The thesis
Channel-pullback runs a familiar pattern: buy dips to the lower regression channel or volume support inside a confirmed uptrend, then exit into the upper channel or resistance. It is a trend-following entry dressed as mean reversion — you are not fading strength, you are paying a discount to rejoin it. The universe is 24 large-cap US names spanning tech, financials, staples, healthcare, and energy (AAPL, MSFT, NVDA, JPM, V, JNJ, PG, XOM, CAT and peers), so the strategy has room to rotate rather than concentrate.
Recent activity
The last week has been quiet. Scheduled runs from Aug 3–10 executed at most two orders a day, with zero rejections — the signal simply isn't firing often. Recent trades read as tight round-trips: a BAC buy at 62.98 (Aug 6) sold at 63.12 the next session, and a CAT position bought at 887.13 (Aug 4) trimmed at 843.72 — a loss taken cleanly rather than held. Paper equity sits at $9,847.53 as of Aug 10, drifting down from $10,034.69 on Aug 5. That's a small live drawdown, consistent with a strategy that grinds rather than lunges.
Backtest and validation
Over 451 days the backtest returned 7.62% (4.19% CAGR) with a Sharpe of 0.40 and a 14.83% max drawdown. Two numbers deserve scrutiny. First, the 39.39% win rate: fewer than four trades in ten are winners, so the edge depends entirely on winners outrunning losers — a classic, but fragile, trend-follower payoff profile. Second, turnover of 2,311%: this book churns aggressively, and the $137 in fees across 137 trades is a real, if modest, tax on a strategy this thin on margin.
Most importantly, validation did not pass. Three of four walk-forward folds were positive, but the spread is alarming: fold 2 (Jan–Jul 2025) lost 11.42% at a Sharpe of −1.70, while fold 3 booked +20.68% at a Sharpe of 3.86. That is not a stable edge — it is one great regime carrying two mediocre ones and one bad one. The out-of-sample return of 3.63% (Sharpe 0.74) is honest but unremarkable. The deflated Sharpe ratio of 0.196 — which discounts for the 7 trials run — is the tell: after accounting for selection, there is little statistically robust alpha here, even though the probabilistic Sharpe of 0.702 looks superficially reassuring.
Verdict
Channel-pullback is a coherent, disciplined strategy that cuts losers and posts a positive full-period return. But it earns its "failed validation" flag. The regime dependence in fold 2, the sub-40% hit rate, and a near-zero deflated Sharpe all point to an edge that is real in good tape and absent in bad. Running it live in small size is a reasonable way to gather forward evidence; sizing it up before it clears validation would be trading hope, not proof.