The thesis
channel-pullback is a hybrid: it hunts for mean-reversion setups inside a trend-following filter. The rule is to buy pullbacks toward the lower regression channel or volume support, but only in confirmed uptrends, then exit at the upper channel or overhead resistance. In principle this captures the best of both worlds — you avoid catching falling knives by requiring an established uptrend, while still buying at a discount rather than chasing breakouts. It trades a 24-name large-cap universe spanning tech, financials, healthcare, staples and energy.
Recent activity
The strategy is live and trading daily. Over the last two weeks it has been a busy but disciplined operator — each scheduled run executes exactly one order with zero rejections. Its favourite recent vehicle has been BAC, which it round-tripped twice: buying 31 shares at $62.98 (Aug 6) and $64.07 (Aug 13), and selling at $64.52 (Aug 14). It also worked CAT and trimmed AAPL, selling 6 shares at $315.09 on Aug 19. Portfolio value has drifted gently lower across the window, from roughly $9,765 (Aug 12) to $9,611 (Aug 19) — a reminder that recent tape has been choppy for this approach.
Backtest and validation
Headline numbers are middling-to-fair. Over 451 days the backtest produced a 7.62% total return (≈4.19% CAGR), a 0.40 Sharpe, and a 14.83% max drawdown. The win rate is a low 39.4% across 137 trades — acceptable for a strategy that aims to win big and lose small, but it leans on a favourable payoff ratio it has not fully proven. Turnover of 2,311% is high, so fees and slippage matter.
The more important story is in the walk-forward validation, which did not pass. Three of four out-of-sample folds were positive, and fold 3 was excellent (+20.68%, Sharpe 3.86). But fold 2 (Jan–Jul 2025) was a genuine failure: -11.42% with a -1.70 Sharpe and a 16% drawdown. Aggregate out-of-sample return was a thin 3.63% at a 0.74 Sharpe.
Most tellingly, after adjusting for the 7 trials run, the deflated Sharpe ratio is just 0.196 — well below the confidence we require. The probabilistic Sharpe of 0.702 is positive but not decisive. In plain terms: the edge is plausible but statistically fragile, and a meaningful share of the backtest return could be luck.
Verdict
Strengths: a coherent thesis, disciplined execution, and one standout regime where it shone. Risks: a losing fold that hurt, sub-40% win rate, high turnover, and a deflated Sharpe that fails our gate. channel-pullback earns its live paper-trading slot as a learning candidate — but it has not yet earned real conviction.