| Date | META D0 | QQQ D0 | META +5d | META +10d | META +20d | Ratio +5d | Ratio +10d | Ratio +20d |
|---|---|---|---|---|---|---|---|---|
| 2012-05-21 | -10.99% | +2.80% | -15.3% | -24.0% | -6.2% | -15.7% | -22.2% | -9.0% |
| 2012-06-29 | -0.83% | +3.15% | +3.4% | -9.2% | -25.6% | +3.7% | -7.8% | -26.3% |
| 2012-07-27 | -11.69% | +2.37% | -11.1% | -8.0% | -18.1% | -12.0% | -10.8% | -22.2% |
| 2012-11-19 | -2.72% | +2.38% | +14.1% | +19.8% | +20.9% | +12.0% | +16.4% | +15.9% |
| 2016-02-16 | -0.39% | +2.25% | +3.8% | +8.1% | +8.9% | +2.4% | +2.4% | +2.4% |
| 2020-04-02 | -0.88% | +2.03% | +10.7% | +13.3% | +27.9% | +2.6% | -2.1% | +11.8% |
| 2022-06-21 | -4.09% | +2.52% | +2.3% | +8.1% | +16.6% | +1.4% | +5.2% | +8.1% |
| 2022-10-21 | -1.16% | +2.34% | -23.7% | -30.2% | -13.8% | -25.3% | -27.3% | -16.7% |
| 2023-05-05 | -0.32% | +2.13% | +0.4% | +5.5% | +16.6% | -0.2% | +1.3% | +6.1% |
| 2024-11-06 | -0.07% | +2.72% | +1.4% | -1.1% | +6.4% | +0.1% | -0.7% | +3.1% |
| 2026-02-06 | -1.31% | +2.11% | -3.3% | -3.7% | -2.1% | -2.0% | -2.3% | -1.8% |
| 2026-05-08 | -1.16% | +2.34% | +0.8% | +0.1% | -4.0% | +1.1% | -0.8% | -4.6% |
| 2026-06-11 | -0.45% | +3.38% | +1.6% | -3.1% | +15.6% | -1.6% | -1.8% | +16.4% |
| 2026-07-30 | -7.95% | +3.30% | +9.4% | +10.4% | +5.9% | +4.7% | +3.1% | +0.4% |
| 2026-08-04 | -0.39% | +3.40% | +1.9% | -7.5% | +nan% | +2.7% | -6.7% | +nan% |
| Horizon | N | META Mean | META Median | META Win% | META Edge | QQQ Mean | QQQ Win% | Ratio Mean | Ratio Win% | Ratio Edge |
|---|---|---|---|---|---|---|---|---|---|---|
| Day 1 | 15 | -0.3% | -0.2% | 0.5 | -0.4% | +0.3% | 0.7 | -0.6% | 0.4 | -0.6% |
| Day 3 | 15 | +1.9% | +1.2% | 0.8 | +1.6% | +1.7% | 0.8 | +0.2% | 0.7 | +0.1% |
| Day 5 | 15 | -0.2% | +1.6% | 0.7 | -0.8% | +1.5% | 0.7 | -1.8% | 0.6 | -1.9% |
| Day 10 | 15 | -1.4% | -1.1% | 0.5 | -2.5% | +2.0% | 0.5 | -3.6% | 0.3 | -3.9% |
| Day 15 | 15 | +0.3% | -0.5% | 0.5 | -1.4% | +3.2% | 0.7 | -2.9% | 0.4 | -3.4% |
| Day 20 | 14 | +3.5% | +6.2% | 0.6 | +1.2% | +4.6% | 0.9 | -1.2% | 0.6 | -1.8% |
| Day 30 | 13 | +1.9% | +4.2% | 0.5 | -1.6% | +5.9% | 0.8 | -3.9% | 0.5 | -5.0% |
| Day 40 | 13 | +3.7% | +4.8% | 0.5 | -0.9% | +7.1% | 0.8 | -3.5% | 0.4 | -4.9% |
| Day 50 | 13 | +3.3% | +3.0% | 0.6 | -2.6% | +7.0% | 0.8 | -3.9% | 0.5 | -5.6% |
Two mechanical scale-out overlays on the same trigger. Enter 100% within two days of the signal, then manage the position in tranches — 20% / 30% / 30%, with the final 20% always exiting at the end of the study window.
Where the deviations come from. For every completed historical precedent we measure how far the trade travelled in your favor (its Maximum Favorable Excursion). The three profit deviations span that favorable-outcome distribution — a low, median and high reading of how far these trades historically ran. (For thin samples these are the literal weakest / typical / strongest outcomes; for larger samples they are robust quartiles, so one outlier can't set the rungs.) The three loss deviations are the exact mirror of the profit deviations — the same levels flipped below the entry, so a +4% first take-profit pairs with a −4% first stop. It's a symmetric bracket. A rung only fills if price actually trades there; unfilled tranches exit with the balance at the window close. Mirrored for shorts (the profit-taker buys the name back lower; the loss-taker covers higher).
⚠️ Small-sample caveat. With only a handful of completed precedents these levels are illustrative, in-sample anchors — a disciplined way to turn "what this pattern historically did" into entry, take-profit and stop levels, not a statistically validated system. The live bracket is the payoff, not the backtest averages.
| Rung | Return | Price | Action | Live status |
|---|---|---|---|---|
| Profit · 3rd dev | +15.0% | 0.944 | sell 30% | pending |
| Profit · 2nd dev | +7.0% | 0.879 | sell 30% | pending |
| Profit · 1st dev | +5.1% | 0.863 | sell 20% | pending |
| ● Entry | 0.0% | 0.821 | buy 100% | Day 18 |
| Loss · 1st dev | -5.1% | 0.779 | cut 20% | ● cut |
| Loss · 2nd dev | -7.0% | 0.763 | cut 30% | ● cut |
| Loss · 3rd dev | -15.0% | 0.698 | cut 30% | pending |
| Signal date | Peak (MFE) | Trough (MAE) | Buy & hold | Profit-taker | P-devs | Loss-taker | L-devs |
|---|---|---|---|---|---|---|---|
| 2012-05-21 | +7.0% | -37.4% | -37.4% | -15.6% | ●●○ | -15.1% | ●●● |
| 2012-06-29 | +5.1% | -45.5% | -36.2% | -28.0% | ●○○ | -14.9% | ●●● |
| 2012-07-27 | +0.0% | -27.1% | -16.0% | -16.0% | ○○○ | -10.8% | ●●● |
| 2012-11-19 | +32.8% | +0.0% | +16.1% | +10.9% | ●●● | +16.1% | ○○○ |
| 2016-02-16 | +6.7% | -3.7% | +6.7% | +6.4% | ●○○ | +6.7% | ○○○ |
| 2020-04-02 | +21.7% | -1.7% | +15.4% | +10.7% | ●●● | +15.4% | ○○○ |
| 2022-06-21 | +8.8% | -9.2% | -0.5% | +2.9% | ●●○ | -3.4% | ●●○ |
| 2022-10-21 | +3.9% | -28.1% | +3.9% | +3.9% | ○○○ | -6.8% | ●●● |
| 2023-05-05 | +15.0% | -0.5% | +11.3% | +9.9% | ●●● | +11.3% | ○○○ |
| 2024-11-06 | +5.3% | -4.1% | +3.6% | +3.9% | ●○○ | +3.6% | ○○○ |
| 2026-02-06 | +0.0% | -15.3% | -6.6% | -6.6% | ○○○ | -9.0% | ●●● |
| 2026-05-08 | +13.0% | -9.8% | +4.3% | +5.3% | ●●○ | -1.0% | ●●○ |
| 2026-06-11 | +20.7% | -3.6% | +2.0% | +8.0% | ●●● | +2.0% | ○○○ |
The study isolates sessions where the Nasdaq-100 (QQQ) closes up more than 2% on the day while Meta Platforms (META) closes in negative territory — a rare divergence given META's historical beta to QQQ of ~1.1–1.3. Over the full sample (May 2012 – present), this condition has occurred 15 times, most recently on Aug 04, 2026.
META mean-reverts toward QQQ. In the 10 trading days following the signal, META posts a median return of -1.1% with a 33% win rate on the META/QQQ spread — meaning in 33% of historical instances, META outperformed QQQ over the following 10 days. The conditional edge over the unconditional base rate is -2.5% at the 10-day horizon.
The spread recovery is most pronounced in the first 5–15 trading days, after which it gradually normalizes toward the base rate. The 20-day ratio win rate is 57%, suggesting the edge persists but diminishes beyond the first two to three weeks.
Long META / Short QQQ spread is the mechanical expression of this signal. The P&L driver is not directional market exposure but the relative reversion of META back toward its expected co-movement with QQQ. Key considerations:
With 15 instances the sample is meaningful but not large. Standard errors on the win rates are approximately ±26 percentage points at the 95% confidence level. The pattern is directionally consistent across sub-periods but individual outcomes show wide dispersion — see the distribution chart. Past conditional probabilities do not guarantee future returns.