The Thesis
Donchian Breakout is a classic trend-following system: buy 20-day high breakouts, exit on 20-day lows. It runs on a 24-name large-cap universe spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), healthcare (UNH, ABBV), and staples (KO, PG, WMT). The logic is deliberately simple and mechanical — it asks nothing of the analyst beyond patience, and it accepts a low hit rate in exchange for the occasional durable trend.
Backtest Snapshot
Over 451 trading days the strategy turned $10,000 into $10,695 — a 6.95% total return, or roughly 3.83% CAGR. That headline masks the texture underneath. The Sharpe of 0.34 is thin, the maximum drawdown of 21.73% is steep relative to the reward, and the win rate is just 38.46% across 108 trades. None of that is disqualifying for trend following — the model expects frequent small losses punctuated by larger wins — but the payoff asymmetry here isn't generous enough to fully justify the drawdown. Turnover of 2,082% is high, and while fees totalled only $108, that churn is a headwind worth watching.
Validation: The Gate It Didn't Clear
Our walk-forward validation is where the story gets interesting — and cautionary. Across four folds, three were positive. The two most recent (July–Dec 2025 and Dec 2025–May 2026) were genuinely strong, returning 14.09% and 11.35% at Sharpes of 2.72 and 1.86 with shallow drawdowns. That recency is encouraging. But fold 2 (Jan–Jul 2025) lost 7.46% at a Sharpe of −1.07, exposing how badly the system bleeds in choppy, trendless regimes.
Crucially, validation did not pass. The deflated Sharpe ratio (DSR) of 0.198 — which discounts performance for the 6 trials run — is far below what we'd want to call an edge robust. A PSR of 0.674 is middling. In plain terms: the out-of-sample results look good, but not so good that we can rule out luck given how many variants were tried. Balanced honesty demands we treat the recent hot streak as a hypothesis, not a conclusion.
Recent Live Activity
The live account tells a quieter tale. The last executed trades clustered in late May and early June 2026 — buys in KO, UNH, ABBV, AAPL, and CAT, with sells trimming MSFT and CAT. Since then, every scheduled run from July 9 through July 16 executed zero trades and logged one or two rejections, with cash pinned at $1,760.89 and total equity drifting between roughly $9,863 and $10,087. No new breakouts are firing, and the book is currently underwater against its $10,000 start.
Verdict
Donchian Breakout is a coherent, well-behaved trend follower with two excellent recent folds — but a failed robustness gate, a punishing drawdown profile, and a stalled live book. It stays live for observation, not conviction. The next trending regime will tell us far more than the last backtest did.