The thesis
PanicFade v4 bets that negative headline shocks are overreacted to and revert toward fundamentals. In its cleanest form it buys oversold large caps — RSI14 below 35 and trading under their SMA20 — that have also been hit by a sentiment reading below -0.4, then sizes into the panic and exits on reversion (RSI14 back above 50 or price reclaiming the SMA20). A hard stop on fresh lows treats continued negativity as a genuine fundamental break rather than a fade opportunity. The idea is grounded in recent q-fin literature on sentiment feedback and media-overreaction reversion (arXiv:2509.11970, 2025), which gives it a more credible footing than a purely price-mined pattern.
What the validation says
The numbers are the strongest part of the story. Across a 1,233-day backtest the strategy returned 37.93% (roughly 6.79% CAGR) with a 62.88% win rate over 530 trades, a Sharpe of 0.96, and a contained 9.3% max drawdown. More importantly, the walk-forward check passed cleanly: all four folds were positive, out-of-sample return came in at 21.7% with an out-of-sample Sharpe of 2.44, and the deflated statistics are reassuring — PSR of 0.985 and DSR of 0.56 across 27 trials. The fold progression is monotonic in a good way, from a barely-positive 1.6% through the 2022 drawdown regime up to 21.7% in the most recent window. A strategy that survives being re-fit and re-tested four times, after deflating for 27 attempts, is not obviously curve-fit.
The honest caveats
Two things temper the enthusiasm. First, the thesis and the implementation have drifted. The improvement log shows the original sentiment gate was structurally unreachable in historical simulation — news coverage is sparse, sentiment defaulted to zero, and the triple-AND never fired. v4's fix redefines panic through price action (deep RSI, distance below SMA20, crash-day drops) with sentiment demoted from a gate to an amplifier. That is a sensible engineering call, but it means the live edge is mostly price-mean-reversion, not the sentiment-overreaction effect the research citation describes.
Second, and more immediately: it hasn't traded. Deployed 2026-08-06 with $10,000, the initial run and the next scheduled run both executed zero trades. That is by design — no name currently clears the panic bar — but a contrarian that only acts in dislocations will sit in cash through calm markets, and its live track record is still a blank page. The 6.79% CAGR also flatters poorly against the turnover (5,807%), so execution assumptions matter; the v4 hardening (live-quote-only accounting, sanity-banded cost stops, verified rotation) directly addresses that risk and earned a cautious reviewer sign-off at risk 0.35.
Verdict
PanicFade v4 is one of the better-validated strategies in the stable: positive in every fold, strong deflated metrics, and unusually careful plumbing. The open questions are whether its real edge is truly sentiment-driven or just oversold-reversion wearing a research citation, and whether the entry bar is so strict that opportunities are rare. It has earned its live slot — now it needs a panic to fade.