Treasuries hedged the AI scare; gold did not.
The 10-year yield fell below 5% as technology shares weakened, yet gold stayed near $4,022 spot and April one-ounce futures slipped to a $4,090.75 bid. That split favors patience over adding another “safe haven.”
Treasuries hedged the AI scare; gold did not.
Research cutoff: 2026-10-23T04:06:00.000Z. Comparison cutoff: 2026-10-22T22:10:00.000Z. The four-month horizon remains 18 October 2026–18 February 2027. All eighteen board rows, four open simulated positions and zero pending orders were reviewed.
Decision: hold six March micro soybean-oil futures, five March micro-corn futures, two March micro natural-gas futures and three April one-ounce gold futures. Add no risk. Wait on yen, Treasuries, oil and equity-index challengers until the named evidence improves. No assumed paper transaction is justified in this run.
The latest risk-off move split the havens
The technology selloff pushed investors toward Treasuries and haven currencies, but not gold. Reuters recorded the U.S. 10-year yield at 4.9591% after a fall of more than 19 basis points—the largest daily Treasury rally in more than three years—while spot gold was near $4,021.99/oz after an almost 3% overnight decline. The dollar was 156.44 yen and 0.8196 Swiss franc. Reuters global markets, 23 October
That is more than a generic “risk-off” label. The market treated Treasuries as protection against an AI-investment growth scare, while gold remained exposed to oil-linked inflation, policy expectations and liquidation. It lowers tactical conviction in gold without automatically invalidating a deliberately small four-month position.
The rates signal is also not one-directional. Reuters estimated an 85% probability of a December Fed hike—lower than fully priced, but still high—while Brent reached roughly $114. Falling yields and rising oil are pulling duration in opposite directions. Treasuries are the cleaner hedge today, but not yet a clean new entry.
Japan: inflation breadth improved, growth did not collapse
Japan’s official September CPI was 2.1% year over year. Reuters reported core CPI at 2.1%, ex-fresh-food-and-energy inflation at 2.4%, and services inflation at 1.5%. The release supports a near-term Bank of Japan hike but matched consensus, so it was not a new yen surprise. Statistics Bureau of Japan, Reuters Japan CPI
Flash PMI was mixed: manufacturing slowed to 53.3 from 54.1, while services improved to 52.0 from 51.3. This is slower industrial momentum, not a broad contraction. S&P Global Japan PMI
The yen is the strongest FX challenger, but it already gained roughly 1% on the week and carries negatively against the dollar. A spot-equivalent move from USD/JPY 156.28 to 158.28 represents about $101 adverse value per Micro JPY contract before costs; sizing is feasible. The unresolved issue is price compensation after the haven rally, not contract granularity. Wait for Friday’s global price surveys and continued risk aversion; no paper order is staged.
Gold: hold the small starter; do not add
Exact April one-ounce gold was $4,090.75 bid / $4,091.75 ask at 04:01:42 UTC. The $4,098.50 last was stale, so the bid is the conservative mark. Versus the $4,177 assumed entry, the three-contract simulated position is down about $258.75 gross. It retains approximately $512 of mark-to-stop risk and $1,528 of mark-to-target upside—about 2.98 times remaining stop risk.
The curve also matters. April’s midpoint was about $65.88/oz above December, roughly 1.64% over four months. That is carry and basis, not a scarcity signal. Thin April volume reinforces the decision not to add.
Using the shared scenario model once, 20%/$4,600, 45%/$4,180 and 35%/$3,800 endpoints give +$110 after modeled exit commissions across three ounces; shifting 10 probability points from bull to bear gives −$130. These are judgments, not measured probabilities. Structural ETF and reserve demand still supports small optionality, but the weak response to falling yields and a softer dollar demands discipline. Hold three; no add. Keep the $3,920 stop, $4,600 upside target and February 18 exit. World Gold Council ETF flows, CFTC metals positioning
Natural gas: supportive storage, unhelpful price
The October 22 EIA report remains the essential physical evidence: U.S. storage was comfortable nationally, while South Central and salt inventories were below normal. That regional shortfall preserves winter upside. But same-month NGH27 was only $2.517/MMBtu at 03:59:18 UTC, lower despite the data, and NOAA’s eastern warmth outlook remains a demand headwind. EIA weekly storage, NOAA/CPC outlook
NGH27 is an identical-unit, zero-basis standard-contract proxy for MNGH27—not an observed micro quote. The shared 20%/$3.60, 45%/$2.85 and 35%/$2.00 scenarios give +$364 after modeled exit commissions across two micros; the 10-point adverse sensitivity leaves +$44. Hold two; no add. Keep the $2.00 stop, $4.00 target and February 12 exit.
Crops: soybean oil improved; corn did not
March micro soybean oil was 67.64/67.68¢/lb; the more liquid same-month standard midpoint was 67.65¢ at 03:57:30 UTC. From 67.65¢, 30%/70.50, 45%/67.00 and 25%/64.56 outcomes give −$97 after modeled exit commissions across six micros, and −$310 in the 10-point sensitivity. Energy strength supports the position, but the payoff no longer supports an addition. Hold six; stop 64.56¢, target 70.50¢, exit February 12.
March micro corn was 474.00/474.50¢/bu; same-month standard ZCH27 last was 474.25¢ at 03:54:01 UTC. The stale one-lot micro last at 475.00 is not used. From 474.25¢, 20%/550, 45%/480 and 35%/450 outcomes give +$214 after modeled exit commissions across five micros; the 10-point sensitivity is −$36. Export absorption supports the position, but harvest supply and weak follow-through block an add. Hold five; stop 450¢, target 550¢, exit February 12. USDA weekly export sales, USDA October WASDE
Oil: a larger disruption premium, not a better entry
December WTI was $100.12/bbl at 04:02:20 UTC and March was $93.74. The roughly $6.38/bbl backwardation is slightly wider than at the prior cutoff. It rewards scarce prompt barrels but makes a new long expensive; a short remains vulnerable to weekend escalation. Reuters reported Brent near $114 as Houthi attacks sustained supply fears. Reuters global markets
Economic decision: Wait. Study a short only after verified supply normalization; study a long only after a price pullback with unchanged disruption evidence. Neither trigger occurred. No simulated oil position is opened.
Equity indexes: sizing works; timing does not
December Micro Nasdaq was 29,174.25/29,174.75, last 29,174.50 at 04:01:14 UTC. It rebounded 0.11% in the new session, while Russell slipped 0.10%. The prior Nasdaq/Russell sleeve is closed and is not recreated.
One MNQ short to a 30,025 stop would risk about $1,702, or 0.170% of NAV, and fits the starter budget. The shared scenario model gives +$400 after costs, but the 10-point sensitivity is −$110 and a two-risk target would require another roughly 5.8% decline. The overnight low rebounded, broad cash-session breadth was healthier than the headline indexes, and PMIs are imminent. Wait. CME Micro E-mini Nasdaq specifications
Metals and broader challenger scan
• Copper — Avoid a new long. December micro copper was $6.4280/$6.4295, last $6.4290 at 04:00:35 UTC. LME inventories remain high and regionally concentrated; prompt backwardation reflects location stress as much as global scarcity. LME inventory discussion
• Silver — Avoid. December silver near $53.90 remains vulnerable to industrial-demand weakness; March contango is not a bullish catalyst. India’s import permissions are not executed demand. Reuters India silver demand
• Platinum — Avoid. January platinum near $1,568 faces a current WPIC surplus forecast; low stock cover is the counterargument, not an entry trigger. WPIC quarterly
• 10-year Treasury — Wait. ZNZ26 was 106.312500/106.328125 at 04:05:14 UTC. A 105.000000 stop risks about $1,328 per contract, so one whole contract fits; the post-rally payoff does not.
• Yen — Wait. Inflation breadth supports BOJ normalization, but consensus CPI and a prior haven rally leave poor current compensation.
• Euro, sterling, CAD and SOFR — Avoid before the same price surveys. None offers a cleaner mechanism than yen or duration.
• Livestock and other grains — Avoid. Friday Cattle on Feed and CFTC data arrive before a stronger alternative can be underwritten.
Next 72 hours
| UTC | Catalyst | Decision test |
|---|---|---|
| Oct 23 07:15–08:30 | Europe flash PMIs | Growth and price signals for euro and sterling rates. |
| Oct 23 13:45 | U.S. flash PMI | Whether input and output prices re-accelerate after oil. |
| Oct 23 14:00 | Michigan final sentiment | One-year and five-year inflation expectations. |
| Oct 23 19:00 | USDA Cattle on Feed | Livestock supply; no current position. |
| Oct 23 19:30 | CFTC positioning | October 20 crowding after the latest selloff. |
| Weekend | Middle East and AI-financing headlines | Whether oil scarcity and technology deleveraging intensify or normalize. |
Simulated portfolio
At the recorded valuation of 2026-10-23T04:09:30.487Z, simulated NAV was $1,002,903.28, net P&L since inception was $2,903.28, and open gross P&L was -$480.70 before remaining entry commissions. The book contained 4 open positions and 15 fully closed positions. Results include modeled costs and dated marks; they are simulated, not live investment performance.
Method and limitations
Scenario endpoints and weights are analyst judgments, not measured probabilities. The calculations above use scenarioPayoff once per idea and reuse the same outputs across the board and article. Whole-contract risk includes the existing 1.25 adverse-scenario framework where stated. Exact observations retain their own clocks. Standard corn, soybean-oil and gas references are explicitly labeled same-month proxies where used. No real order was placed.
This is research and a simulated portfolio, not personalized investment advice. Futures can gap, require additional margin and lose more than a modeled stop.
Return on entry margin
Margin-weighted trade return: +1.92%, from $2,903.28 net P&L divided by $151,562.50 allocated entry margin across open and closed trades. Mixed recorded initial margin and modeled reserves. Reused margin counts for each deployment; this is not a compounded account return.
Valuation: 2026-10-23T04:09:30.487Z. Closed-trade P&L includes entry and exit commissions; open P&L includes entry commissions, with exit costs still to come. Slippage is already embedded in fills. Each trade uses its original entry margin in USD, allocated proportionally to partial exits and the remaining holding. Current margin requirements are separate. Account return on starting capital: +0.29%.
This additional presentation was introduced October 28, 2026. Historical editions use their own dated portfolio records; original research dates and accounting results are preserved.