The thesis
PanicFade v4 bets on a well-documented asymmetry: negative headline shocks tend to be overreacted to, then revert toward fundamentals. The rules are deliberately mechanical. It buys names where sentiment drops below -0.4 and the price is already technically oversold (RSI14 under 35, trading below its 20-day average), then sizes into the panic. It exits 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 universe is 24 liquid large caps (AAPL, NVDA, JPM, XOM, and peers), which keeps slippage and gap risk contained. The approach is grounded in recent sentiment-feedback literature (arXiv:2509.11970, 2025), not a hunch.
Recent activity
The last two weeks have been busy and cash-heavy. The scheduled runs show a steady drip of one-to-two executions per day with zero rejections, and total equity hovering between roughly $9,600 and $9,750 — essentially flat to slightly soft, with over half the book ($5,639 of $9,617 on 13 Nov) parked in cash. Trading clustered in NVDA and CAT. NVDA was round-tripped repeatedly (sold at $177.08, re-bought at $166.95, in and out again near $164). CAT illustrates the strategy's sharp edge: bought at $621.97 and sold the next day at $585.79, then re-entered at $561.74 and exited at $559.24 — two small losses where the fade met a name that kept falling. That is the hard-stop logic working as designed, but it is also a reminder that "panic" and "fundamental break" look identical on day one.
Validation and backtest
This is where PanicFade v4 earns a second look. Across four walk-forward folds spanning 2021–2026, all four were positive, and performance improved over time: +1.6%, +3.9%, +10.1%, and +21.7%, with fold Sharpes climbing from a weak 0.19 to a strong 2.44. The out-of-sample Sharpe of 2.44 and a Probabilistic Sharpe Ratio of 0.985 are excellent; crucially, the Deflated Sharpe of 0.56 holds above 0.5 even after penalising for 27 trials, so the result is not obviously a data-mining artefact. Full-period return is 37.93% over 1,233 days, max drawdown a modest 9.3%, win rate 62.9%.
Strengths and risks
Strengths: a clean validation story, controlled drawdowns, and a readable, literature-backed edge. The risks are just as clear. That 37.93% is only a 6.79% CAGR — respectable, not spectacular — and turnover is a staggering 5,807%, making the strategy acutely sensitive to fees and slippage that a backtest understates. The 2021–2022 fold barely broke even (Sharpe 0.19), hinting the edge thins in trending-down regimes. And live, the book is flat-to-down over the fortnight while sitting on heavy cash. Promising, well-governed, unproven in production.