Thesis
Donchian Breakout is a textbook trend-following system: buy the 20-day high, exit on the 20-day low. It runs across a 24-name universe of US large caps spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), healthcare (UNH, ABBV) and staples (KO, PG, WMT). The premise is simple and well-documented — let winners run, cut losers at a mechanical stop — and it makes no forecast beyond "price that breaks range tends to keep going."
Recent activity
The live book is currently idle. The last executed fills were on 2026-08-13 (bought 5 V at $362.35, sold 3 UNH at $399.25). Since then, every scheduled run from 2026-09-02 through 2026-09-09 reports 0 executed, 1–2 rejected — signals are firing but not clearing. Cash has sat unchanged at $1,510.88 while total equity drifted from $10,084.48 down to $9,823.99, dipping below the $10k starting mark. In a trend follower this stall is expected during rangebound tape: no new 20-day highs, nothing to buy; no 20-day lows, nothing to sell.
Backtest
Over 451 days the strategy returned 6.95% (final equity $10,695), a 3.83% CAGR. The character is classic trend following: a 38.46% win rate across 108 trades, meaning most positions lose small and a few winners carry the book. Sharpe is a modest 0.34, and the 21.73% max drawdown is the headline risk — you must stomach a fifth of capital evaporating to collect that return. Turnover of ~2,083% signals heavy churn, though fees ($108 total) stayed contained.
Validation
This is where the picture splits. Walk-forward testing across 4 folds was 3-of-4 positive, and the out-of-sample slice returned 11.35% at a 1.86 Sharpe — genuinely strong. Fold 3 (+14.09%, Sharpe 2.72) and Fold 4 (+11.35%, Sharpe 1.86) were excellent; Fold 1 was flat (+0.73%); Fold 2 was ugly (-7.46%, Sharpe -1.07), confirming the strategy bleeds in choppy, trendless regimes.
Despite that, the gate reads passed: false. The Probabilistic Sharpe Ratio is a middling 0.674, but the Deflated Sharpe Ratio — which penalises for the 6 trials run — collapses to 0.198. In plain terms: once we account for how many variants were tried, we can't be confident the edge is real rather than lucky.
Verdict
Donchian Breakout is a coherent, honest strategy with real out-of-sample promise, but two flags keep it on probation: a deep drawdown profile and a deflated Sharpe that fails our multiple-testing check. The recent run of rejected orders is worth watching — if signals keep firing without filling, the live results will tell us little either way. For now: promising, not proven.