Spurious Alpha
A toy that lets you bet on a market superstition — the moon, Mondays, prime-numbered dates — watch it beat the market, then watch it lose to doing nothing. Real S&P 500 data, and a lesson in why a rising tide makes any rule look like genius.
Everyone has a theory about what moves the market — the moon, Mondays, Mercury in retrograde, the shape of a chart. Most of them take about ten seconds to test against real prices, and when you test them, something strange happens. They all seem to work. Buy the S&P only on prime-numbered dates and you'd have made money. Buy it only when a coin says so and you'd have made money. Full moons, Tuesdays, dates with a 7 in them — money, money, money.
The tool below lets you try it. Stake some money, pick a superstition, and watch two real years of the market reward it. Then find out what doing nothing would have done instead.
Real SPY total-return prices, 2015–2025. Pick a hunch, run it, then reveal it against buy-and-hold — and test whether it survives five separate two-year windows. Open it full-screen.Why every hunch "works"
There are three ways to fool yourself.
1 — Spurious correlation. Two unrelated things line up by chance. Tyler Vigen has collected hundreds: the number of people who drown in swimming pools each year rises and falls with the number of films Nicolas Cage releases. Nobody thinks one causes the other.
2 — Spurious regression. Two things that both grew over the same years can look linked even when neither touches the other — the S&P 500 and the price of a house both roughly tripled since 2010. To tell a real relationship from a coincidence you compare how they move week to week, not how high each has climbed. Compare the heights, and everything that rose together looks connected.
3 — The missing benchmark. A rising tide floats all boats, and a rising market floats every superstition with it. While stocks climb, any rule that keeps you invested part of the time makes money — the moon, Mondays, a coin flip, they all "work." The only honest way to judge a rule in a trending market is against the one line you never drew: buy-and-hold. Leave that line off the chart, and a guess looks exactly like a forecast.
The third is the whole point of the tool. The original full-moon post shows a moon strategy climbing across a clean stretch of bull market. It's true, and it's empty: over that same window, all thirteen superstitions made money, and 58 of 60 purely random rules did too. The market rose 48%. Nothing invested could have failed.
So the moon strategy that started all this? Run it against a decade and it ranks fifth of thirteen — beaten by prime-numbered dates, and by doing nothing at all.The tool uses real SPY total-return prices from Yahoo Finance, 2015–2025. The headline window is January 2024 – December 2025, when the market rose 48%. That it rose is the point, not a flaw: the illusion only appears while the market is rising, which is why nobody posts moon-phase charts after a bad year. Moon phase is computed from a 29.53-day cycle; no transaction costs, which only makes the hunches look worse.