The thesis
channel-pullback is one of our hand-written, built-in strategies (Reviewer: "Approved, risk 0"). The idea is classic trend-following mean reversion: in a confirmed uptrend, buy pullbacks toward the lower regression channel or volume support, then exit at the upper channel or resistance. It trades a 24-name large-cap US universe — AAPL, MSFT, NVDA, JPM, V, JNJ, XOM and friends — so it is betting on dips within strong names, not on picking obscure turnarounds.
Recent activity
The strategy went live for Season 2 at a $10,000 reset on 2026-10-29. The scheduled runs tell a measured story: 0 trades that day, then two buys on 10-30 (MSFT ×3 @ $522.07 and AAPL ×5 @ $344.84), and 0 again on 11-02. That is the point of a pullback system — it waits, and most days it simply does nothing. Equity has drifted from $10,000 to $9,904.86, about -1% over the first few sessions, with $6,674.22 still in cash. Far too early to judge; mostly it signals the strategy is picky about entries.
The backtest
Over 451 days the full-sample backtest returns 7.62% (final equity $10,761.52, CAGR 4.19%), with a Sharpe of 0.40 and a max drawdown of 14.83%. The win rate is a low 39.39% across 137 trades — yet the strategy is net positive, which is the signature of a working mean-reversion edge: it loses small often and wins larger occasionally, exiting into strength. Turnover is a heavy 2,311%, though fees are modest ($137 total, no FX cost) given the liquid universe.
Walk-forward validation — the hard part
Here is where I temper the optimism. Validation did not pass. Three of four folds were positive, but the dispersion is alarming:
- Fold 1: +6.53% (Sharpe 1.25)
- Fold 2: -11.42% (Sharpe -1.70, 16% drawdown)
- Fold 3: +20.68% (Sharpe 3.86)
- Fold 4: +3.63% (Sharpe 0.74)
Out-of-sample return averages 3.63% at Sharpe 0.74 — better than the full-sample Sharpe, but driven heavily by one spectacular fold. The decisive numbers are the probabilistic ones: PSR is a borderline 0.702, but the Deflated Sharpe Ratio is just 0.196 after accounting for 7 trials. In plain terms, once we penalise for how many variants were tested, we cannot be confident the edge is real rather than lucky.
Verdict
channel-pullback is a coherent, cheap-to-run strategy with a genuine asymmetric-payoff profile and a sensible, liquid universe. Its strength is patience and small losses; its weakness is regime sensitivity — Fold 2 shows the dip-buying logic can bleed badly when the uptrend assumption breaks. The failed validation and low DSR mean it should be treated as a modest, probationary allocation, not a conviction bet. We will be watching whether live behaviour resembles the steady folds or the one that hurt.