The thesis
donchian-breakout runs one of the oldest ideas in systematic trading: buy the 20-day high, sell the 20-day low. It is pure trend following across a 24-name universe of US large caps — the usual megacap tech (AAPL, MSFT, GOOGL, NVDA), financials (JPM, V, MA), staples (PG, KO, WMT, COST) and industrials (CAT, HON). The bet is simple and honest: ride breakouts, cut losers quickly, and accept that most trades will be small losses in exchange for a few large winners.
The backtest bears out that profile. Over 451 days the strategy returned 6.95% (final equity $10,695), a CAGR of 3.83%, on a low win rate of 38.46% across 108 trades. A sub-40% hit rate is not a red flag here — it is exactly what a breakout system should look like. The concern is elsewhere: a Sharpe of just 0.34 and a 21.73% max drawdown mean investors are taking equity-like pain for bond-like reward. Turnover of 2,083% also tells you this book churns hard; fees of $108 are small today but scale with trading.
Recent activity
The last six scheduled runs (21–28 October) tell a revealing story: zero executions and a string of rejections each day, with cash parked at $2,241.07 and total equity drifting in a narrow band below ~$9,800. The earlier executed trades — buying WMT and MSFT, trimming NVDA and CAT through September and October — show the engine does fire when breakouts appear. Right now it simply isn't finding qualifying 20-day highs, so it sits on its hands. That is the strategy working as designed, not malfunctioning, but it does mean the live book is currently underwater relative to a $10k start.
Backtest vs. validation
Here is the honest part. Validation failed. The walk-forward split is genuinely encouraging — 3 of 4 folds positive, with folds 3 and 4 posting +14.09% (Sharpe 2.72) and +11.35% (Sharpe 1.86), and an out-of-sample return of 11.35% that actually beats the full-sample 6.95%. But fold 2 lost 7.46% with a negative Sharpe, and the deflated statistics are what sink it: a PSR of 0.674 is mediocre, and the Deflated Sharpe Ratio of 0.198 — adjusting for 6 trials — says we cannot be confident the edge survives the multiple-testing penalty.
Verdict
Trust the gate. The strategy has a coherent, time-tested thesis and genuinely strong recent folds, but the DSR is telling us the full-sample Sharpe is too thin to distinguish from luck once you account for the search. Keep it live as a paper candidate, watch whether it converts the next real breakout cleanly, but don't scale capital into it until the deflated metrics clear the bar.