What the Strategy Does
Dual-momentum is a relative-strength system: at each rebalance it ranks a 24-stock large-cap universe by 60-day return and holds the strongest trending names, exiting any position whose trend breaks. The universe spans mega-cap tech (AAPL, MSFT, NVDA, GOOGL), financials (JPM, V, MA), healthcare (UNH, ABBV), consumer staples (WMT, KO, PG), and energy (XOM, CVX) — a diversified but clearly US-equity-only sandbox.
The thesis is not novel; dual-momentum has decades of academic literature behind it. Its value in this context is as a baseline: a well-understood, hand-written rule that the strategy-lab validation stack should be able to score sensibly.
Backtest Snapshot
Over 451 days of in-sample history the strategy compounded to +23.5% total return (12.52% annualised CAGR) against a 15.67% maximum drawdown and a Sharpe of 0.95. Those are respectable headline numbers for a momentum sleeve — positive risk-adjusted returns with drawdown contained under 16%.
The win rate of 28.8% on 136 trades is a tell-tale momentum signature: most individual exits are losers, but the winners run far enough to produce net positive expectancy. High turnover (2,638%) confirms the strategy is actively rotating through its universe rather than buy-and-hold.
Walk-Forward Validation: Where It Gets Complicated
The four-fold walk-forward tells a more textured story:
| Fold | Period | Return | Sharpe | Max DD |
|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +5.85% | 1.23 | 3.92% |
| 2 | Jan 2025 – Jul 2025 | −7.31% | −1.05 | 17.15% |
| 3 | Jul 2025 – Dec 2025 | +25.59% | 3.32 | 4.04% |
| 4 | Dec 2025 – May 2026 | +13.34% | 2.15 | 7.43% |
Three of four folds are positive, and the most recent out-of-sample window (Fold 4, OOS return +13.34%, OOS Sharpe 2.15) is genuinely encouraging. But Fold 2's 17.15% drawdown and negative Sharpe illustrate that the strategy has a real regime vulnerability — the early-2025 environment, likely characterised by choppy or mean-reverting price action, was hostile to trend-following.
The Validation Verdict: Not Yet
The platform has flagged this strategy as validation: failed. Two statistics drive that call:
- PSR 0.893 — the Probabilistic Sharpe Ratio asks whether the observed Sharpe is statistically greater than a benchmark (usually 0 or 1). At 0.893 the answer is probably yes, but with meaningful uncertainty.
- DSR 0.476 — the Deflated Sharpe Ratio goes further, penalising for multiple trials (6 in this case). With a DSR below 0.5 the evidence that this Sharpe survives trial-count inflation is weak. That is the primary failure gate.
In plain terms: with six parameter combinations explored and a relatively short live history, there is not yet enough statistical evidence to distinguish genuine edge from a well-fitted curve.
Deployment Status
Dual-momentum went live on 2026-05-31 with a $10,000 paper allocation. The first two scheduled runs (May 31 and June 1) executed zero trades — the strategy found no positions meeting its entry criteria, leaving the full allocation in cash. This is expected behaviour for a rotational system; it will deploy into equities when its momentum filter fires.
Bottom Line
Dual-momentum is a competent, readable baseline with legitimate historical returns. The failed DSR gate is the right call: six trials and 451 days of data is not a long enough runway to be confident the edge is real rather than mined. The playbook here is to let it run live in paper mode, accumulate additional out-of-sample folds, and revisit validation once the DSR crosses the 0.5 threshold. If the Fold 4 quality (Sharpe > 2) persists in live trading, that case becomes considerably stronger.