Thesis & Mechanics
Donchian breakout is one of the oldest systematic trend-following rules: enter long when price closes at a 20-day high, exit when it closes at a 20-day low. No forecasting, no sentiment — just price structure. The strategy runs across a 24-name universe of U.S. large-caps spanning technology, financials, healthcare, consumer staples, energy, and industrials, which provides diversification without drifting into illiquid territory.
The simplicity is a feature. There are no free parameters to overfit beyond the single 20-day window, and the rule is fully mechanical, making it straightforward to audit and explain.
Backtest Snapshot
Over the 451-day backtest period, the strategy returned +6.95% on a $10,000 notional, finishing at $10,695. Annualised, that's a 3.83% CAGR — below the S&P 500's typical pace, but not the primary measure for a trend system designed to participate in directional moves while cutting losses quickly.
The concerning figures are the Sharpe ratio (0.34) and max drawdown (−21.73%). A Sharpe below 0.5 means the strategy earns very little risk-adjusted return, and a drawdown exceeding 21% can be psychologically punishing even in a paper-trading context. The win rate of 38.46% across 108 trades is expected for a trend-follower — these systems are designed to lose small and win large — but it means extended cold streaks are baked in.
Turnover deserves a flag: at 2,083% annualised, the strategy churns through its book roughly 20× per year. In a live account with realistic commissions, this would materially erode returns.
Walk-Forward Validation
The four-fold walk-forward tells a more nuanced story:
| Fold | Period | Return | Sharpe | Max DD |
|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +0.73% | 0.21 | −6.71% |
| 2 | Jan 2025 – Jul 2025 | −7.46% | −1.07 | −15.58% |
| 3 | Jul 2025 – Dec 2025 | +14.09% | 2.72 | −3.14% |
| 4 | Dec 2025 – May 2026 | +11.35% | 1.86 | −6.01% |
Three of four folds are positive, and the two most recent — which are the closest proxy for current market conditions — are genuinely strong. Fold 3's Sharpe of 2.72 and fold 4's 1.86 suggest the strategy found its footing in a trending environment. The out-of-sample return of +11.35% with a Sharpe of 1.86 is the headline number bulls will point to.
However, the validation gate did not pass. The Probabilistic Sharpe Ratio (PSR) of 0.674 means there is roughly a 1-in-3 chance the observed edge is noise. More critically, the Deflated Sharpe Ratio (DSR) of 0.198 — which penalises for the six parameter trials run — drops that confidence sharply. The strategy has not yet demonstrated the statistical robustness required for autonomous capital deployment under Headmars' approval framework.
Live Activity
The strategy deployed on May 31, immediately buying 5 shares of MSFT at $450.24. On June 2 it added 7 shares of AAPL at $314.77. As of the most recent run, the portfolio sits at $9,951.90 (cash $5,545.41), a shallow −0.5% from inception — well within normal noise for a two-day-old trend system.
Balanced Verdict
Strengths: Classic, non-overfit rule; recent momentum is real; large-cap universe limits liquidity risk; improving fold trajectory.
Risks: Below-threshold DSR; high turnover would punish live execution; fold 2's −7.5% drawdown shows the strategy struggles in choppy, mean-reverting regimes. More live history across a full market cycle is needed before treating these results as durable.