# Hold the yen thesis. Keep the entry gates intact.

Research cutoff: Sep 7, 2026, 7:30 p.m. Toronto. Status: archived.

Yen appreciation has reduced remaining upside. Platinum/gold crossed its numerical level, but asynchronous quotes and the pending WPIC report leave the trade unconfirmed.

## Market outlook
Resilient activity. Constrained energy supply.
Subjective research scenarios, not portfolio weights. Sugar short and platinum/gold both lose in the supply-shock scenario.

## What changed
- **Yen became more priced:** Priced-in score rose to 4/5 as December yen appreciated approximately 1.47% from the baseline.
- **Numerical crossing; no activation:** The indicative platinum/gold ratio exceeded 0.414, but WPIC and valid close/retest gates remained pending.
- **ULSD/RBOB approached its target:** At $1.4362/gal, only $0.0138 remained to the old $1.45 target.

## Trade theses

### Japanese yen — Hold thesis
The yen thesis strengthened, but appreciation consumed remaining upside. Japan’s revised GDP was due 20 minutes after the monitor cutoff. The September 7 decision was to hold without adding and wait for the catalyst.
**Contracts:** Long CME December 2026 yen · 6JZ26
**Quote:** 0.0065445 USD per JPY. CME ten-minute-delayed intraday feed, September 7 at 19:16:57 EDT. Retrieved at 19:26 EDT. No September 7 settlement.
**Change:** +1.47%. vs 0.006450
**Entry:** Add only at 0.00635–0.00639 with intact policy evidence and after the GDP release.
**Targets:** 0.00670–0.00680; unreached at the snapshot.
**Stop:** 0.00618. No breach identified from the quoted snapshot.
**Invalidation:** Adverse Japan/U.S. policy divergence undermining the yen-appreciation thesis.
**Counterargument:** Strong U.S. growth and rates, Japan’s energy-import burden, and a heavily anticipated BOJ hike limit additional upside.
**Catalyst:** Japan revised Q2 GDP: September 7, 19:50 EDT. BOJ policy and U.S. inflation remain material.
**Roll:** 6JZ26 last trade: December 14. Review a roll to 6JH27 before then; the January 6 horizon extends beyond December expiry.
**Market evidence:** Priced in rose from 3.5 to 4/5. Latest cited CFTC yen leveraged-fund net position was −102,188 as of September 1; the rally is not proof of measured short covering.
**Expected return:** September 6 modeled expectation: +2.5% of notional. September 7 mechanical remaining-value sensitivity: approximately +1.0–1.1%, keeping the prior scenario outcomes fixed. This is not a newly estimated forecast.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME yen quotes](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/69)
[Japan GDP calendar](https://www.esri.cao.go.jp/en/sna/kouhyou/kouhyou_top.html)
[CFTC financial futures](https://www.cftc.gov/dea/futures/financial_lf.htm)

### Sugar short — Wait
The only new conditional promotion from the September 6 full analysis. Crowded scarcity expectations may disappoint, but a fresh failed retest and physical confirmation are still required. The trade remains unactivated.
**Contracts:** Short ICE March 2027 Sugar No. 11 · SBH27
**Quote:** 19.05¢/lb. Vendor-reported September 4 settlement. Latest displayed trade: 18.98¢ at 12:59 EDT on September 4. ICE U.S. softs were closed September 7; this is not an unchanged Monday trade.
**Change:** Carried. no new Monday quote
**Entry:** A new failed retest of 19.20–19.40¢/lb. September 4’s high predates the thesis and cannot activate it retroactively.
**Targets:** 18.00¢, then 17.50¢/lb.
**Stop:** 20.15¢/lb after activation.
**Invalidation:** Material deterioration in exportable supply, worsening Brazilian production, or export restrictions.
**Counterargument:** Expensive oil can encourage ethanol production and reduce sugar output. Crowding alone does not justify a short.
**Catalyst:** Fresh sugar production/export evidence and the effect of expensive oil on ethanol allocation.
**Roll:** March 2027 contract: no roll required before the retained January 6 research horizon.
**Market evidence:** Priced in unchanged at 2/5. September 1 sugar managed-money net long: 233,771 contracts, up 35,754 weekly. No new CFTC report at the cutoff.
**Expected return:** September 6 modeled expectation: +2.6%; target/stop asymmetry 2.1:1 at the preferred reference entry. September 7 assessment: slightly worse qualitatively due to ethanol risk; model not recalculated.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[SBH27 quote page](https://www.wsj.com/market-data/quotes/futures/SBH27)
[CFTC agricultural futures](https://www.cftc.gov/dea/futures/ag_lf.htm)
[ICE holiday calendar](https://www.ice.com/publicdocs/futures/IFUS_Trading_Hours_Holiday_Calendar.pdf)

### Wheat / soybeans — Wait
The price is inside the proposed band, but the entry also requires post-WASDE export and basis confirmation. Soybean-specific sourcing risk adds a counterargument to the short soybean leg.
**Contracts:** Long 2 March 2027 wheat · ZWH27 / short 1 March 2027 soybeans · ZSH27
**Quote:** 749.25¢ / 1,330.25¢ per bushel. September 4 settlement benchmarks. Ratio = 2 × 749.25 ÷ 1,330.25 = 1.126480. No refreshed trade in the September 7 monitor.
**Change:** Carried. inside price band
**Entry:** 2W/S of 1.115–1.135, after WASDE and export/basis confirmation.
**Targets:** 2W/S of 1.18–1.20.
**Stop:** 2W/S of 1.08 after activation.
**Invalidation:** Normalized Black Sea shipping without deliverable-wheat demand, or soybean-specific supply damage.
**Counterargument:** Larger world wheat supply and soybean sourcing constraints can work against the relative thesis.
**Catalyst:** September 8 export inspections and Crop Progress; September 11 WASDE, beyond the 72-hour monitor window.
**Roll:** March wheat and soy first notice: February 26, 2027, beyond January 6. The old December/November expression was superseded.
**Market evidence:** Priced in unchanged at 2.5/5. September 1 managed-money net positions: wheat +14,904; soybeans +234,920. Contract counts specify the expression, not an allocation.
**Expected return:** September 6 modeled expectation: +2.3% of total absolute spread-leg notional; approximately 1.6:1 target/stop asymmetry. September 7 expected return was slightly worse qualitatively; no refreshed model.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME wheat](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/323)
[CME soybeans](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/320)
[USDA WASDE](https://www.usda.gov/about-usda/general-information/staff-offices/office-chief-economist/commodity-markets/wasde-report)
[FAO cereal balance](https://www.fao.org/worldfoodsituation/csdb/en/)

### Platinum / gold — Wait
Relative momentum improved and the indicative ratio exceeded 0.414. The WPIC report, a qualifying close, and a defended retest were still required. No trade was activated.
**Contracts:** Long 5 January 2027 platinum · PLF27 / short 1 February 2027 gold · GCG27
**Quote:** $1,860.2 / $4,488.7 per oz. September 7 delayed feeds: platinum 18:20:27 EDT; gold 19:16:45 EDT. The legs are approximately 56 minutes apart. PL price ÷ gold price = 0.414418, indicative and not an executable quote.
**Change:** +1.26%. vs 0.409276
**Entry:** Supportive WPIC report, a close above 0.414, then a defended 0.410–0.414 retest.
**Targets:** PL/G ratio of 0.445–0.460.
**Stop:** PL/G of 0.390 after activation.
**Invalidation:** Disappearance of the deficit and tight-stock premise.
**Counterargument:** A growth shock can favor gold over industrial platinum. The 5:1 package roughly matches initial dollars, not volatility or correlation.
**Catalyst:** WPIC Platinum Quarterly: September 9; exact release time unverified in the monitor.
**Roll:** PLF27 first notice: December 31; roll beforehand. February gold extends beyond the January 6 horizon.
**Market evidence:** Priced in rose from 3 to 3.5/5. Uneven quote timing limits confirmation. The prior December-gold expression was replaced with February gold, changing the ratio.
**Expected return:** September 6 modeled expectation: +1.2% of total absolute leg notional, approximately 1.7:1 asymmetry. The September 7 rally reduced remaining payoff; model not recalculated.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME platinum](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/446)
[CME gold](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/437)
[WPIC quarterly](https://platinuminvestment.com/supply-and-demand/platinum-quarterly)

### SOFR flattener — Wait
A half-basis-point flattening versus the baseline did not provide fresh policy confirmation. The old event-entry setup remains cancelled; the structural idea is only a watch.
**Contracts:** Long March 2027 SOFR · SR3H27 / short December 2026 SOFR · SR3Z26
**Quote:** 95.805 / 95.940. September 7 delayed feeds: March 19:15:21 EDT; December 19:11:59 EDT. March-minus-December implied-rate slope: 13.5 bp. No September 7 settlement.
**Change:** −0.5 bp. vs 14.0 bp
**Entry:** Post-data flattening confirmation. No entry from an overshoot alone; no complete new numeric entry band supplied.
**Targets:** A refreshed target was not specified in the latest supplied monitor.
**Stop:** 16 bp implied-rate slope after activation.
**Invalidation:** Failure of post-data policy evidence to support flattening; the old payroll catalyst setup had already failed.
**Counterargument:** Resilient U.S. growth and inflation risk weaken the immediate easing or growth-break case.
**Catalyst:** U.S. PPI September 10; CPI September 11; FOMC September 16.
**Roll:** The December 2026/March 2027 package must be reassessed as the front leg approaches its settlement period; no unchanged four-month hold assumed.
**Market evidence:** Priced in unchanged at 2/5. Asynchronous leg marks; no new policy or package-liquidity confirmation.
**Expected return:** Prospective return was marginally worse after the small move. No updated numeric four-month estimate supplied.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME SOFR](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/8462)
[BLS release schedule](https://www.bls.gov/schedule/2026/09_sched.htm)
[FOMC calendar](https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm)

### Natural gas calendar — Wait
The curve changed little and the premium remained below the required $1.10. Winter and physical-market evidence are still required before a calendar fade.
**Contracts:** Short January 2027 natural gas · NGF27 / long March 2027 · NGH27
**Quote:** $3.910 / $2.912 per MMBtu. September 7 delayed feeds: January 19:00:03 EDT; March 18:04:21 EDT. January-minus-March premium: $0.998/MMBtu, indicative. March is a stale leg.
**Change:** +$0.003. vs $0.995 premium
**Entry:** January-minus-March premium of at least $1.10/MMBtu, with supportive winter and physical evidence.
**Targets:** No refreshed numeric target specified in the latest supplied monitor.
**Stop:** Premium of $1.35/MMBtu after activation.
**Invalidation:** Winter or physical-market conditions that support a sustained January scarcity premium.
**Counterargument:** Winter weather and physical shortages can overwhelm apparently comfortable aggregate inventory.
**Catalyst:** EIA natural-gas storage September 10, 10:30 EDT, plus winter and physical-market evidence.
**Roll:** NGF27 last trade: December 29. Review exit or roll during December 18–23 rather than carrying the January leg unchanged.
**Market evidence:** Priced in unchanged at 3/5. Curve broadly unchanged. The stale March quote limits interpretation of the package.
**Expected return:** Flat to slightly worse qualitatively. A small entry-price improvement was offset by winter risk; no updated numeric model.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME natural gas](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/444)
[EIA storage calendar](https://ir.eia.gov/ngs/schedule.html)

### ULSD / RBOB — Wait
The relative trend strengthened, but the difference reached just $0.0138 below the original $1.45 target. The remaining payoff deteriorated sharply; no chase.
**Contracts:** Long January 2027 ULSD · HOF27 / short January 2027 RBOB · RBF27
**Quote:** $4.1787 / $2.7425 per gallon. Both September 7 delayed feeds stamped 19:16:45 EDT. Difference: $1.4362/gal. Same-second feeds improve comparability but do not establish executable package depth.
**Change:** +$0.0816. vs $1.3546
**Entry:** A verified pullback toward $1.12–$1.18/gal.
**Targets:** $1.45/gal; only $0.0138 away at the snapshot.
**Stop:** $1.03/gal after activation.
**Invalidation:** Relative distillate normalization that removes the tightness supporting ULSD over RBOB.
**Counterargument:** Much physical tightness is already priced; chasing close to the target leaves little compensation for reversal risk.
**Catalyst:** Holiday-adjusted EIA petroleum report: September 10 at 12:00 EDT.
**Roll:** January HO/RB last trade: December 31. Exit or roll before the retained January horizon.
**Market evidence:** Priced in rose from 4 to 4.5/5. Matching timestamps do not prove spread liquidity or a verified change in daily correlation.
**Expected return:** Expected return deteriorated sharply near the old target. No new four-month numeric estimate was supplied.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[CME ULSD](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/426)
[CME RBOB](https://www.cmegroup.com/CmeWS/mvc/quotes/v2/429)
[EIA petroleum schedule](https://www.eia.gov/petroleum/supply/weekly/schedule.php)

### VIX calendar — Wait
No new calendar-specific edge emerged. The latest official benchmark remains outside the preferred entry band, and the unchanged package does not fit the four-month horizon.
**Contracts:** Long October 2026 VIX · VXV26 / short December 2026 VIX · VXZ26
**Quote:** 18.1384 / 18.9750. September 4 official Cboe settlements. October-minus-December = −0.8366. The monitor also recorded untimestamped indicative marks implying −0.82; those are not used as the primary quote.
**Change:** Carried. official benchmark
**Entry:** A verified spread of −1.05 to −1.20 points, with a calendar-specific edge.
**Targets:** No refreshed numeric target specified in the supplied monitor.
**Stop:** −2.00 points after activation.
**Invalidation:** Failure to establish a calendar-specific edge or a structure compatible with the intended horizon.
**Counterargument:** Event timing, volatility repricing, and roll costs can outweigh an apparent calendar discount.
**Catalyst:** Inflation and central-bank event timing must support the specific October/December package.
**Roll:** October VIX expires October 21; December expires December 16. The package cannot remain unchanged to January 6.
**Market evidence:** Priced in unchanged at 3.5/5. Untimestamped intraday marks lack adequate auditability for an actionable entry.
**Expected return:** Flat to lower qualitatively; no updated numeric expectation was supplied.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[Cboe official settlements](https://www.cboe.com/markets/us/futures/market-statistics/settlement/futures/daily)
[Cboe VIX futures](https://www.cboe.com/tradable-products/vix/vix-futures/)

### OAT / Bund — Wait
Exact December contracts were identified, but cheapest-to-deliver, DV01 matching, and basis remain unverified. No tradable hedge ratio or expected return can be inferred from the unmatched prices.
**Contracts:** December 2026 Euro-OAT · FOATZ26 / December 2026 Euro-Bund · FGBLZ26; tightening watch, hedge ratio unverified
**Quote:** OAT 114.31 / Bund 121.78. Vendor September 7 latest closes; displayed times 15:59:57 / 16:03:25 with timezone unspecified. These are separate futures prices, not a verified spread.
**Change:** −0.43 / −0.41. separate leg changes
**Entry:** Proper DV01 matching and verified reversal evidence. The expression is not ready for execution.
**Targets:** Not estimable from unmatched prices.
**Stop:** Original yield-spread conditions cannot be tested using unmatched futures prices.
**Invalidation:** Failure of the properly specified tightening thesis; definition awaits CTD/DV01 and basis verification.
**Counterargument:** CTD changes, basis, and incorrect duration matching may dominate the apparent country-spread thesis.
**Catalyst:** ECB September 10: decision at 08:15 EDT, press conference at 08:45 EDT.
**Roll:** Legacy September last trading: September 8. The current watch uses December; December last trade December 8 requires a roll for the January horizon.
**Market evidence:** Not rated. Leg changes of −0.43 and −0.41 are not portfolio or spread P&L.
**Expected return:** Not estimable until the futures spread and hedge are specified.
**Data limitations:** Historical September 7 research; observations retain their original dates and have not been freshly verified.
[OAT vendor history](https://www.investing.com/rates-bonds/euro-oat-historical-data?cid=1181013)
[Bund vendor history](https://www.investing.com/rates-bonds/euro-bund-historical-data?cid=1181008)
[Eurex specifications](https://www.eurex.com/ex-en/markets/int/long-term-interest-rates/fix/government-bonds/Euro-OAT-Futures-34684)
[ECB calendar](https://www.ecb.europa.eu/press/calendars/weekly/html/index.en.html)

## Retired views
### Crude normalization
Physical normalization unconfirmed. The front-leg short remains exited.
September 7 delayed legs: 92.66 / 86.37 at 19:16:59 / 19:16:56 EDT. Indicative spread $6.29, up $0.27 from $6.02 and above the old $6.25 invalidation. This was not an executed stop.
### Dow / Nasdaq rotation
Relative confirmation weakened. Keep the rotation thesis exited.
September 7 delayed marks: 53,104 / 29,536.00. Indicative ratio 1.797941, approximately −0.53% versus the baseline. No automatic reversal.
### Cocoa long
Prior exit remains in force. No new evidence justifies reinstatement.
September 6 review benchmark: vendor-reported September 4 settlement 6,188. No fresh September 7 quote was supplied. The prior rebound reduced entry value while supply and weather evidence conflicted.
### CAD / EUR
Keep the thesis retired. No replacement or reverse trade established.
September 6 review quoted 0.72585 / 1.16545 without a comparable verified prior cross. No fresh September 7 quote or actionable reinstatement was supplied.

## Catalyst calendar
- **07 SEP 19:50: Japan revised Q2 GDP.** Pending at the September 7 cutoff; outcome not incorporated [Source](https://www.esri.cao.go.jp/en/sna/kouhyou/kouhyou_top.html)
- **08 SEP 11:00: USDA export inspections.** Holiday-delayed release in supplied monitor [Source](https://www.ams.usda.gov/market-news/grain)
- **08 SEP 16:00: USDA Crop Progress.** Holiday-delayed release [Source](https://www.nass.usda.gov/Publications/Calendar/reports_by_date.php?month=09&year=2026)
- **09 SEP Unverified: WPIC Platinum Quarterly.** Physical-balance confirmation required [Source](https://platinuminvestment.com/supply-and-demand/platinum-quarterly)
- **10 SEP 08:15 / 08:45: ECB decision / press conference.** European rates and FX policy expectations [Source](https://www.ecb.europa.eu/press/calendars/weekly/html/index.en.html)
- **10 SEP 08:30: U.S. Producer Price Index.** Inflation and policy-expectation risk [Source](https://www.bls.gov/schedule/2026/09_sched.htm)
- **10 SEP 10:30: EIA natural-gas storage.** Storage balance and winter premium [Source](https://ir.eia.gov/ngs/schedule.html)
- **10 SEP 12:00: EIA petroleum report.** Holiday-adjusted release time [Source](https://www.eia.gov/petroleum/supply/weekly/schedule.php)

## Portfolio results
This edition predates the simulated portfolio, or its matching historical valuation is unavailable. No portfolio values have been substituted.