The index is deceptively calm while the stocks inside it are swinging hard — AI sentiment flips from doom to euphoria within days, and oil plus bond-market shocks hit sectors very differently. Dispersion trades profit when single-stock volatility runs hot

Verdict: A validated dispersion setup that hasn't triggered — wait for the entry, don't chase it

The verdict is wait: this is a rules-based dispersion strategy whose entry has not fired. The strongest case for the trade is the realized backtest — a 46.4% return over 60 months on 42 trades at a 50% win rate with a 13.9% worst drawdown, holding up across the 24-month (23.0%) and 12-month (12.1%) windows — in exactly the regime Bloomberg described on September 27, 2026, where AI and oil tremors are boosting single-stock volatility relative to a calm index. The strongest case against is that the edge is crowded (per the idea's own thesis, hedge funds are already in this dispersion trade) and unconfirmed live: NVDA closed at $225.07, about $3.39 above its 20-day EMA of $221.68, so the required dip-and-reclaim close hasn't happened, and ATR data is unavailable so that gate cannot be verified; meanwhile Q2 free cash flow fell 56% sequentially to $21.4B and insiders showed roughly $569.5M of net open-market selling in the June 30, 2026 filing period. A full-size close at or below $208.89 on the NVDA leg, or an ADX (14) rise above 25, would flip this to avoid. **Conviction breakdown:** Thesis support 70 — the hot-single-stock-vs-calm-index regime the thesis needs is confirmed by two same-day Bloomberg reports. Trade readiness 35 — the EMA cross sequence is unmet and the ATR gate is unverifiable right now. Risk quality 55 — a fixed 2.4% stop against a 4.9% target gives roughly 2-to-1 reward-to-risk, but exits were filled on daily bars, so fills and drawdown are approximate. Backtest evidence 65 — realized results across three windows, though 50% win rates mean long losing runs are the base case, not the tail. Fundamentals trend 55 — revenue up 17.9% sequentially to $96.2B and 74.9% gross margin, offset by falling cash conversion, a 9.4-point net-margin compression to 62.0%, and a debt-to-equity jump from 0.04 to 0.15.

Trade now: conditions are close but not all live

NVDA closed at $225.07, about 3.4 points (1.5%) above its 20-day EMA of $221.68. The strategy's long entry needs three things at once: a fresh close crossing above the 20-day EMA, ADX (14) below 20, and ATR (14) above 4.5. The ADX gate is already met — ADX reads 14.5 versus the 20 threshold. But price is currently *above* the EMA, so the crossing condition cannot trigger until price dips back and closes back above it; the live distance to the EMA is $3.39. The ATR reading is not currently available, so that gate cannot be confirmed either. In short: nothing to buy today. What "wait" means concretely: stand aside until a daily close first drops to or below $221.68 and then closes back above it, with ADX still under 20 and ATR confirming above 4.5. If that fires, the position risk controls are mechanical: a 2.4% stop loss on entry price (roughly $216.30 from the current EMA level) and a 4.9% take profit (roughly $232.55), an effective reward-to-risk of about 2-to-1. The same conditions apply to the QQQ and SPY legs, where ADX (19.5 and 20.9) is also near or below the threshold. The completed backtest on this rule set over five years on the NVDA pair produced a 46.4% total return across 42 trades at a 50% win rate, with a worst drawdown of 13.9% — exits were filled on daily bars rather than intrabar data, so treat the drawdown and win rate as approximate. That track record supports waiting for the entry rather than anticipating it: the edge comes from taking the signaled trades, not pre-positioning. Note also that no robust parameter setup was established — the sensitivity evaluation exceeded its time budget — so the published thresholds are what you trade, nothing tuned.

The bull case: regime tailwinds and a rule set that has actually traded through them

The macro premise in the idea checks out against recent reporting. Per the Bloomberg piece from September 27, 2026 on tremors from AI to oil boosting the hedge fund dispersion trade, single-stock volatility is running hot relative to index volatility — exactly the environment this rules-based version is designed to monetize. A companion Bloomberg piece the same day, on AI whiplash jolting stocks as sentiment lurches from fear to greed, describes the fast sentiment reversals that keep single-name ranges wide while the index looks deceptively calm. The backtest evidence is realized, not hypothetical. Over a 60-month daily window the strategy traded 42 times, won 50% of those trades, returned 46.4%, with a worst peak-to-trough drawdown of 13.9%. Over the last 24 months it returned 23.0% across 35 trades (45.7% win rate, 12.5% max drawdown), and over the…

Scores

• Conviction score breakdown: 56
• Thesis support: 70
• Trade readiness: 35
• Risk quality: 55
• Backtest evidence: 65
• Fundamentals trend: 55

Watch items

• NVDA — Close vs 20-day EMA
• NVDA — ADX (14)
• NVDA — ATR (14)
• NVDA — Close vs second support
• SPY — ADX (14)
• QQQ — Close vs 20-day EMA