Thesis Recap
Channel-pullback is a trend-continuation strategy: it buys large-cap equities from a 24-name universe when price pulls back to the lower boundary of a linear regression channel — ideally coinciding with volume support — and exits at the upper channel or a defined resistance level. The setup requires a confirmed uptrend, so the strategy is inherently selective and avoids chasing breakouts.
Backtest at a Glance
Over 451 trading days, channel-pullback compounded to a 7.62% total return (4.19% annualised) on a starting book of roughly $10,000, finishing with $10,761 in final equity. Those numbers sound reasonable in isolation, but the supporting metrics complicate the picture:
| Metric | Value |
|---|---|
| Sharpe ratio | 0.40 |
| Max drawdown | 14.83% |
| Win rate | 39.39% |
| Turnover | 2,311% |
| Trades | 137 |
A Sharpe of 0.40 is below the 0.5 threshold most systematic desks treat as a minimum for live capital. A 39% win rate is workable for a trend-following strategy — wins need to be meaningfully larger than losses — but the 14.83% max drawdown relative to a 7.62% total return implies the strategy spent extended periods underwater relative to its cumulative gains.
Fold-by-Fold: Where It Works (and Where It Doesn't)
The walk-forward validation splits the backtest into four roughly equal out-of-sample windows. Three of the four folds were profitable, which is encouraging, but the dispersion is wide:
- Fold 1 (Aug 2024 – Jan 2025): +6.53%, Sharpe 1.25 — strong, controlled drawdown of 6.16%
- Fold 2 (Jan 2025 – Jul 2025): −11.42%, Sharpe −1.70 — the strategy's worst period, with a 16.09% peak-to-trough decline
- Fold 3 (Jul 2025 – Dec 2025): +20.68%, Sharpe 3.86 — exceptional; the strategy's clear best environment
- Fold 4 (Dec 2025 – May 2026): +3.63%, Sharpe 0.74 — the live/most-recent fold, positive but muted
Fold 3's outsized return skews the aggregate meaningfully. Strip it out and the picture is far more modest. Fold 2's deep loss suggests the strategy struggles in range-bound or whipsaw conditions where trend confirmation signals deteriorate.
Validation result: not passed. With a Probabilistic Sharpe Ratio of 0.702 and a Deflated Sharpe Ratio of just 0.196 across 7 parameter trials, the strategy has not cleared the statistical bar set by the platform's auto-deploy gate. The DSR in particular — designed to penalise for parameter search across multiple trials — is a meaningful flag.
Recent Activity
The strategy has been running daily scheduled sweeps throughout the first two weeks of July. Activity has been light: four consecutive sessions (July 2–6) logged zero executions with one rejection each day, suggesting the market's current positioning isn't offering many clean channel-touch entries. The week of July 7–9 saw four executions: a DIS round-trip (bought at $96.32, sold at $95.85 — a small loss), a COST exit at $910.50, a WMT entry at $112.99, and an AAPL exit at $314.01. Portfolio NAV has ranged between $10,155 and $10,333 over the observed window, with cash sitting elevated at $5,868 as of July 9 — indicating the strategy is running roughly 58% to cash, consistent with few qualifying setups.
Strengths and Risks
Strengths: The universe is clean large-cap names with deep liquidity. The thesis is grounded in well-understood mean-reversion-within-trend dynamics. Three of four out-of-sample folds were profitable, and the live fold (Fold 4) is the basis for the reported OOS Sharpe of 0.74 — the strategy's best risk-adjusted figure.
Risks: Regime dependence is the primary concern. Fold 2's −11% drawdown shows the strategy can go badly wrong in choppy markets. Turnover of 2,311% signals frequent rebalancing, which amplifies transaction cost sensitivity at scale. The failed validation gate, driven mainly by the low DSR, means the strategy has not demonstrated robustness across its parameter space — a necessary condition before treating the backtest results as reliable forward expectations.
Bottom Line
Channel-pullback is a coherent, rules-based strategy with a defensible thesis and a positive (if unspectacular) headline return. It is currently running conservatively in paper-trade mode, which is appropriate given the failed validation. The path to live capital requires either a broader data window that confirms the Fold 3 environment wasn't an outlier, or parameter refinement that tightens the DSR. Watch the live fold metrics over the coming quarters before drawing stronger conclusions.