The Thesis
PanicFade v4 makes a specific, testable bet: negative headline shocks are overreacted to and tend to revert toward fundamentals. The rule set is disciplined rather than vibes-driven — it buys names hit by sentiment below −0.4 only when they are already technically oversold (RSI14 under 35 and trading below their 20-day moving average), and it sizes into the panic. Exits are equally mechanical: it takes profit on reversion (RSI14 back above 50, or price reclaiming the SMA20) and hard-stops on fresh lows under continued negativity, treating that as evidence of a genuine fundamental break rather than a sentiment blip.
The design leans on recent sentiment-feedback literature (arXiv:2509.11970, 2025, and corroborating JRFM work) on asymmetric overreaction and reversion. Crucially, it fishes in a pond of 24 mega-cap, liquid names — AAPL, MSFT, JPM, UNH, XOM and peers — where 'panic' is more likely to be noise than terminal decline. That universe choice is doing a lot of the risk management.
Recent Activity
The live paper book has been quiet. Over the past week most scheduled runs executed nothing (0 trades on 17, 18, 22 and 24 September), with two bursts of activity: three fills on 21 September — buying HD and NKE, trimming WMT — and a single UNH buy on 23 September. That cadence is consistent with a strategy that waits for its setup rather than forcing trades.
Less flattering is the equity line. Total account value has drifted down over the week, from roughly $9,987 on 18 September to $9,898 on 24 September, with over half the book ($5,766) sitting in cash. The live account is, for now, running slightly soft — a useful reality check against the backtest.
Backtest and Validation
On paper the history is strong: 37.93% total return over 1,233 days (a 6.79% CAGR), a 0.96 Sharpe, a contained 9.3% max drawdown, and a 62.88% win rate across 530 trades.
The validation is where PanicFade v4 earns real credibility. Four-fold walk-forward testing returned positive in all four out-of-sample windows, and the results improved over time: 1.6% → 3.88% → 10.14% → 21.7%, with Sharpe climbing 0.19 → 0.66 → 0.97 → 2.44. The most recent fold (through August 2026) is the best. A Probabilistic Sharpe Ratio of 0.985 says the edge is very unlikely to be zero.
Strengths and Risks
The strengths are genuine: a coherent, literature-grounded thesis, mechanical entries and exits, and a rare clean sweep of positive out-of-sample folds. The risks deserve equal billing. The CAGR is modest for the drawdown and effort involved, and turnover of 5,807% signals heavy churn — fine at flat fees, but fragile if slippage rises. The deflated Sharpe of 0.56 (across 27 trials) is far more sober than the headline 0.96, a reminder that some of the shine is search bias. And the live account's gentle bleed is the honest counterpoint to a glowing backtest. Promising, well-built, but unproven where it counts most — in production.