| MDY ATH date | SPY dd | MDY +21d | MDY +252d | SPY +252d | Ratio +63d | Ratio +252d |
|---|---|---|---|---|---|---|
| 2026-08-12 | -0.1% | — | — | — | — | — |
| 2026-06-04 | -0.3% | +0.3% | — | — | — | — |
| 2026-04-20 | -0.2% | -2.6% | — | — | -2.3% | — |
| 2026-02-06 | -0.7% | -4.4% | — | — | -3.6% | — |
| 2026-01-05 | -0.4% | +3.1% | — | — | +5.4% | — |
| 2024-11-22 | -0.5% | -5.4% | -0.2% | +15.5% | -6.7% | -13.6% |
| 2024-10-16 | -0.3% | +1.6% | +3.4% | +16.7% | -1.2% | -11.4% |
| 2024-03-28 | -0.0% | -4.3% | -2.4% | +8.6% | -7.5% | -10.1% |
| 2024-02-12 | -0.0% | +4.7% | +14.2% | +23.3% | +1.5% | -7.3% |
| 2021-10-22 | -0.1% | +2.9% | -15.6% | -15.1% | -2.7% | -0.6% |
| 2021-04-14 | -0.3% | -0.1% | -2.2% | +7.9% | -6.9% | -9.4% |
| 2021-02-09 | -0.1% | +4.0% | +6.8% | +17.1% | +1.1% | -8.8% |
| 2020-12-15 | -0.2% | +6.0% | +21.9% | +29.1% | +7.6% | -5.6% |
| 2020-11-09 | -0.5% | +7.5% | +40.3% | +33.7% | +9.8% | +5.0% |
| 2020-02-13 | -0.1% | -36.3% | +23.0% | +18.7% | -11.7% | +3.7% |
| 2020-01-14 | -0.2% | +1.3% | +18.7% | +18.2% | -14.3% | +0.5% |
| 2018-07-09 | -2.2% | +0.4% | -1.8% | +9.6% | -5.6% | -10.3% |
| 2018-06-06 | -2.8% | +0.1% | -3.6% | +5.7% | -1.8% | -8.8% |
| Signal date | 15d thrust | Ratio +63d | Ratio +126d | Ratio +252d | MDY +252d |
|---|---|---|---|---|---|
| 2026-06-10 | +4.2% | — | — | — | — |
| 2024-07-24 | +4.5% | -2.8% | -2.8% | -9.0% | +8.4% |
| 2023-12-04 | +4.7% | -0.5% | -4.8% | -4.8% | +28.4% |
| 2022-04-29 | +4.1% | +0.8% | +3.1% | -1.7% | -0.3% |
| 2022-02-18 | +4.0% | +1.0% | +0.6% | +7.6% | +0.4% |
| 2020-10-15 | +4.2% | +11.7% | +12.9% | +6.3% | +38.4% |
| 2020-07-30 | +4.2% | +0.1% | +8.8% | +5.7% | +45.2% |
| 2020-05-26 | +4.6% | -3.6% | +2.9% | +8.9% | +54.8% |
| 2016-11-17 | +4.1% | +0.6% | -2.4% | -3.4% | +16.1% |
| 2009-12-22 | +4.2% | +3.1% | +4.6% | +10.0% | +26.2% |
| 2008-12-22 | +4.3% | +2.3% | +5.3% | +10.0% | +44.5% |
| 2008-02-04 | +4.2% | +1.8% | +6.0% | +1.0% | -37.2% |
| 2000-12-08 | +5.8% | +3.7% | +7.2% | +12.2% | -4.1% |
| 2000-05-02 | +4.7% | +3.9% | +9.9% | +22.9% | +9.3% |
| 2000-01-06 | +4.4% | +5.4% | +9.6% | +23.4% | +17.0% |
| 1999-04-27 | +4.0% | +5.4% | +2.6% | +11.1% | +21.2% |
| 1999-01-04 | +4.6% | -12.5% | -4.6% | -5.3% | +13.2% |
| 1998-10-29 | +6.2% | -1.9% | -2.0% | -3.5% | +22.2% |
| Horizon | N | MDY med | win | SPY med | win | Ratio med | win |
|---|---|---|---|---|---|---|---|
| 10d | 85 | +0.2% | 54% | +0.3% | 56% | -0.2% | 40% |
| 21d | 84 | +1.0% | 60% | +1.0% | 62% | -0.2% | 46% |
| 42d | 84 | +1.9% | 65% | +1.9% | 71% | -0.3% | 42% |
| 63d | 83 | +2.1% | 63% | +3.3% | 67% | -0.7% | 42% |
| 126d | 82 | +5.0% | 70% | +4.7% | 76% | -1.3% | 39% |
| 252d | 80 | +8.6% | 74% | +11.2% | 81% | -2.3% | 30% |
| Horizon | N | Ratio med | win | MDY med | win |
|---|---|---|---|---|---|
| 10d | 18 | +0.4% | 72% | +1.9% | 61% |
| 21d | 18 | -0.2% | 50% | +1.8% | 61% |
| 42d | 18 | +0.4% | 56% | +2.6% | 78% |
| 63d | 17 | +1.0% | 71% | +4.7% | 76% |
| 126d | 17 | +3.1% | 71% | +8.7% | 76% |
| 252d | 17 | +6.3% | 65% | +17.0% | 82% |
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 | +4.4% | 0.975 | sell 30% | pending |
| Profit · 2nd dev | +2.9% | 0.961 | sell 30% | pending |
| Profit · 1st dev | +1.9% | 0.951 | sell 20% | ● filled |
| ● Entry | 0.0% | 0.934 | buy 100% | Day 55 |
| Loss · 1st dev | -1.9% | 0.916 | cut 20% | ● cut |
| Loss · 2nd dev | -2.9% | 0.907 | cut 30% | ● cut |
| Loss · 3rd dev | -4.4% | 0.892 | cut 30% | pending |
| Signal date | Peak (MFE) | Trough (MAE) | Buy & hold | Profit-taker | P-devs | Loss-taker | L-devs |
|---|---|---|---|---|---|---|---|
| 1998-10-29 | +3.7% | -2.6% | -2.5% | +0.0% | ●●○ | -2.4% | ●○○ |
| 1999-01-04 | +0.0% | -11.9% | -9.4% | -9.4% | ○○○ | -4.5% | ●●● |
| 1999-04-27 | +5.6% | +0.0% | +3.9% | +3.4% | ●●● | +3.9% | ○○○ |
| 2000-01-06 | +14.8% | -0.2% | +4.7% | +3.5% | ●●● | +4.7% | ○○○ |
| 2000-05-02 | +4.1% | -0.4% | +2.6% | +2.5% | ●●○ | +2.6% | ○○○ |
| 2000-12-08 | +4.9% | -3.3% | +4.4% | +3.5% | ●●● | +1.0% | ●●○ |
| 2008-02-04 | +1.9% | -2.7% | +1.2% | +1.4% | ●○○ | +0.6% | ●○○ |
| 2008-12-22 | +4.4% | -0.1% | +3.3% | +3.2% | ●●● | +3.3% | ○○○ |
| 2009-12-22 | +3.7% | -0.9% | +3.6% | +3.0% | ●●○ | +3.6% | ○○○ |
| 2016-11-17 | +2.6% | -0.1% | +0.1% | +0.5% | ●○○ | +0.1% | ○○○ |
| 2020-05-26 | +1.7% | -6.4% | -4.6% | -4.6% | ○○○ | -3.5% | ●●● |
| 2020-07-30 | +2.5% | -3.6% | +1.8% | +1.8% | ●○○ | -0.3% | ●●○ |
| 2020-10-15 | +12.2% | -0.0% | +10.9% | +4.8% | ●●● | +10.9% | ○○○ |
| 2022-02-18 | +1.3% | -3.6% | +1.2% | +1.2% | ○○○ | -0.7% | ●●○ |
| 2022-04-29 | +2.0% | -2.0% | +0.6% | +0.9% | ●○○ | +0.1% | ●○○ |
| 2023-12-04 | +2.9% | -3.2% | -0.1% | +1.2% | ●●○ | -1.3% | ●●○ |
| 2024-07-24 | +1.3% | -4.8% | -2.6% | -2.6% | ○○○ | -3.1% | ●●● |
This study tests two distinct signals on the same instrument. Idea 1 fires when the S&P MidCap 400 (MDY) closes at a fresh all-time high while the S&P 500 (SPY) sits below its own high — a mid-cap-leadership "poke." Idea 2 fires when the MDY/SPY ratio reclaims its 65-day moving average and is simultaneously up more than 4% over the trailing 15 sessions — a relative-momentum "thrust." They reach opposite conclusions, and the reason is the most useful thing on this page.
The breadth poke is genuinely bullish — but for the whole market, not for mid-caps specifically. Twelve months out, MDY is higher 80% of the time (median +8.6%) and SPY is higher 81% of the time (median +11.2%). Healthy broadening tends to precede broad gains. But SPY out-gains MDY: the MDY/SPY ratio's 12-month median is -2.3% with just a 30% win rate. The per-event log makes it concrete — the ratio's one-year return is negative in nearly every recent instance. The intuitive read ("mid-caps are leading, so buy MDY versus SPY") is backwards: a single-day new high is often an isolated poke near the top of a range that mean-reverts. Trade it long the market, not long the spread.
As literally specified — a same-day cross above the 65d MA and +4%/15d — the setup has occurred only twice in 30 years (Apr 1999, Jul 2024); n=2 is untradeable, because by the time the ratio is up 4% it has usually crossed the MA days earlier. The faithful tradeable read (reclaim within the last ~10 sessions, still thrusting) gives 18 events, and here the ratio momentum does carry: 12-month median +6.3%, win rate 65%, versus a roughly flat-to-negative baseline.
Be clear about what the edge is, though. Stripping the 65d-MA filter entirely leaves the result unchanged — thrust alone (n=38) prints +6.0% at 12 months versus +6.0% for above-MA-plus-thrust, and thrust while below the MA (n=5) is actually the best at +8.3%. So the moving-average reclaim is confirmation, not cause: the signal is the 15-day relative thrust. We keep the MA in the headline definition because it is what you asked for and it improves entry timing slightly, but the page would be dishonest if it implied the MA earns its keep.
Same ratio, opposite results. Idea 1 conditions on a point (one new-high print); Idea 2 conditions on a path (15 sessions of accumulation). The point catches isolated pokes that revert; the path catches genuine regime shifts that continue. This is the general lesson worth more than either win rate: the signal lives in the conditioning, not the instrument.
The +6%-ish aggregate edge on Idea 2 is not evenly earned — it is a bet on a small/mid-cap-favorable regime. The era split is stark: the thrust signal printed +11% to +23% across 1999–2001 (post-bubble value rotation) and +6% to +9% through 2020, but essentially every signal from 2021 through 2024 lost — −8.9% (Mar 2021), −9.3% (Oct 2022), −4.8% (Dec 2023), −9.0% (Jul 2024), −14.6% (Nov 2024) on a 12-month basis. Strip 1999–2001 and 2020 and the edge largely disappears. The live June 2026 signal is firing immediately after the worst losing streak this setup has ever had, during the tail of a mega-cap-dominated tape. Treat Idea 2 as a conditional trade — worth taking only when mid/small-cap breadth is genuinely expanding, not as an always-on signal.
MDY closed at a new all-time high on Aug 12, 2026 with SPY about 1.4% below its own high (Idea 1 active). The ratio's recent-reclaim-plus-thrust variant fired Jun 10, 2026, with the ratio up -2.0% over the trailing 15 sessions and sitting at 0.896 (Idea 2 active). Per the regime caveat above, the honest verdict is constructive on the market (Idea 1) but only conditionally constructive on mid-cap relative outperformance (Idea 2) — the historical edge is real but has been absent for three years.