The thesis
Dual-momentum is a trend-following strategy with a simple mandate: hold the strongest trending names by 60-day return, and exit on a trend break. It runs against a 24-name large-cap universe spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT), healthcare (JNJ, UNH), and energy (XOM, CVX). The design bet is classic momentum: let winners run, cut positions when the trend rolls over, and accept a low hit rate in exchange for asymmetric payoffs.
Backtest performance
Over 451 days the strategy returned 23.5%, growing a $10,000 book to $12,349.63 — a 12.52% CAGR. Sharpe came in at 0.95, with a max drawdown of 15.67%. The win rate is a low 28.79% across 136 trades, which is exactly what you'd expect from a momentum system: most trades are small losers, and a minority of large winners carry the return. Turnover is heavy at 2,638%, though fees stayed modest at $136 with no FX cost.
Validation — the strategy does not pass
Here is the important caveat: the strategy failed our validation gate. Across four walk-forward folds, three were positive and one was sharply negative. The out-of-sample return was 13.34% with a strong OOS Sharpe of 2.15, and the probabilistic Sharpe ratio (PSR) is a healthy 0.893. But the deflated Sharpe ratio (DSR) is only 0.476 — once you penalise for the six trials run during development, the edge is no longer statistically convincing. That deflation is the honest reason the gate says no.
Fold 2 (Jan–Jul 2025) is the scar: -7.31% with a -1.05 Sharpe and a 17.15% drawdown. Momentum strategies are vulnerable to choppy, mean-reverting regimes where trends fail to persist, and that fold captured one. By contrast, Fold 3 returned 25.59% at a 3.32 Sharpe — evidence of how regime-dependent this approach is.
Recent activity — sitting in cash
Since going live, the strategy has been conspicuously quiet. The last six scheduled runs (17–24 September) all report 0 executed, 0 rejected, with the full $10,000 in cash. No name is currently clearing the momentum threshold, so the trend filter is keeping it flat. That is arguably the system working as designed — refusing to force trades in the absence of a clear trend — but it also means zero live P&L to date and no real-world confirmation of the backtested edge.
Verdict
Dual-momentum has an intuitive thesis and a solid-looking backtest, but the numbers demand humility. The failed DSR, the double-digit drawdown in Fold 2, and the sub-30% win rate all point to a strategy that is regime-sensitive and prone to painful stretches. Its current all-cash stance is prudent, not a red flag — but until it deploys capital and validates live, treat the 23.5% figure as a hypothesis, not a track record.