The thesis
The donchian-breakout strategy runs one of the oldest ideas in systematic trading: buy the 20-day high, exit on the 20-day low. It trades a 24-name universe of US large caps spanning tech (AAPL, MSFT, NVDA), payments (V, MA), healthcare (JNJ, UNH, PFE), staples (KO, PG, WMT, COST) and industrials/energy (CAT, XOM, CVX). The logic is pure trend-following: ride what is breaking out, cut what is breaking down, and accept that most trades will be small losers paid for by a few large winners.
Recent activity
The tape bears that character out. Late-September runs put capital to work — buying 10 NVDA at $230.43 and 3 CAT at $825.98 while rotating out of KO, MA, V and UNH. Since then, though, the book has stalled. Every scheduled run from October 1 through 6 executed zero trades while rejecting one or two signals apiece. The cause is visible in the ledger: cash sits pinned at $813.35 against a total equity near $10,093. The strategy is seeing breakouts it cannot afford to buy. That is a capital-allocation problem, not a signal problem, but it blunts the engine precisely when fresh trends may be forming.
Backtest and validation
Over 451 days the backtest returned +6.95% (final equity $10,695), a 3.83% CAGR with a Sharpe of just 0.34. The win rate is a lowly 38.46% across 108 trades — unremarkable for trend-following, where asymmetry matters more than hit rate — but the 21.73% max drawdown is a real cost of admission, and turnover of 2,082% means the approach churns hard for its return.
Crucially, the validation gate did not pass. Across four walk-forward folds, three were positive, and the out-of-sample stretch looks genuinely good (+11.35%, Sharpe 1.86; the prior fold managed +14.09% at Sharpe 2.72). But fold 2 lost 7.46% at a Sharpe of -1.07, and the deflated Sharpe ratio of 0.198 — adjusted for 6 trials — signals the headline numbers may not survive multiple-testing scrutiny. The probabilistic Sharpe of 0.674 is suggestive, not decisive.
The balance
Strengths: the recent two folds are strong and consistent, drawdowns in the good regimes stayed shallow (3–6%), and the rules are transparent and hard to overfit. Risks: a sub-0.2 DSR, a 22% peak drawdown, a losing fold, and a live book that is currently too cash-constrained to act on its own signals.
Verdict: a credible classical strategy showing real out-of-sample promise, but one that has not cleared the bar — and needs its sizing fixed before we judge the live edge.