The thesis
PanicFade v4 is a contrarian mean-reversion strategy. Its wager is that negative headline shocks are overreacted to and revert toward fundamentals, so it buys names hit by sentiment below -0.4 while already oversold — RSI14 under 35 and trading below the SMA20 — and sizes into the panic. It exits on reversion (RSI14 back above 50, or price reclaiming the SMA20) and hard-stops fresh lows on continued negativity, reading that as a genuine fundamental break rather than noise. The universe is 24 mega-cap names (AAPL, MSFT, JPM, JNJ, XOM and peers), which keeps the fade anchored to liquid, well-covered stocks. The thesis is grounded in recent q-fin literature on sentiment feedback and media-overreaction (arXiv:2509.11970, 2025).
Validation
The walk-forward record is the strongest part of the story. Across four folds spanning 2021–2026 the strategy was positive every time, with returns climbing fold to fold: 1.6%, 3.88%, 10.14%, and 21.7%. The most recent out-of-sample fold posted a 2.44 Sharpe at just a 2.6% max drawdown. Deflated metrics are reassuring given 27 trials were run: a Probabilistic Sharpe Ratio of 0.985 says the positive Sharpe is very unlikely to be luck, and the Deflated Sharpe of 0.56 survives the multiple-testing haircut, if not emphatically. The static reviewer approved it at a low 0.35 risk score.
Recent activity
As of the 2026-10-29 scheduled run, PanicFade executed zero trades and holds $10,000 in cash — a flat start consistent with the Season 2 reset. No panic conditions met its entry bar, which is the strategy working as designed: it should sit idle when nothing is oversold. The development history is worth noting. Early versions never traded in simulation at all — the triple-AND sentiment gate was structurally unreachable because historical per-symbol news was sparse, so sentiment defaulted to zero. The fix redefined panic by price action (deep RSI, distance below SMA20, crash-day drops) with sentiment demoted to an amplifier. Versions 2 through 4 then hardened execution: haircut-adjusted sell proceeds, sanity-banded cost-basis inference, verified rotation, and typeof-guarded defaults.
Strengths and risks
Strengths: genuinely clean validation, four-for-four positive folds, strong PSR, a shallow 9.3% full-period drawdown, and a 62.88% win rate over 530 trades.
Risks: the full-period Sharpe is only 0.96 and CAGR a modest 6.79% — the headline 37.93% is accumulated over 1,233 days. Turnover is very high at 5,807%, so frictions matter, though fees were a slim $530. Most importantly, the validated engine is really a price-defined panic fader; the sentiment edge central to the thesis was never testable in history and remains unproven live. The accelerating fold returns are encouraging but could equally reflect a mean-reversion-friendly regime. A flat, patient start is fine — but the live sentiment layer still has everything to prove.