How much smaller our misses have been than a naive forecast (last published value) and than consensus. Above zero is better. Colored only when the 90% range excludes zero.
Full record and tests →The Lead · Research Insight
Tested every possible starting day in the S&P 500 since 1927: invest a lump sum all at once, or spread it out over a fixed window. Investing all at once wins about 6 times out of 10 over a year, more often over longer windows, but spreading it out gives a measurably smoother ride, and wins more often right after the market has run up a long way. Includes an interactive calculator.
Insight
Running the same statistical tests across 502 stocks, 5 indices, and 10 Treasury/credit series confirms fat tails everywhere, but also surfaces a quieter finding: the most "extreme" stocks by the numbers usually aren't…
Insight
Daily S&P 500 returns since 1927 are far from normally distributed, with heavier tails, negative skew, and a return profile shaped disproportionately by a handful of extreme days. What that means for anyone modeling…