The thesis
Dual-momentum is one of the oldest ideas in systematic investing dressed in a simple rule: hold the strongest trending names by 60-day return, and exit when the trend breaks. The universe is a conservative slate of 24 U.S. large caps — mega-cap tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT) and healthcare (JNJ, UNH). It is a momentum engine bolted to a blue-chip chassis, which limits blow-up risk while still chasing the leaders.
Backtest scorecard
Over 451 days the backtest turned a notional $10,000 into $12,349.63 — a 23.5% total return, or roughly 12.5% CAGR. The Sharpe of 0.95 is respectable but not exceptional, and the 15.67% max drawdown is meaningful for a portfolio that only touches large caps.
Two numbers deserve attention. The win rate is just 28.79% across 136 trades — normal for trend-following, where a handful of large winners carry many small losers, but it demands discipline to stomach. And turnover is a striking 2,638%, meaning the book churns its capital roughly 26 times over the period. Fees stayed modest ($136 total, no FX cost), but that churn is a persistent drag and a fragility signal.
Validation: it did not pass
This is where the story turns cautious. The walk-forward audit failed. Across four folds, three were positive but one — the Jan–Jul 2025 stretch — lost 7.31% with a −1.05 Sharpe and a 17.15% drawdown. The out-of-sample return was 13.34%, healthy on its own, but the aggregate statistics tell the real tale: the Deflated Sharpe Ratio came in at 0.476, below the confidence threshold once you account for the 6 trials run. The PSR of 0.893 is encouraging, yet the DSR — which penalises selection across multiple attempts — says we cannot rule out that the edge is a product of the search itself.
The fold dispersion reinforces this. Fold 3 posted a 3.32 Sharpe; fold 2 was outright negative. That is a wide spread for a strategy meant to be regime-robust, and it suggests performance is heavily dependent on trend-friendly windows.
Live activity: an empty book
Most telling is what the strategy has not done. The last six scheduled runs (2026-09-21 through 2026-09-28) each executed zero trades, leaving the account fully in cash at $10,000. Under its own rules that is defensible — if nothing is trending, dual-momentum should hold cash rather than force a position — but a week of complete inactivity means the live track record is currently generating no signal to evaluate.
Verdict
Dual-momentum has an honest, well-understood thesis and a headline return that looks good in isolation. But the failed validation and low DSR argue against treating the 23.5% as reliable forward expectation, and the recent flat-cash stretch means the live book has yet to prove itself. Watch it — do not yet trust it. The right next step is more out-of-sample data and a hard look at whether the high turnover is buying anything the fees don't take back.