The thesis
Donchian-breakout is a textbook trend-following system: it buys when a name in its 24-stock large-cap universe prints a fresh 20-day high, and exits when price falls back to a 20-day low. The universe spans mega-cap tech (AAPL, MSFT, NVDA), payments and financials (JPM, V, MA), staples (PG, KO, WMT, COST) and industrials (CAT, HON) — a liquid, diversified pool well suited to a breakout approach that needs clean, sustained moves to pay for the many small losses it will take along the way.
Recent activity
The last fortnight has been quiet on execution and loud on rejections. Across six scheduled runs from 14–21 October, the strategy fired signals but executed almost nothing: five of six runs logged zero fills with two or three rejections each, the lone exception being 15 October (2 executed, 1 rejected). Cash has been pinned at $2,241.07 and total paper equity drifted between roughly $9,734 and $9,794 — below the ~$10,000 base implied by the backtest. The repeated rejections suggest breakout signals are outrunning available capital; with little dry powder, the book simply can't take the trades it wants.
The fills that did go through tell the trend-following story. A 10-share NVDA sell at $217.62 on 15 October closed out an entry at $230.43 from 30 September — a small clipped loss, exactly the kind the system is designed to accept. Rotations into WMT and MSFT and out of CAT and KO round out the recent tape.
Backtest and validation
Over 451 days the backtest returned 6.95% (final equity $10,695), a 3.83% CAGR, with a Sharpe of just 0.34 and a punishing 21.73% max drawdown. The win rate is 38.46% across 108 trades — low, but normal for breakout systems that lose small often and win big occasionally. Turnover of 2,083% is heavy, though fees ($108, no FX cost) stay modest at this scale.
Walk-forward validation is the most telling section — and it did not pass. Three of four folds were positive, and the two most recent out-of-sample windows were genuinely strong: +14.09% at Sharpe 2.72, then +11.35% at Sharpe 1.86. But fold 2 bled −7.46% at Sharpe −1.07, and the deflated Sharpe ratio of 0.198 — which penalises the six trials run — falls short of significance even though the probabilistic Sharpe reads 0.674.
Strengths and risks
The strengths are real: a disciplined, mechanical rule set, strong recent out-of-sample behaviour, and loss control consistent with the design. The risks are equally real: a sub-0.5 full-sample Sharpe, a drawdown north of 20%, a failed validation gate, and a live book currently too cash-constrained to act on its own signals. This is a strategy with a credible edge in trending regimes and a clear vulnerability in choppy ones — promising enough to keep live and monitor, not yet proven enough to lean on.