The thesis
channel-pullback is a classic mean-reversion play riding a trend. It buys pullbacks toward the lower regression channel or a volume-support zone only inside confirmed uptrends, then exits at the upper channel or overhead resistance. The universe is 24 large-cap US names spanning tech (AAPL, MSFT, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE), and staples (PG, KO, WMT). It's a sensible, well-worn idea: buy dips in things that are still going up. The question is always whether the edge survives out of sample.
Recent activity
The strategy is live and trading daily. Over the last two weeks it has cycled quickly through positions — buying and selling AAPL around the $308–$315 band, a round-trip in BAC near $62–$63, plus entries in JPM, UNH, and others. This is high-churn behaviour, consistent with the backtest's eye-watering 2,311% turnover. The paper account tells a sobering story: total equity sits at roughly $9,618 as of the 3 September run, below the $10k starting line. Recent scheduled runs executed one trade or none at a time, so the drift is gradual rather than dramatic.
Backtest performance
Over 451 days the strategy returned 7.62% (final equity $10,761), a 4.19% CAGR, with a Sharpe of 0.40 and a 14.83% max drawdown. The win rate is a low 39.4% across 137 trades — meaning the average winner must meaningfully outrun the average loser for the strategy to stay positive. That's typical of trend-following mean reversion, but it leaves little margin. Fees totalled $137 with no FX cost; at this turnover, trading friction is a real headwind, not a rounding error.
Validation: the strategy fails our gate
This is where the case weakens. Walk-forward validation across four folds returned passed: false. Three of four folds were positive, but the picture is uneven:
- Fold 1 (Aug 2024–Jan 2025): +6.53%, Sharpe 1.25
- Fold 2 (Jan–Jul 2025): −11.42%, Sharpe −1.70 — the strategy broke
- Fold 3 (Jul–Dec 2025): +20.68%, Sharpe 3.86 — carried the whole result
- Fold 4 (Dec 2025–May 2026): +3.63%, Sharpe 0.74
Out-of-sample return averaged 3.63% with an OOS Sharpe of 0.74 — respectable, but the aggregate leans heavily on one exceptional fold. The Probabilistic Sharpe Ratio of 0.70 is encouraging, yet the Deflated Sharpe Ratio is just 0.196 after adjusting for 7 trials. A DSR that low means we cannot confidently reject the possibility that the edge is a product of selection.
Verdict
Strengths: a coherent, interpretable thesis; positive full-sample and OOS returns; three of four folds green. Risks: a failed validation gate, a single fold doing the heavy lifting, a sub-40% win rate, punishing turnover, and a live account currently underwater. channel-pullback is worth keeping on paper to gather live evidence, but the numbers do not yet justify real capital. The deflated Sharpe is the tell — treat the 7.62% headline with healthy scepticism.