The thesis
donchian-breakout runs one of the oldest ideas in systematic trading: buy the 20-day high, exit on the 20-day low. It is pure trend-following — no forecast, no mean-reversion, just a rule that rides momentum and cuts when the channel breaks the other way. The strategy trades a universe of 24 US large-caps spanning tech, financials, healthcare, staples, and energy (AAPL, NVDA, JPM, UNH, XOM and peers), so it is diversified across sectors rather than a single-name bet.
Backtest performance
Over 451 days the strategy returned 6.95% (final equity $10,695 on a $10k base), a 3.83% CAGR. The headline numbers are honest about the trade-off: a Sharpe of 0.34, a 21.73% max drawdown, and a 38.46% win rate across 108 trades. That low hit rate is expected for trend-following — the edge lives in a handful of large winners paying for many small losers — but the drawdown is steep relative to the return, and turnover is punishing at 2,083%, meaning the book churns roughly twenty times over. Fees ($108) are small in absolute terms but scale directly with that churn.
Validation: a split decision
Here the picture gets more interesting. Across 4 walk-forward folds, 3 were positive, and the two most recent folds are genuinely strong: fold 3 returned +14.09% (Sharpe 2.72) and fold 4 +11.35% (Sharpe 1.86), the latter being the out-of-sample tail. Fold 2, however, lost 7.46% (Sharpe -1.07) — the strategy struggled badly through the early-2025 window.
The out-of-sample Sharpe of 1.86 looks tempting, but our gate fails the strategy, and the deflated statistics explain why. The probabilistic Sharpe (PSR 0.674) is only moderately confident the true Sharpe exceeds zero, and once we deflate for the 6 trials run, the DSR falls to 0.198 — well short of conviction. In plain terms: after accounting for how many variants were tested, we can't rule out that the good folds are luck.
Recent activity: a stalled engine
The live sleeve tells its own cautionary story. Every scheduled run from 2026-08-25 through 2026-09-01 executed zero trades and rejected one or two each time, with cash pinned at $1,510.88. Over that stretch total equity drifted down from $10,261 to $10,019 — the book is fully invested, unable to act on new signals, and slowly bleeding. The last genuine executions were in mid-August (buys in V and MA, sells in UNH and AAPL).
Verdict
The strengths are real: a coherent, well-understood thesis; sector-diversified; two recent out-of-sample folds with strong Sharpes. The risks are equally real: a failed validation gate, a low deflated Sharpe, a 22% drawdown, heavy turnover, and a live account that has stopped transacting. This is a strategy to keep watching — not yet one to lean on.