The thesis
PanicFade v4 is a contrarian mean-reversion strategy with a specific, defensible premise: negative headline shocks tend to be overreacted to and revert toward fundamentals. Operationally, it buys large caps hit by a sentiment reading below −0.4 that are already technically oversold — RSI14 under 35 and trading below their 20-day moving average — and sizes into the decline. It exits on reversion, when RSI14 climbs back above 50 or price reclaims the SMA20, and hard-stops names that keep printing fresh lows under continued negativity, treating that as a genuine fundamental break rather than a buyable panic. The design is grounded in recent sentiment-feedback research (arXiv:2509.11970) and runs on a tight 24-name universe spanning mega-cap tech, financials, healthcare, staples, and energy.
Recent activity
The book was dormant for most of late September, then acted. On Oct 1 it executed two trades: a 5-share MCD buy at $231.79 and a 21-share BAC sell at $53.60. That BAC exit is instructive — the strategy had bought the same 21 shares at $59.86 on Sept 15, so it booked a loss of roughly $131 as the position kept falling. That is exactly the "fundamental break" stop the thesis describes, cutting a losing fade rather than waiting for a reversion that never came. The WMT round-trip went the other way: bought at $103.12 on Aug 28, sold at $107.48 on Sept 21 for a modest gain on reclaimed strength. Sept 21 also added HD and NKE; Sept 23 added UNH. Notably, the paper total drifted down over the week — from $9,922.84 (Sept 25) to $9,784.14 (Oct 1) — while cash sat near $5,720, meaning the engine is holding a lot of dry powder.
Backtest and validation
Across 1,233 days and 530 trades the backtest returns 37.93% (6.79% CAGR), a 62.88% win rate, and a 9.3% max drawdown, for a Sharpe of 0.96. More reassuring than the headline is the walk-forward structure: all four out-of-sample folds are positive, and returns improve across them — 1.6%, 3.88%, 10.14%, and 21.7% — with the latest fold posting a 2.44 Sharpe. The deflated Sharpe ratio of 0.56 survives 27 trials, and the probabilistic Sharpe ratio of 0.985 both suggest the edge is not simply overfit noise.
Strengths and risks
The strengths are real: a clean, researched thesis; validation that holds up out-of-sample; shallow drawdowns; and a stop discipline that actually fired this month. The risks are equally real. Turnover is extreme at 5,807%, so slippage and real-world costs will bite harder than the flat fee model implies. A 0.96 overall Sharpe is merely decent, and the escalating fold returns may partly reflect a friendlier recent regime rather than pure skill. And the live book is modestly underwater this week — a fitting reminder that a strategy built to buy fear needs that fear to revert. Worth watching closely; not yet worth betting the firm on.