, versus 46% of null dates. Kemal Ataturk, Fleming, Boeing, Picasso, Roosevelt, Joyce, Stravinsky, Goddard — all born while Saturn was between 30° and 60° ecliptic longitude. Keynes, Chanel, Mayer, and Merrill — born in the same window but after Saturn had moved into Gemini and Cancer.

Window 1910–14. Twenty-five of 55 months are Saturn-in-Taurus months. Five of 8 subjects (62%) were born in those months, versus 45% of null dates. Mother Teresa, Reagan, Lucille Ball, Turing, and Rosa Parks on the Taurus side. Shockley, Watson Jr., and Salk on the other.

Window 1939–42. This is the striking one. Twenty-four of 42 months are Saturn-in-Taurus months. Six of 6 subjects (100%) were born in those months, versus 58% of null dates. Lennon, Pelé, Bruce Lee, Dylan, Ali, Aretha Franklin — every single one born while Saturn was in Taurus. Not one subject in this window was born after Saturn left.

Pooled result: 19 of 26 subjects born in Saturn-in-Taurus windows were born specifically in Saturn-in-Taurus months (73%), versus 48% of null dates. Fisher’s exact test: odds ratio 2.98, p = 0.008.

The signal is not just “born in the right era.” It’s “born in the right months within the right era.” The subjects preferentially land in the specific 2.5-year sub-windows where Saturn occupies Taurus, not just the surrounding 5-year periods.

Saturn’s exact position

For the curious — here’s where Saturn actually was for each of the 19 subjects. Taurus spans 30° to 60° ecliptic longitude.

They’re spread across the full 30-degree range. No clustering at a particular degree — just “in Taurus.” That’s consistent with the astrological feature encoding (sign placement) rather than some narrow astronomical coincidence.

The within-window enrichment test

Here’s the test that matters most for the era-artifact question.

We asked: among people born in Saturn-in-Taurus windows (both subjects and null dates), are subjects more likely to actually have Saturn in Taurus?

If the signal is an era artifact, the answer should be no. Within the same era, subjects and null dates should have similar Saturn-in-Taurus rates, because they’re drawing from the same pool of birth dates.

The answer is yes.

Among those born in Saturn-in-Taurus windows: 19 of 29 subjects (66%) have Saturn in Taurus, versus 688 of 1,825 null dates (38%). Fisher’s exact test: odds ratio 3.14, p = 0.002.

The era explains why more subjects are in these windows at all — 29% of subjects versus 18% of null dates are born in Saturn-Taurus eras (p = 0.006). But even after accounting for that era concentration, subjects within those eras are still significantly more likely to have Saturn specifically in Taurus. The era gets you into the window. It doesn’t explain why you land in the Taurus months.

What we’re not claiming

The same caveats from yesterday still hold, and we want to add a new one.

We’re not claiming this is a large sample. The within-window test has 26 subjects. The pooled monthly test has 19. These are real statistical departures, but they’re built on small numbers. A different list of 100 influential people could produce a different result.

We’re not claiming to have eliminated all confounds. Birth-month seasonality within eras — maybe the spring and summer of 1881 were particularly good times to be born for economic or public health reasons that have nothing to do with Saturn. We’ve shown the signal isn’t at the era level, but we haven’t ruled out every possible month-level confounder.

We’re not claiming Saturn causes anything. There is no known physical mechanism. If this pattern is real and generalizable, the explanation is almost certainly demographic or cultural, mediated through some variable we haven’t measured that happens to correlate with Saturn’s position. That would still be interesting — it would mean Saturn’s cycle is a useful proxy for some real-world variable — but it wouldn’t be magic.

We are claiming the pattern survives the tests designed to kill it. Bonferroni correction, Benjamini-Hochberg FDR, permutation testing, and within-window era control. Each of these is a different way of asking “is this noise?” and each one says no.

What’s next

The honest next step is replication. Can a different dataset of influential people reproduce this pattern? Nobel laureates, Fortune 500 founders, heads of state — any curated list where we didn’t pick the names. If Saturn in Taurus shows up in an independent dataset, the era-artifact explanation gets much harder to sustain. If it doesn’t, then it’s a quirk of the TIME 100 and we’ll say so.

The GA is still coming. But we’re not going to feed it the data until we know whether the strongest signal in the dataset is real or an artifact. Garbage in, garbage out — even if the garbage has a p-value of 0.0001.

The methodology, continued

Building on yesterday’s technical section. All new analysis uses the same data (100 subjects, 10,000 null dates, seed 42) and the same feature matrix.

Permutation test. 10,000 label shuffles. For each permutation, randomly assign 100 of the 10,100 charts as “subjects,” compute the number with Saturn in Taurus, record. Permutation p-value = fraction of permutations where count ≥ 19. Result: 0/10,000. Seed 42 for reproducibility.

Within-window analysis. Saturn-in-Taurus orbital windows defined as year ranges where any month has >50% Saturn-in-Taurus prevalence in the null dates: 1881–85, 1910–14, 1939–42. “Saturn-in-Taurus months” defined as months where >50% of null dates have Saturn in Taurus. Within-window Fisher’s exact test conditions on being born in a Saturn-Taurus window, then tests whether subjects are more likely to be born in Saturn-in-Taurus months specifically.

Saturn longitude. Raw longitudes from Swiss Ephemeris, stored in the charts table. Taurus = 30.0° to 59.999° ecliptic longitude. All 19 subjects verified to fall within this range.