Thesis
Donchian-breakout is a textbook trend follower: it buys when price closes at a fresh 20-day high and exits when price sinks to a 20-day low. The idea is to capture sustained directional moves and cut positions before a trend fully reverses. The strategy runs across a 24-name universe of megacap and blue-chip equities — AAPL, MSFT, NVDA, JPM, V, MA, UNH, XOM, CAT and peers — a liquid, well-behaved pool where channel breakouts are less prone to microstructure noise.
Recent activity
The strategy is live but quiet. Across scheduled runs from August 10–17, execution has been thin: most days show zero fills with one to three rejected orders, and total portfolio value has drifted in a narrow band around $9,750–$9,850 on roughly $1,510 of idle cash. The last real activity was August 12–13, when it bought V (5 sh @ $362.35) and MA (3 sh @ $559.85) while selling AAPL (7 sh @ $301.65) and trimming UNH. That pattern — rotating into financials on fresh highs while exiting names rolling over — is exactly what the channel logic prescribes. The high rejection count suggests either position-sizing or cash constraints are blocking signals, worth a closer look operationally.
Backtest and validation
Over 451 days the backtest returned 6.95% (3.83% CAGR) on 108 trades, ending at $10,695 equity. The headline caveat is a Sharpe of just 0.34 against a 21.73% max drawdown — a rough return-to-risk profile. The 38.46% win rate is normal for trend following, where a minority of large winners is expected to carry many small losers, but the shallow Sharpe means those winners weren't large enough here. Turnover of 2,083% is high and generated $108 in fees; frictions matter for this style.
Walk-forward validation is more encouraging in places and is why the story is nuanced. Three of four folds were positive, and the two most recent out-of-sample windows are genuinely strong: fold 3 returned 14.09% (Sharpe 2.72) and fold 4 returned 11.35% (Sharpe 1.86) with a shallow 6% drawdown. Fold 2, however, lost 7.46% (Sharpe -1.07), exposing the classic trend-following weakness — choppy, range-bound regimes where breakouts fail repeatedly.
The verdict
The validation gate failed. Despite a healthy Probabilistic Sharpe (PSR 0.674), the Deflated Sharpe Ratio came in at just 0.198 — below the bar once you account for the 6 trials behind this configuration. In plain terms: the recent out-of-sample strength is real but not distinguishable enough from luck given how many variants were searched. That is the right conservative call.
Bottom line
Strengths: coherent trend-following logic, a liquid universe, and two consecutive strong OOS folds. Risks: a sub-0.35 full-sample Sharpe, a 21% drawdown, regime sensitivity (fold 2), high turnover, and a failed deflated-Sharpe check. The improving fold trajectory earns it continued paper-trading observation — but not an unconditional promotion until the edge survives the multiple-testing penalty.