Thesis
channel-pullback is a mean-reversion strategy dressed in trend-following clothes. It buys pullbacks to the lower regression channel or volume support only within confirmed uptrends, then exits at the upper channel or resistance. In principle this is a disciplined idea: you are trying to buy dips in things that are still structurally rising, rather than catching falling knives. The universe is a conservative, large-cap slate — 24 names spanning tech (AAPL, MSFT, NVDA), financials (JPM, BAC, V, MA), healthcare (JNJ, UNH, PFE), and consumer staples (PG, KO, WMT, COST).
Recent Activity
The strategy is live and trades on a daily scheduled cadence. Over the last week most runs executed zero or one order, and the paper account has hovered around $9,500–$9,620 in total value — notably below a $10k notional start, so live paper performance is currently underwater. Recent fills cluster in financials and AAPL: it round-tripped BAC (bought 30 shares at $63.04 on 9/03 after selling 31 at $62.48 the day before) and repeatedly cycled AAPL through late August (buys near $308–$315, sells near $311–$314). The tight, rapid in-and-out pattern in the same names is consistent with its mean-reversion mechanics — but also with churn.
Backtest & Validation
The backtest looks superficially fine: +7.62% total return, final equity $10,761, a 4.19% CAGR over 451 days, and a 14.83% max drawdown. Look closer and the cracks show. The win rate is just 39.39% across 137 trades — this strategy loses more often than it wins and relies on asymmetric winners. Turnover is a staggering 2,311%, and while fees were only $137, that level of activity is a headwind that thinner-spread live conditions can punish.
Walk-forward validation is where it fails outright. Across four folds, three were positive — a solid +6.53% (Sharpe 1.25), an excellent +20.68% (Sharpe 3.86), and a modest +3.63% (Sharpe 0.74) — but fold 2 (Jan–Jul 2025) cratered -11.42% with a Sharpe of -1.70 and a 16% drawdown. Out-of-sample return averaged 3.63% at a 0.74 Sharpe, respectable but far from the folds' best. The killer is the deflated metrics: with 7 trials, the Deflated Sharpe Ratio is 0.196 against a full-sample Sharpe of only 0.40. Probabilistic Sharpe (0.702) clears a low bar, but DSR does not — the edge does not survive a correction for selection.
Verdict
Strengths: a coherent, risk-aware thesis; a clean large-cap universe; and genuinely strong performance in favourable regimes (fold 3). Risks: a sub-40% win rate, extreme turnover, one regime (early-2025) that inflicted double-digit losses, and a failed validation gate driven by a weak deflated Sharpe. This is a strategy that can work when trends are clean and fail hard when they break. Keep it on a short leash and small size until it shows it can survive a fold-2-type regime live.