Strategy Thesis
Donchian breakout is one of the oldest systematic trend-following rules in quantitative finance. The logic is deliberately simple: enter a long position when price closes at a 20-day high, exit when it closes at a 20-day low. No fundamental overlay, no sentiment filter — pure price structure. The universe is 24 liquid US large-caps spanning technology, financials, healthcare, consumer staples, energy, and industrials, giving the system broad sector exposure without venturing into illiquid names.
Backtest Performance
Over 451 days (roughly August 2024 through late May 2026), the strategy returned +6.95% on a $10,000 notional, compounding at a 3.83% CAGR. Those headline figures are modest, and the supporting statistics confirm the bumpy ride underneath: a Sharpe ratio of 0.34, a max drawdown of 21.73%, and a win rate of just 38.46% across 108 trades.
A sub-40% win rate is entirely consistent with trend-following — the model is designed to cut losses quickly and let winners run, not to be right most of the time. What matters is whether the average winner is large enough relative to the average loser. With 451 days of data and 108 trades, that sample is thin enough that the answer remains genuinely uncertain.
Turnover came in at a very high 2,082%, suggesting the 20-day window generates frequent entries and exits. At $1-per-trade fees the total cost was $108 — negligible — but in a real account with wider spreads or larger size, friction would compound meaningfully.
Walk-Forward Validation
The four-fold walk-forward tells a more nuanced story than the aggregate:
| Fold | Period | Return | Sharpe | Max DD |
|---|---|---|---|---|
| 1 | Aug 2024 – Jan 2025 | +0.73% | 0.21 | 6.71% |
| 2 | Jan 2025 – Jul 2025 | −7.46% | −1.07 | 15.58% |
| 3 | Jul 2025 – Dec 2025 | +14.09% | 2.72 | 3.14% |
| 4 | Dec 2025 – May 2026 | +11.35% | 1.86 | 6.01% |
Three of four folds were positive, and the two most recent periods are genuinely impressive — Fold 3 and Fold 4 each posted Sharpe ratios above 1.8 with contained drawdowns. The out-of-sample return of +11.35% (Fold 4) is the strongest single-period result in the entire history.
However, the platform's validation gate still flagged the strategy as not passed. The Probabilistic Sharpe Ratio (PSR) of 0.674 means roughly a one-in-three chance the true Sharpe is zero or negative after accounting for estimation noise. The Deflated Sharpe Ratio (DSR) of 0.198 — which penalises for the number of parameter trials tested — is considerably lower. With only six trials and 451 days of data, there simply isn't enough evidence yet to distinguish genuine edge from lucky clustering.
Live Deployment (First Five Days)
The strategy launched May 31 with $10,000 and has placed four trades: MSFT, AAPL, ABBV, and CAT — all buys triggered by 20-day high breakouts. Portfolio value as of June 4 stands at $9,917, a fractional drawdown from inception. Activity is consistent: one to two signals per day, with the system correctly filtering rejected orders on June 4.
Risk Factors
- Momentum crowding: A 24-stock large-cap universe means breakouts often coincide across names during broad risk-on moves, concentrating correlated exposure.
- Thin statistical base: 108 backtest trades and only ~15 months of history leave wide confidence intervals around every metric.
- Fold 2 underperformance: The January–July 2025 period lost 7.5% with a Sharpe of −1.07. Understanding what market regime drove that loss matters more than averaging it away.
Outlook
Donchian breakout is a legitimate, time-tested framework — the recent OOS performance is a genuine signal worth monitoring. But the validation flag is not bureaucratic noise: it is the system saying the evidence base is still too small to trust the edge is real. Give it another two to three folds of live data before drawing firm conclusions.