Thesis
channel-pullback is a mean-reversion strategy that operates within trends rather than against them. It buys pullbacks toward the lower regression channel or volume support in confirmed uptrends, then exits into the upper channel or overhead resistance. The universe is 24 large-cap US names spanning tech (AAPL, MSFT, NVDA), financials (JPM, V, MA), healthcare (UNH, JNJ), staples (PG, KO, COST) and industrials (CAT, HON). It is currently marked live.
Recent activity
The strategy runs on a near-daily schedule, typically executing one trade per run. Over the last week it has stayed close to flat: total equity has hovered in a tight band around $10,010–$10,066, ending its 2026-07-24 run at $10,010.03 with $2,678.74 in cash. That is a book doing a lot of turning for little net movement.
The recent tape leans defensive. It round-tripped UNH — bought near $419.30 on 2026-07-15, sold at $433.57 on 2026-07-21 — and PG, bought at $146.44 on 07-14 and sold at $150.62 on 07-16, both clean winners consistent with the buy-support/sell-strength logic. It also trimmed AAPL at $327.54 the day after buying at $315.56. Fresh positions in UNH ($420.31), PG ($145.99) and KO ($81.44) show the same staples-and-healthcare tilt.
Backtest and validation
Over 451 days the backtest returned 7.62% (final equity $10,761.52), a 4.19% CAGR, with a Sharpe of 0.40 and a max drawdown of 14.83%. Win rate is a low 39.39% across 137 trades — but the positive return despite sub-40% hit rate signals healthy payoff asymmetry: winners are meaningfully larger than losers, exactly what a sell-into-strength exit should produce. The cost is turnover of 2,311% and $137 in fees; this is an active book, and slippage would matter in live size.
The honest headline is that validation did not pass. Walk-forward across four folds shows the fragility: three folds were positive, one badly negative. Fold 1 (+6.53%, Sharpe 1.25) and Fold 3 (+20.68%, Sharpe 3.86) were excellent, but Fold 2 (2025-01-22 to 2025-07-05) lost 11.42% at Sharpe -1.70 with a 16% drawdown. Out-of-sample aggregate return was +3.63% (OOS Sharpe 0.74) — positive, but a fraction of the full-sample figure.
The statistics tell the same story. The Probabilistic Sharpe Ratio of 0.70 is respectable, but the Deflated Sharpe Ratio of 0.196 — which penalizes for the 7 trials run — is well short of conviction. In plain terms: once we account for how many variants were tested, the observed edge is not distinguishable enough from luck to auto-promote.
Verdict
channel-pullback has a coherent thesis, disciplined exits, and a real payoff profile in favorable regimes. But regime dependence is its weakness — one adverse fold erases a lot — and the deflated statistics say the edge is thin. Keep it live on paper, watch behavior through a full down-cycle, and treat the failed gate as a reason to iterate, not to deploy capital yet.