The Thesis
donchian-breakout is textbook trend-following: buy a 20-day high, exit on a 20-day low. The logic is to ride momentum when price pushes into new territory and cut the position when the trend rolls over. It runs on a diversified universe of 24 US large-caps spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (KO, PG, WMT) and industrials (CAT, HON). There is nothing exotic here — and that is part of the appeal. The strategy is now flagged live.
Recent Activity
The live tape tells a measured story. Over the scheduled runs from 23–30 September, executions were scarce: three of the six logged runs placed zero trades, while rejections kept piling up (two rejected on the 30th, two on the 28th, two on the 25th). When the breakout did fire, it acted decisively — buying 10 shares of NVDA at 230.43 and 3 of CAT at 825.98 on the latest runs, while trimming KO (23 shares at 86.90) and WMT. Portfolio total equity drifted from $9,932.29 on the 28th to $9,802.28 on the 30th.
That high rejection count is worth flagging. A trend system that rarely finds a valid entry — or keeps getting orders bounced — risks sitting in cash and missing the very moves it is designed to capture.
Backtest and Validation
Over 451 days the backtest returned 6.95% (final equity $10,695.11, CAGR 3.83%). The honest numbers underneath are less flattering: a Sharpe of 0.34, a 21.73% max drawdown, and a 38.46% win rate across 108 trades. Low win rate is normal for breakout systems — the design pays off through a few large winners — but pairing it with a sub-0.4 Sharpe and a drawdown north of 20% is a rough ride for the return delivered. Turnover of 2,083% also means the strategy churns hard, though total fees were a modest $108.
Crucially, validation did not pass. The walk-forward split is genuinely mixed: three of four folds were positive, and folds 3 and 4 were excellent (14.09% at Sharpe 2.72; 11.35% at Sharpe 1.86). But fold 2 lost 7.46% at a Sharpe of −1.07, exposing how badly the system can behave in the wrong regime. The deflated Sharpe ratio of 0.198 — adjusting for the 6 trials run — and a PSR of 0.674 simply do not clear the bar.
The Balanced View
The encouraging signal is the out-of-sample performance: an 11.35% OOS return at a 1.86 Sharpe suggests the edge is real when the trend cooperates. The caution is that it cooperates inconsistently, the drawdowns are steep, and the strategy's live behaviour so far looks hesitant. This is a credible classic worth keeping live and watched — but not one to size up until it proves it can fire cleanly and survive a hostile fold without the damage seen in early 2025.