S&P 500 Strategy Study
It did not reliably beat the S&P 500. The main version returned 9.6% a year against 9.2% for the index, a difference too small to tell from luck, and the data flatters it. The trend overlay is the useful part: it gave up return but cut the worst fall from -51% to -16%.
| Strategy | Yearly return | Volatility | Worst fall | Extra return vs SPY | t-statistic | Beat SPY over 5 years |
|---|---|---|---|---|---|---|
| S&P 500 (SPY) | 9.2% | 15.3% | -50.8% | — | — | — |
| 75% SPY + 25% stock picks | 9.6% | 14.2% | -48.8% | +0.17% | 0.37 | 49% of the time |
| Same, with the trend overlay | 8.8% | 10.1% | -16.3% | -1.06% | -0.50 | 42% of the time |
| Stock picks only | 10.3% | 13.4% | -42.9% | +0.73% | 0.39 | 45% of the time |
Extra return is the average monthly gap to SPY, annualised. A t-statistic of about 2 or more is the usual bar for a result that is unlikely to be luck; none of these reach it.
| Strategy | 1998–2006 | 2007–2015 | 2016–2026 |
|---|---|---|---|
| S&P 500 (SPY) | 5.7% / -45% | 6.3% / -51% | 14.7% / -24% |
| 75% SPY + 25% stock picks | 7.8% / -34% | 6.6% / -49% | 13.7% / -22% |
| Same, with the trend overlay | 8.7% / -14% | 8.6% / -16% | 9.0% / -16% |
| Stock picks only | 13.7% / -12% | 7.0% / -43% | 10.4% / -21% |
Each cell shows yearly return, then worst fall. The advantage came before 2016. Since then the index has led.
| Rule | Yearly return | Worst fall |
|---|---|---|
| Buy and hold | 9.1% | -82% |
| Trend rule | 7.4% to 9.2% | -52% to -47% |
| Volatility rule | 8.8% to 9.0% | -67% to -66% |
| Trend and volatility | 7.2% to 9.1% | -38% to -34% |
Monthly S&P data (Robert Shiller's series). Ranges show cash earning 0% at one end and the 10-year Treasury yield at the other, because this history has no T-bill rate. The same pattern appears: smaller falls, no reliable extra return.
Results built 2026-11-02 from saved model runs. Daily prices from Yahoo Finance; index membership from github.com/fja05680/sp500. Educational research, not investment advice.