The thesis
PanicFade v4 bets on a well-documented behavioural edge: negative headline shocks are systematically overreacted to and tend to revert toward fundamentals. The rules are disciplined rather than clever. It buys names carrying sentiment below -0.4 that are already technically oversold — RSI14 under 35 and trading below their 20-day moving average — and scales into the fall. Exits are equally mechanical: it takes profit on reversion (RSI14 back above 50 or price reclaiming the SMA20) and hard-stops fresh lows under continued negativity, treating that as a genuine fundamental break rather than noise.
The design is grounded in recent literature on sentiment asymmetries and negative-media overreaction (arXiv:2509.11970, 2025), and the strategy restricts itself to 24 liquid large caps — AAPL, MSFT, NVDA, JPM, UNH, WMT and the like. That universe choice matters: fading panic works best in names with durable fundamentals to revert toward.
Validation: the strong point
This is where PanicFade v4 earns attention. It passed a 4-fold walk-forward test with all four folds positive, and the out-of-sample return of 21.7% at a 2.44 Sharpe is genuinely impressive. Crucially, the folds improve monotonically — 1.6%, 3.9%, 10.1%, 21.7% — suggesting the edge has held, and arguably strengthened, into the most recent window.
The overfitting guards are reassuring. A Probabilistic Sharpe Ratio of 0.985 and a Deflated Sharpe Ratio of 0.56 across 27 trials say the result is unlikely to be luck, even after penalising for the search. Max drawdown stayed contained at 9.3% full-sample and just 2.6% in the latest fold.
The caveats
Balance demands two flags. First, the headline 37.93% total return is over ~1,233 days — a 6.79% CAGR. That is respectable and low-drawdown, but it is not the fireworks the top-line number implies. The full-sample Sharpe of 0.96 is honest and modest; the 2.44 lives mostly in the final fold.
Second, turnover is enormous at 5,807%. With 530 trades the strategy is highly active, and while modelled fees are trivial here ($530, no FX cost), any real-world slippage or spread would bite harder than the backtest suggests.
Live so far
The live book is quiet and, candidly, slightly underwater. Recent scheduled runs mostly execute nothing — several days of "0 executed" — reflecting that panic setups are rare by design. It did act on 21 September (buying HD, NKE and trimming WMT) and again on the 23rd (adding UNH), but total equity has drifted between roughly $9,898 and $9,977, below the $10,000 backtest starting base. That is a small sample and consistent with a strategy that waits for dislocation rather than forcing trades.
Verdict
PanicFade v4 is a well-validated, low-drawdown contrarian strategy with a credible academic thesis and unusually clean out-of-sample evidence. The risks are pedestrian rather than fatal: modest absolute returns, punishing turnover, and a live start that has yet to prove the backtest travels. Worth watching — patiently, which is rather the point.