Strategy Overview
Donchian-breakout follows one of systematic trading's most durable principles: buy strength, cut weakness. The strategy enters when a symbol from its 24-stock large-cap universe clears a new 20-day high and exits when price falls to a 20-day low. No sentiment overlay, no macro filter — pure price action. Running live, it has completed 451 days of tracked history across tech, healthcare, financials, consumer staples, energy, and industrials.
Recent Activity
The strategy has been quiet over the past week. Scheduled runs from June 17 through June 24 each returned zero executions, with two to three signals rejected per session. The portfolio sits at roughly $9,494 with $1,761 in cash — the engine is holding existing positions and waiting for clean breakout setups.
The last meaningful activity came in early June. The strategy opened ABBV at $218.75, CAT at $941.58, AAPL at $314.77, and MSFT at $450.24 across a five-day window. Then on June 9–10, it rotated: MSFT was exited at $404.21 (below the May entry) and CAT at $857.78 (also below entry), while UNH at $412.62 and KO at $83.69 were added. Those quick MSFT and CAT exits within days of entry illustrate the strategy's mechanical discipline — when price hits the 20-day low channel, it sells without hesitation, regardless of the holding period.
Backtest & Validation
Across 108 trades, donchian-breakout returned 6.95% (3.83% annualised) with a Sharpe ratio of 0.34. The 38.46% win rate is consistent with trend-following orthodoxy — this style profits from a small number of large winners, not frequent wins. High turnover (2,083%) underscores how actively the engine cycles through the universe.
The max drawdown of 21.73% is the primary concern. A strategy returning under 7% cumulative while sustaining drawdowns above 21% implies a reward-to-risk profile that has not yet proven itself over a sufficiently long horizon.
The four-fold walk-forward breakdown is instructive. Fold 2 (January–July 2025) was the worst stretch: −7.46% at a Sharpe of −1.07 and the highest trade count of any fold (38), suggesting a choppy environment that generated repeated false breakouts. Fold 3 (July–December 2025) then delivered 14.09% at a Sharpe of 2.72, and the out-of-sample Fold 4 followed with 11.35% at a Sharpe of 1.86. The divergence between folds is wide, which is both encouraging (improving) and cautionary (volatile regime sensitivity).
Despite the strong out-of-sample numbers, the strategy did not pass formal validation. The Probabilistic Sharpe Ratio (PSR) sits at 0.674 and the Deflated Sharpe Ratio (DSR) at 0.198 — with six trials and a short track record, the observed Sharpe cannot yet be distinguished from noise with high statistical confidence.
Strengths and Risks
Strengths: The trend-following thesis has deep empirical support. Three of four walk-forward folds are profitable, and the two most recent folds show materially better risk-adjusted returns, suggesting the strategy may be finding its regime. Mechanical execution is consistent — the engine respects channel exits without discretion.
Risks: The 21.73% drawdown is large relative to the cumulative return. The DSR remains well below conventional thresholds, and the short backtest window (under 16 months) limits statistical confidence. In a live funded account, the high turnover rate could amplify friction costs beyond the flat $1-per-trade simulation assumption. A single choppy period — as Fold 2 demonstrated — is enough to erase a full fold's gains and then some.