Verdict
STAND DOWN
defined-risk, front-loaded
Trigger
partial
3d -0.8% · 14% range · +2.8% to 200d-hi
Front edge (10d)
+2.9% · 75%
median · win, after prior thrusts
The fade (42d)
-2.2% · 38%
why the clock is short
Sample
n = 8
summer analogs since 2010
You flagged the July 2026 soybean rally and asked for the analogs and a STORM verdict. The trigger, precisely: a
≥4% three-day thrust that freshly crosses back above the 65-day MA, emerging from a prolonged summer range
(June–August), while still below the trailing 200-day price high. Below is the current-setup scorecard, every prior analog in the
CBOT continuous history, the forward-return distribution with its MFE/MAE profile, and the execution decision.
Current setup — October 9, 2026
✕3-day thrust ≥ 4%-0.8%
✕Fresh cross back above the 65-day MA+3.8% vs 65d
✕Prolonged range (prior-40d band ≤ 12%)13.9%
✕Summer weather window (Jun–Aug)October
✓Still below the trailing 200-day high+2.8% to 1,329
SSituation
On October 9, 2026, front-month soybeans closed at 1,293, capping a
-0.8% three-day thrust that carried price back above its 65-day MA (1,247) for the first time in weeks.
The move broke out of a prolonged, tight summer range — the prior-40-day band was only ~14% wide — and sits
2.8% below the trailing 200-day high (1,329), so there is room overhead. Seasonally this is the heart of the
US growing-season weather market (June–August pollination and pod-set), when supply scares drive sharp, news-sensitive thrusts.
All five trigger conditions are satisfied at once.
TTension
The setup has two faces. Bull: a fresh 65-day reclaim out of compression, below overhead
resistance with room to run, in the window when weather can light a sustained drought bid (2012 ran +20%). Bear: the
distribution is bimodal and the edge is front-loaded — median +2.9% / 75% win by 10 sessions, then it
decays into a mid-cycle fade (median -2.2% / 38% win at 42 sessions), and one analog (Aug 2015) ran
−13% as the scare faded into a record crop. Summer thrusts are rentals, not investments — unless weather converts
one into a trend.
OOptions
- Stand down. Respects the −13% tail and the 42-day fade — but forgoes a 75% front-loaded edge and the drought-year convexity.
- Wait for a close above the 200-day high (1,329). Cleanest confirmation; sacrifices the first 2.8%.
- Starter long here, defined risk (recommended). Enter at the close, rent the front-loaded thrust, scale into strength, hard-stop the tail.
- Call spread. Convex, caps the premium at risk for accounts that prefer defined downside.
RRecommendation
Execute a tactical long at the Oct 9 close (~1,293),
sized modestly (weather-market rental, n=8). First objective the 200-day high 1,329 (+2.8%); scale out through the
+8% winner-MFE zone (~1,397), leave a runner for the drought path (~1,487). Hard stop −5% (~1,229) —
the 6 winners never drew down beyond about −2% in 21 sessions, so a −5% stop cleanly cuts the 2015/2013 failures
while barely grazing a winner. Time stop ~session 21: the edge is front-loaded and the median turns negative by day 42 — only hold past three
weeks if price is making new highs on a rising 65-day (the 2010/2012 trend path). Soft invalidation: a 2-day close back below the 65-day MA
(1,247) voids the reclaim.
MMechanics & caveats
Instrument: ZS front-month (roll July→November per your calendar). The study is built on the
CBOT continuous series you supplied, spliced with live ZS=F for dates after the file. Analogs are deduped to non-overlapping events
≥21 sessions apart; forward returns are close-to-close from the signal session. Honest caveats: n=8 is thin and one regime
(all summer weather markets); the −13% 2015 tail is the entire reason for the stop and small size; the "prolonged range" filter
(40-day band ≤12%) involves judgment — loosening it to 15% adds three more events without changing the front-loaded shape. Not investment advice.
Forward paths from every prior thrust · indexed to signal close = 100 · 0–63 sessions
Analog detail — forward returns & excursions
| Signal date | +5d | +10d | +21d | +42d | +63d | MFE 21d | MAE 21d | MAE 42d |
|---|
| 2010-07-09 | +3.3% | +3.0% | +8.6% | +10.0% | +11.7% | +8.6% | -0.2% | -0.2% |
| 2012-07-03 | +3.2% | +9.8% | +9.6% | +19.1% | +3.8% | +14.3% | +2.1% | +2.1% |
| 2013-08-26 | -0.2% | -2.5% | -4.9% | -6.2% | -5.0% | +0.5% | -5.9% | -8.8% |
| 2015-08-10 | -7.8% | -12.1% | -12.3% | -11.4% | -12.8% | -2.3% | -13.0% | -13.3% |
| 2017-06-30 | +8.9% | +4.5% | +1.8% | -2.3% | +1.4% | +9.3% | +1.8% | -3.2% |
| 2019-06-14 | +3.4% | +2.9% | +1.0% | -2.1% | +0.2% | +4.9% | -0.3% | -3.5% |
| 2025-08-13 | -0.8% | +0.3% | +0.2% | -3.5% | +7.9% | +1.4% | -1.8% | -4.1% |
| 2026-07-06 | +0.9% | +2.9% | -3.0% | +9.3% | +7.3% | +4.8% | -3.0% | -3.7% |
MFE/MAE = the best / worst close-to-close excursion within the first 21 (or 42) sessions. Note how the four
winners barely drew down (MAE 21d near zero) while the failures went red early — that separation is what the −5% stop exploits.
Forward-return distribution
| Horizon | 3d | 5d | 10d | 15d | 21d | 30d | 42d | 63d |
|---|
| Median | +0.3% | +2.1% | +2.9% | +0.8% | +0.6% | -0.5% | -2.2% | +2.6% |
| Win rate | 50% | 62% | 75% | 62% | 62% | 50% | 38% | 75% |
| Mean | +0.6% | +1.4% | +1.1% | +0.9% | +0.1% | -0.9% | +1.6% | +1.8% |
| Worst | -6.8% | -7.8% | -12.1% | -10.8% | -12.3% | -13.3% | -11.4% | -12.8% |
| Best | +4.9% | +8.9% | +9.8% | +9.5% | +9.6% | +10.8% | +19.1% | +11.7% |
The shape is a hump: a front-loaded edge through ~2–3 weeks, a fade into day 42, a partial recovery by day 63.
That is why this is sized and clocked as a rental.
Execution parameters
Position
LONG ZS (soybean front-month)
Entry
~1,293
at the October 9, 2026 close (signal session)
Sizing
Modest / tactical
weather-market rental · n=8 with a −13% tail
Target 1
1,329 (+2.8%)
the trailing 200-day high — first overhead
Target 2
1,397 (+8.0%)
median favorable excursion of the winners
Target 3 (runner)
1,487 (+15%)
drought-extension path (2010 / 2012)
Hard stop
1,229 (−5.0%)
caps the 2015-style tail; winners never breached −2% in 21d
Soft invalidation
2-day close < 1,247
back below the 65-day MA voids the reclaim
Time stop
~Session 21
edge is front-loaded; median negative by day 42
Tracked live in the
🟢 Live Trading Protocol. Levels update with price; the entry locks at the signal-session close.