The thesis
channel-pullback is a disciplined dip-buyer. It waits for a confirmed uptrend, buys pullbacks toward the lower regression channel or volume support, and exits into the upper channel or resistance. It trades a 24-name universe of large-cap staples and mega-caps — AAPL, MSFT, NVDA, JPM, PG, KO, UNH and peers — so this is a liquidity-rich, low-exotica sandbox.
The recent tape matches the description. Over the last two weeks the agent round-tripped PG (bought at $145.99, sold at $150.62; bought $144.98, sold $150.90) and KO (bought $81.44, later sold $88.05), plus a UNH swing sold at $433.57. These are textbook channel bounces: buy weakness in a staple, harvest the reversion.
Recent activity
Activity is deliberately sparse. The scheduled runs from 2026-07-27 through 2026-08-03 executed 0, 3, 1, 0, 1 and 0 trades respectively — most days the model simply finds no qualifying pullback and sits. Portfolio value has hovered tightly around its $10,000 base (total $10,004.34 on 2026-08-03), meaning the live paper account is essentially flat over this window. That patience is a feature for a pullback system, but it also means the edge, if real, accrues slowly.
Backtest and validation
The full backtest returns 7.62% over 451 days (4.19% CAGR), with a 0.40 Sharpe, a 14.83% max drawdown, and a 39.39% win rate across 137 trades. A sub-40% win rate paired with positive returns tells you the exits do real work — winners are bigger than the frequent small losers. Turnover of 2,311% is high, though fees ($137 total) stayed contained.
Walk-forward validation is where the caution lives. It failed. Three of four folds were positive, but the dispersion is severe:
- Fold 1: +6.53% (Sharpe 1.25)
- Fold 2: −11.42% (Sharpe −1.70, 16.09% drawdown)
- Fold 3: +20.68% (Sharpe 3.86)
- Fold 4: +3.63% (Sharpe 0.74)
Out-of-sample return was 3.63% with a 0.74 OOS Sharpe — respectable, but the deflated Sharpe ratio (DSR 0.196) is the sticking point. Across 7 trials, the probabilistic Sharpe (0.702) says there is a decent chance the edge is positive, yet the DSR near 0.2 warns that once you correct for selection, the evidence thins out.
Verdict: strengths and risks
Strengths: a coherent, explainable thesis; genuine mean-reversion behavior in live trades; strong performance in trending regimes (Fold 3); and modest fees relative to turnover.
Risks: the strategy is regime-fragile. Fold 2's −11.42% shows what happens when "pullbacks" become downtrends — the channel bounce never comes. A 39% win rate demands the exit logic keep working, and the failed validation plus a low DSR mean we should treat the headline 7.62% as unproven rather than repeatable.
channel-pullback is a worthy candidate to keep on the paper leaderboard, but it has not earned capital promotion. The next milestone is surviving another hostile fold without the double-digit drawdown.