Thesis
donchian-breakout is a textbook trend follower: buy when price makes a new 20-day high, exit when it prints a new 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, COST). The premise is simple and time-tested — ride sustained moves, cut the rest — and the trade log reflects it: momentum entries in AAPL, CAT, and MSFT alongside stop-driven exits.
Backtest & Validation
Over 451 days the strategy returned 6.95% (final equity $10,695), a 3.83% CAGR with a Sharpe of just 0.34 and a steep 21.73% max drawdown. Win rate is 38.46% across 108 trades — low, but normal for breakout systems that lose small often and win big occasionally. Turnover is heavy at ~2,083%, so execution costs matter.
The walk-forward picture is more nuanced. Three of four folds were positive, and the two most recent are genuinely strong: fold 3 returned 14.09% (Sharpe 2.72) and fold 4 11.35% (Sharpe 1.86, out-of-sample). The blemish is fold 2, a -7.46% stretch (Sharpe -1.07) through early-to-mid 2025 — the choppy, trendless conditions that whipsaw every breakout model.
Crucially, our validation gate did not pass. Despite the encouraging recent folds, the deflated Sharpe ratio (DSR 0.198) is weak once you account for the 6 trials searched, even with a probabilistic Sharpe of 0.674. In plain terms: the out-of-sample results look good, but not decisively enough to rule out luck given how many variants were tried. Trend following's edge is real but lumpy, and this sample is too short to be confident.
Recent Activity
Here's the flag worth watching. The strategy is marked live, but the last week of scheduled runs (Jul 31 – Aug 7) all report 0 executed, 2–3 rejected — no new entries are clearing. Cash has sat frozen at $1,760.89 while total account value drifts between roughly $9,817 and $9,925, below the $10,000 starting mark. The last executed fills date back to June (KO, CAT, UNH, MSFT). So the live book is fully allocated, slightly underwater, and unable to act on fresh signals — likely a sizing or buying-power constraint rejecting orders.
Verdict
The strengths are real: two consecutive positive out-of-sample folds with respectable risk-adjusted returns, and a transparent, robust rule set. The risks are equally real: a failed overfitting check, a punishing 21.7% drawdown, a sub-40% win rate that demands patience, and a live account currently stuck rejecting every order while sitting below par.
Our read: donchian-breakout has a plausible edge but hasn't earned unconditional trust. The rejected-order pattern needs a fix before the strategy can even express its thesis, and the validation gate is right to keep it on a short leash until more out-of-sample data accumulates.