marketcorollary

Skeptical market research · refreshed from live data

Do market signals actually predict returns?

Two questions, one method. Do external events predict markets? No. Do the academic risk premia? Yes — smaller than advertised, and painfully. Every number here is computed from public data and updated automatically.

Explore the factors See the folklore battery
Factors clearing t > 3
9 of 12
vs 1 of 27 folk signals — and that one is untradeable Santa
Folk signals that survive FDR
1 of 35
turn-of-year — untradeable
Best factor Sharpe
0.81
Momentum within small caps
Data coverage
2026-05
factors from 1949-01 · daily prices to 2026-07-29

Two studies

The same permutation-tested, false-discovery-corrected machinery, pointed at the believable and the absurd.

9/12 significant

The Factor Reference

Twelve equity risk premia — market, size, value, profitability, investment, momentum and more — scored on return, risk, cost drag, post-publication decay and out-of-sample survival.

Open the explorer →

 1/35 survives

The Almanac of Everything Else

The moon, Mercury (and Venus, and Mars) retrograde, eclipses, supermoons, great comets, geomagnetic storms, the groundhog, Brood X cicadas, papal conclaves, the Super Bowl, the skyscraper curse — 35 market superstitions, re-tested from raw prices through the present.

See what survives →

Questions this answers

Short, cited answers. The detail — and the code — sits behind each study.

Does the Super Bowl Indicator predict the stock market?
No. Scored consistently, it is right 53% of the time — worse than the 67% base rate you get by always guessing the market rises. It fails outright out-of-sample (32% since 1998).
Do the moon, Mercury retrograde or eclipses move stock prices?
No. Across 98 years of S&P and 141 years of Dow data, none of the celestial indicators — moon phases, three planets' retrogrades, solar and lunar eclipses, supermoons, great comets, geomagnetic storms, sunspots — are statistically significant, and several point the opposite way from the folklore.
Which market signal actually survives rigorous testing?
Of 35 signals in the battery, only turn-of-year (Santa Claus) seasonality survives false-discovery correction — and it is calendar-internal and untradeable, not an external predictor.
Do academic investment factors still work after publication?
Mostly weaker. 9 of 12 factors clear the t>3 significance bar, but most premia shrank after their founding paper. Small-cap value and momentum held up best; the size premium is effectively dead.
Is the US stock market overvalued right now?
By the cyclically-adjusted P/E (CAPE), yes on a historical basis: CAPE is 31 versus a long-run median of 20 (90th percentile).
Can market timing beat buy-and-hold?
Not for return. The 200-day moving average roughly matches buy-and-hold on return but cuts the worst drawdown dramatically. Timing rules buy risk reduction, not extra return.