Calculated based on data from CRSP US Stock Database, © Center for Research in Security Prices (CRSP), The University of Chicago Booth School of Business, and from Compustat, © S&P Global Market Intelligence.
U.S. common stocks at or above the NYSE 20th size percentile, eleven published factors, monthly, ten basis points of traded value per trade. The last five years were set aside before any result was read. The configuration was chosen by a rule fixed in advance, committed, and then run on the holdout once.
| In sample | Holdout |
|---|
| Period | August 1969 to December 2020 | January 2021 to December 2025 |
| Net Sharpe ratio | 1.50 | −0.37 |
| 95% bootstrap interval | 1.12 to 1.90 | −1.20 to 0.40 |
| Annual return | 20.0% | −5.4% |
| Maximum drawdown | −41% | −44% |
| Alpha against Fama–French five factors and momentum, per month | 1.45% (t 7.1) | −0.34% (t −0.7) |
It lost money out of sample, and the two intervals do not overlap.
In sample, the deflated Sharpe ratio was 1.000 and the probability of backtest overfitting 0.071. Both were right about what they measure: the result was not an accident of picking one of six methods. Neither measures whether the backtest's assumptions hold. The in-sample numbers were flagged as too good before the holdout was run: turnover above 100% a month at a flat ten basis points, one-month reversal as a main input, no market impact, and no borrow fee. Why the result changed sign was not investigated, because any further analysis of those five years would be a second look.