The Thesis
Donchian-breakout runs a classic trend-following rule: buy a 20-day high breakout and exit on a 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 logic is deliberately simple and well-understood, which makes it a useful baseline for how a textbook breakout system behaves on today's mega-caps.
Recent Activity
The agent has gone quiet on execution. Scheduled runs on 2026-10-01, 10-02 and 10-05 each reported 0 executed, 1 rejected, and the book is sitting on just $813.35 in cash against a total of roughly $10,016. That cash constraint is the likely culprit behind the rejections — fresh breakout signals can't be filled when the account is nearly fully invested.
The last real activity came late September. On 09-30 the agent bought 10 NVDA at $230.43 and trimmed 1 WMT; on 09-29 it added 3 CAT at $825.98 while selling out 23 KO at $86.90. Mid-September saw rotations into AAPL and V and exits from MA, V and UNH. The pattern is consistent with the thesis: rotating capital toward names making new highs and cutting positions rolling over.
Backtest and Validation
The headline numbers are honest but unexciting. Over 451 days the strategy returned 6.95% (3.83% CAGR) to a final equity of $10,695, with a Sharpe of just 0.34 and a maximum drawdown of 21.73%. The win rate is 38.46% across 108 trades — low, but normal for trend-following, which relies on a few large winners to pay for many small losses. Turnover is a steep 2,083%, though fees totalled only $108.
The cross-validation picture is more interesting. Three of four folds were positive, and the two most recent folds were genuinely strong: fold 3 returned +14.09% at a 2.72 Sharpe, and fold 4 +11.35% at 1.86. The painful stretch was fold 2 (2025-01 to 2025-07), which lost -7.46% at a -1.07 Sharpe — a whipsaw regime where breakouts repeatedly failed.
Crucially, validation did not pass. Despite an out-of-sample Sharpe of 1.86, the deflated Sharpe ratio sits at 0.198 and the probabilistic Sharpe at 0.674, with 6 trials considered. In plain terms: the full-sample edge is too thin and too drawdown-heavy to clear the bar with confidence, even if recent folds flatter it.
Verdict
The strengths are real — a transparent rule, encouraging recent out-of-sample behaviour, and controlled costs. The risks are equally real: a sub-0.4 full-period Sharpe, a 20%+ drawdown, a low win rate that demands patience, and a failed validation gate. The current cash crunch is a practical reminder that the live book needs room to act on its own signals. For now, donchian-breakout is a respectable baseline worth watching, not yet a strategy worth trusting capital to.