The Thesis
channel-pullback is a disciplined dip-buyer with a trend filter. It waits for confirmed uptrends, then buys pullbacks to the lower regression channel or volume support, aiming to exit at the upper channel or overhead resistance. The idea is to harvest the "buy-the-dip" edge only when the broader trend is on your side — a sensible way to avoid catching falling knives. It runs a 24-name universe of large-cap U.S. equities spanning tech, financials, healthcare, staples, and energy (AAPL, MSFT, NVDA, JPM, UNH, KO, XOM, CAT, and more).
Backtest Performance
Over 451 days the strategy returned +7.62% (final equity $10,761.52 on a $10K base, ~4.19% CAGR). The headline numbers are honest but unspectacular: a Sharpe of 0.40, a max drawdown of 14.83%, and a win rate of just 39.39% across 137 trades. A sub-40% hit rate that still turns a profit tells you this is an asymmetric-payoff engine — winners run further than losers cut — which is exactly what a channel-exit rule is designed to do. Less flattering is the 2,311% turnover: this book churns hard, and at that trade frequency, slippage and fees (137 total here) are a real tax on the edge.
Validation: The Honest Red Flag
Our walk-forward validation marks this strategy failed, and it's worth being candid about why. Three of four folds were positive, but the dispersion is severe:
- Fold 1 (Aug 2024–Jan 2025): +6.53%, Sharpe 1.25
- Fold 2 (Jan–Jul 2025): −11.42%, Sharpe −1.70, 16.1% drawdown
- Fold 3 (Jul–Dec 2025): +20.68%, Sharpe 3.86
- Fold 4 (Dec 2025–May 2026): +3.63%, Sharpe 0.74
Out-of-sample results (+3.63% return, 0.74 Sharpe) are respectable but thin. The deciding numbers are the deflated statistics: with 7 trials in the search, the Deflated Sharpe Ratio is only 0.196 and the Probabilistic Sharpe Ratio 0.702 — not enough confidence that the in-sample edge survives the multiple-testing penalty. Fold 2's double-digit loss shows the strategy's Achilles' heel: choppy, trendless, or falling markets, where "pullbacks" keep pulling back.
Recent Live Activity
Live trading has been calm. The last week of scheduled runs executed 0–2 orders each — recent fills include buying KO (24 @ $81.44) and UNH (4 @ $419.30), trimming PG, AAPL, and DIS — with total account value oscillating narrowly between roughly $9,959 and $10,059. A 2026-07-20 run executed nothing at all. That's consistent with a selective, patient system waiting for setups rather than forcing trades.
Verdict
channel-pullback is a coherent, well-behaved strategy with a clear edge in trending regimes and sensible risk during them (Folds 3 and 4). But the validation gate is right to hold it back: the return is modest, the Sharpe is low, turnover is punishing, and Fold 2 proves it can hemorrhage in the wrong tape. It's a promising candidate for further refinement — tighter regime filters or lower turnover — not yet a conviction allocation.