The thesis
donchian-breakout runs one of the oldest ideas in systematic trading: buy when price makes a new 20-day high, exit when it makes a new 20-day low. It trades a 24-name large-cap universe spanning tech, financials, healthcare, staples and energy (AAPL, MSFT, NVDA, JPM, V, JNJ, XOM, CAT and others). The bet is simple and honest — ride persistent trends, cut the rest — and the strategy makes no pretense of predicting direction.
Backtest profile
Over 451 days the strategy returned 6.95% (final equity $10,695), a 3.83% CAGR. That headline hides a rough ride: a 21.73% max drawdown against a single-digit return is a poor pain-to-gain ratio, and the 0.34 Sharpe confirms the returns are not well paid for their volatility. The 38.5% win rate across 108 trades is entirely expected for a breakout system — trend-followers lose often and small, then win occasionally and large. More concerning is turnover of 2,083%: the book churns hard, and while fees were only $108 with no FX cost here, that level of trading leaks performance in any live setting with real spreads.
Validation: did not pass
The strategy failed our walk-forward gate, and it's worth being candid about why. Three of four folds were positive, and folds 3 and 4 were genuinely strong (+14.09% at Sharpe 2.72, then +11.35% at Sharpe 1.86). But fold 2 (Jan–Jul 2025) lost 7.46% at a Sharpe of −1.07 — a chop-driven whipsaw regime that breakout systems hate. The deciding numbers are the risk-adjusted ones: a deflated Sharpe ratio (DSR) of 0.198 across 6 trials and a probabilistic Sharpe of 0.674 are not enough confidence that the edge is real rather than a lucky back half. The out-of-sample Sharpe of 1.86 looks encouraging in isolation, but the DSR exists precisely to discount that.
Recent live activity
Paper trading tells a quieter story. The last six scheduled runs executed almost nothing — typically 0 executed, 1–2 rejected — with cash pinned at $813.35 and total equity drifting between roughly $9,792 and $10,093, i.e. around or slightly below a $10k start. The rejections alongside flat cash strongly suggest the book is fully invested and unable to fund new breakout signals. Earlier fills (late Sept) show the system behaving as designed: buying NVDA and CAT into strength, rotating out of KO, MA and V.
Verdict
donchian-breakout is a clean, interpretable trend-follower whose better folds hint at a real regime-dependent edge. But a failed validation gate, a 0.34 full-period Sharpe, a 22% drawdown and cash-starved live runs mean it should stay on a short leash — a candidate for sizing and cash-management fixes before it earns any conviction.