The thesis
Channel-pullback is a disciplined mean-reversion strategy dressed in trend-following clothes. It only buys names already in a confirmed uptrend, then waits for price to dip back to the lower regression channel or a volume-support level before entering. Exits target the upper channel or overhead resistance. In plain terms: it tries to buy the dip in things that are already working, and sell into strength. The tradable universe is 24 large-cap US names spanning tech, financials, healthcare, staples and energy.
Recent activity
The strategy is live and trading a paper book that currently sits at roughly $10,025 total value, barely above its $10,000 start. Cadence is calm: scheduled runs over the past week executed zero or one trade each, with nothing rejected. The recent tape is notably defensive — buys in Procter & Gamble, Coca-Cola and UnitedHealth, and round-trips in UNH (bought at 419-420, sold at 433.57) and PG (bought near 146, sold at 150.62). A CAT entry at 931.39 and an AAPL trim at 327.54 round out the book. This is a strategy nibbling at staples pullbacks, not chasing momentum.
Backtest and validation
Over 451 days the backtest returned 7.62% (CAGR 4.19%), ending at $10,761 across 137 trades. That is respectable, but the quality metrics temper the enthusiasm. The Sharpe ratio is a soft 0.40, max drawdown reached 14.83%, and the win rate is just 39.39% — this is a strategy that loses more often than it wins and relies on winners outrunning losers. Turnover of 2,311% is heavy, meaning fees and slippage will bite in live conditions.
More importantly, the strategy failed our walk-forward validation. Across four folds it was positive in three, but the results are wildly inconsistent: fold 1 +6.53%, fold 2 −11.42% (Sharpe −1.7), fold 3 a stellar +20.68% (Sharpe 3.86), and fold 4 +3.63%. Out-of-sample return averaged 3.63% with a 0.74 Sharpe — decent, but the deflated Sharpe ratio (DSR) of just 0.196, adjusted across 7 trials, is well short of confidence. The probabilistic Sharpe ratio of 0.70 is suggestive but not decisive.
Strengths and risks
The encouraging read: the edge held up in three of four out-of-sample windows and produced positive OOS returns, and the defensive universe tilt should cushion violent selloffs. The pullback-in-uptrend logic is well-grounded and interpretable.
The risks are real. The single −11.42% fold shows the strategy can be badly wrong when uptrends fail and "support" keeps breaking — exactly the regime mean-reversion buyers fear. The low win rate and high turnover make it sensitive to execution costs, and the weak DSR signals that its headline return may owe as much to trial selection as to durable skill.
Verdict
Channel-pullback is a plausible, transparent idea that has not yet earned an auto-deploy stamp. It stays live on a paper book so we can watch it work in real regimes — but until the out-of-sample and deflated-Sharpe numbers firm up, treat its 7.62% as a hypothesis, not a track record.