The thesis
The donchian-breakout strategy runs one of the oldest ideas in systematic trading: buy when price sets a new 20-day high, exit when it makes a new 20-day low. It is pure trend following, applied to a universe of 24 US large caps spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), staples (KO, PG, WMT, COST) and energy (XOM, CVX). The appeal is simplicity and the hope that a handful of sustained moves pay for many small losers.
Backtest performance
Over 451 days the strategy returned 6.95%, ending at $10,695 of equity on a $10k base — a CAGR of just 3.83%. The risk-adjusted picture is weaker than the headline: a Sharpe of 0.34 and a 21.73% max drawdown mean the return was earned with meaningful volatility. The win rate of 38.46% across 108 trades is low, but that is expected for breakout systems that rely on asymmetry rather than hit rate. One number deserves a flag: turnover of 2,083%. That is heavy churn, and although total modelled fees were only $108, real-world slippage on that activity would bite harder than the backtest suggests.
Validation: the gate it missed
Our walk-forward validation returned passed: false, and the detail explains why the call is nuanced rather than damning. Three of four folds were positive, and the most recent two were strong — fold 3 (+14.09%, Sharpe 2.72) and fold 4 (+11.35%, Sharpe 1.86). The out-of-sample window returned 11.35% at a Sharpe of 1.86, genuinely encouraging. But fold 2 (Jan–Jul 2025) lost 7.46% at a Sharpe of −1.07, exposing how the system bleeds in choppy, directionless markets. The deflated Sharpe ratio of 0.198 — adjusted for the 6 trials run — is what sinks the gate: once you account for selection effort, the edge is not statistically convincing. The PSR of 0.674 is suggestive but short of conviction.
Recent live activity
Live trading has been quiet and, frankly, frustrated. The last six scheduled runs executed only a handful of orders while rejecting one or two each time — a pattern that points to cash or position constraints rather than a lack of signals. Recent fills include buying NVDA (10 @ $230.43) and CAT (3 @ $825.98) while trimming KO, V and MA. Notably, total account value has drifted between $9,792 and $9,932 — below the $10k start — so the live paper account is modestly underwater even as the backtest shows a gain.
Verdict
Donchian-breakout is behaving exactly as a trend-follower should: patient, drawdown-prone, and dependent on a few big winners. The strong recent folds are real, but the failed validation and persistent order rejections argue for caution. Watch whether the capital-constraint problem gets resolved — and whether the 2,083% turnover survives contact with live costs — before reading much into the next up-leg.