Theme map

Where themes stand

Active 18

• Fed / Inflation / Rates

  ◆◆◆

  Tracks how inflation and labor data change the expected path of Fed policy and market interest rates.

  highactive

  October 10 update: held on the weekly reconciliation, with the week ending roughly where it started. The 30-year closed a one-year-high 5.67% on October 7 before strong 10-year and 30-year auctions pulled it back to 5.60% on October 8; effective fed funds held 3.88%, a nine-month high (209 sessions). The front end kept easing: the 2-year fell 11bp in five sessions to 4.77% and 2s10s widened to 0.51% on October 7, a second bull-steepening signal. The map's own checks split: 30-year below 5.60% on October 8 settled MISS (closed 5.60, right on the line), while 10-year below 5.35% on October 9 settled HIT (closed 5.24). Over the last 20 sessions 10-year breakevens and WTI decoupled (correlation -0.03 versus +0.62 over 60), so oil is no longer feeding the inflation-expectations leg. Next check: CPI on October 14, with the 2-year below 4.85% and 10-year below 5.35% through October 14 and the 10-year below 5.30% through October 16 all open; a 10-year close above 5.35% is the flip condition. October 7 update: held at high after a week that ended roughly where it started on the reconciliation (the 10-year up 11bp to 5.28% and the 30-year up 14bp on the week, outweighing one soft payrolls print). The long end then pushed again: on October 5 the 10-year closed 5.31% and the 30-year 5.66%, both the highest closes in a year, so the map's own 30-year-above-5.50%-on-October-3 check settled HIT (closed 5.63 on October 2). The composition changed on October 6: short-term yields fell faster than long ones, the 2-year down 10bp in five sessions and 2s10s widening 11bp to 0.48%, a bull steepening that prices some easing at the front end while term premium keeps the long end near its highs. Effective fed funds held 3.88%, a nine-month high. The October 7 $39 billion 10-year auction and the Fed minutes are the next tests. Next check: the open 30-year-below-5.60%-through-October-8 and 10-year-below-5.35%-through-October-9 checks, then the 2-year holding below 4.85% through October 14; a 30-year close back under 5.50% is the weekly flip condition that would lower conviction, while a weak 10-year auction pushing the 10-year through 5.35% restores the long-end raise. October 2 update: lowered near-term after soft payrolls interrupted a week of fresh long-end extremes. The 30-year closed a one-year-high 5.64% on September 30 and the 10-year 5.29%, also a one-year high, after a 24bp five-session rise in the 30-year, the largest in the stored 120, with 2s10s steepening to 0.41%–0.46% and effective fed funds still at a nine-month-high 3.88%. The map's own below-0.40% 2s10s-through-September-30 check settled MISS (closed 0.41). Soft September payrolls (+29,000; unemployment 4.2%) then cut the case for another near-term hike even as the long end stayed elevated near 5.61%–5.63% into October 2. Composition therefore shifted from a pure hike-path raise to term-premium-led long rates with a softer front-end policy read. Next check: 30-year holds below 5.60% through October 8; a close back through 5.60% with payroll revisions or services prices re-heating would restore the hike leg, while a 30-year under 5.45% with 2s10s compressing marks the yield spike as fading. September 30 update: raised as the long end printed fresh extremes into Core PCE day. After the September 24 one-year highs (2-year 4.87%, 10-year 5.18%, 30-year 5.47%), the curve kept bear-steepening: on September 29 the 10-year closed 5.26% and the 30-year 5.59%, both the highest closes in a year, each up 30bp in five sessions, the biggest such moves in the stored 120, while 2s10s steepened another 12bp to 0.37% and reverse-repo balances jumped 10.99 points in five sessions. Effective fed funds held 3.88%, the highest in about nine months across 202 sessions. The prior below-0.40%-through-September-30 2s10s check is therefore breached on the steepening side with the 30-year well clear of 5.45%, handing more weight to term premium and long-end supply rather than a policy-led flatten. Next check: 30-year holds above 5.40% through October 3; a close back under 5.40% with 2s10s compressing again would mark this as a one-week overshoot rather than a new regime. September 26 update: the hike arrived and the repricing moved out the curve. The Fed raised a quarter point to 3.75%-4%, its first increase in three years, and it landed as priced: stocks and bonds both rose on the decision, effective fed funds printed 3.88% on September 24, the highest in about nine months, and 2s10s closed 0.27% on September 16, the tightest gap in a year, so the market declined to price a long run of further increases. The following week reversed that composition. On September 24 the 2-year closed 4.87%, the 10-year 5.18%, and the 30-year 5.47%, all one-year highs; the 10-year's 24bp five-session rise is the largest in the stored 120, the 2-year added 20bp and the 30-year 18bp, and 2s10s steepened 11bp to 0.36%. The window therefore moved from bear flattening to bear steepening led by the 10-year rather than the 30-year, and the 2-year continuation check (above 4.50% through September 23) reads as met with room. Next check: 2s10s holding below 0.40% through September 30, the delta's own condition, now with 4bp of room; a close through 0.40% with the 30-year above 5.45% on weak long-dated demand hands the argument back to term premium, and a 2-year back under 4.65% lowers it. September 16 update: the front end repriced three more times and every one printed a one-year high. The 2-year closed 4.56%, 4.63%, and 4.65% on September 10, 11, and 14, with five-session moves of 17bp, 29bp, and 28bp, the last two the largest in the stored 120, while the 10-year closed 4.95%, 4.96%, and 4.97%, also one-year highs, and traded above 5% for the first time since 2007. 2s10s flattened 10bp, then 9bp, then 8bp, so all three windows read as bear flattening: the market is repricing the next few meetings, not the long run. Hike odds went from better than 73% to 81% by September 11 and to near-certain into the September 16 decision, with Brent at $107-109 supplying the argument. The standing check that the 2-year hold above 4.40% through September 18 is met with room. Next check: the 2-year holding above 4.50% through September 23; a 2-year back under 4.40% lowers it, and a 30-year outrunning the front end on weak long-dated demand hands the argument back to term premium. September 12 update: the market repriced the next few meetings rather than the long run. The 2-year closed 4.56% and the 10-year 4.95% on September 10, both one-year highs, with the 30-year at 5.37%; over five sessions the 2-year added 17bp and the 10-year 16bp while 2s10s flattened 10bp to 0.33%, each among the largest handful of the last 120 five-session moves, which is bear flattening rather than a term-premium leg. August CPI printed 3.71% year over year and futures went from roughly 60% odds of a September hike on September 8 to better than 73% by September 11, having spent the summer arguing over cut timing. The composition is the finding: breakevens moved about 1bp and gold fell into a confirmed oil supply shock, so a crude move through $100 was priced as a tax on growth, not as inflation. The standing labor check settled MISS, since a cooling read needed the 2-year to fall and stay lower after the employment report and it rose instead. Next check: the 2-year holding above 4.40% through September 18, with a hike delivered on September 16 while breakevens stay near 2.35% keeping this a front-end repricing; a 2-year back below 4.30% lowers it, and a 30-year above 5.45% on weak long-dated demand hands the argument back to term premium. September 5 update: the front-end leg held and the FX leg that carried last window's raise came off. The 2-year printed 4.39% on September 1 and 2, the highest closes in the stored 252 sessions and a 22bp five-session move (z +2.2), then ended the week exactly where it started at 4.34%; the 10-year added 4bp to 4.77% after its own 252-session high at 4.79%, while the 30-year added only 3bp to 5.25%, so the repricing was belly-led rather than term-premium-led and breakevens rose 4bp to 2.35% with effective fed funds unchanged at 3.63%. The data kept services carrying the argument, August payrolls +162,000, unemployment 4.1%, ISM Services prices paid 72.6, the hottest since 2022, but the dollar confirmation reversed, with USD/JPY down 2.42% to 156.22, and the 10-year note net short grew 90,452 to 897,236 contracts (report date September 1), so a supply accident and a squeeze stay hard to separate. Next check: the 2-year holding above 4.30% through September 11 CPI, with a 2-year through 4.45% and breakevens above 2.45% raising and a 2-year below 4.25% on cooling core lowering; a 30-year above 5.35% on weak long-dated auction demand hands the argument back to term premium. September 2 update: the front end took the argument back and the curve bear-flattened. The 10-year closed August 31 at 4.75%, the highest close in the stored 252 sessions, the 2-year added 17bp to 4.34%, 2s10s narrowed 7bp to 40bp, and the 30-year fell 5bp to 5.22%, the opposite composition of August's term-premium signature, with breakevens up only 3bp to 2.35%, so nominal yields did the work rather than inflation compensation. The dollar confirmed this time at +0.58% to 99.57 with USD/JPY at 160.06, so the raise condition that fired in FX alone last window now has rates alongside it. Next check: the 2-year holding above 4.30% into the September 4 employment report and above 4.25% through September 11, with ISM Services prices paid on September 3 the first read on whether services pressure survives without energy; a 30-year back above 5.30% hands the argument back to term premium, a 2-year under 4.10% lowers it. August 29 update: the chair called inflation not meaningfully slowing and recommitted to 2%, and the curve did the opposite. 2s10s flattened 11bp to 39bp through August 28 with breakevens down 3bp to 2.31%, the 30-year at 5.19% and the 2-year at 4.20% on August 27, effective fed funds unchanged at 3.63%, against July core PCE 3.34% and core CPI 2.79%. The raise condition, sticky follow-through in yields, dollar, and USD/JPY after the print, fired in FX only and failed in rates; seven high-scored wallets lifting no-cuts-in-2026 from about 79¢ to 87.6¢ on $68,915 (narrative-reactive, manipulation risk 30) agrees with the curve rather than testing it. Next check: the 2-year back above 4.20% and the 30-year above 5.25% together raises; a cooler core print with the front end under 4.10% lowers. August 28 update: the long end gave back the August build and the front end slipped under the level the last raise depended on. The 30-year closed August 25 at 5.17%, below both the 5.25% mid-month level and the month's 5.31% high, while the 2-year eased to 4.17% and 2s10s flattened 3bp to 47bp, a parallel easing rather than the front-led firming that followed core PCE at 3.34%, so neither end is currently carrying the inflation argument. Next check: whether the 30-year stays below 5.25% and the 2-year below 4.20% through the September 4 employment report rather than backing up on the next long-dated auction, with September 1 manufacturing prices paid the first read on whether services pressure survives without energy. August 26 update: core firmed just as the fuel that was supposed to be driving it fell, and the pressure moved to the front end. July core PCE rose to 3.34% year over year from 3.29% with headline at 3.70%, in the same week Brent fell 5.19% to $86.57, services prices, not energy, now carry the whole inflation argument. The curve reversed the prior two windows' signature: the 2-year added 5-7bp to 4.24% and the 10-year 6bp to 4.74% while the 30-year fell 5bp to 5.23%, flattening 2s10s to 46bp from 50bp, and speculative 10-year shorts were cut 44,596 to 875,800 (report date August 18). Next check: whether ISM Services prices paid stays firm on September 3 and the 2-year holds above 4.20% into the September 4 jobs report; a 30-year back above 5.30% hands the argument back to term premium, a 2-year under 4.10% lowers it. August 22 update: the long-end thrust stalled and the work passed to inflation compensation. The 30-year closed at 5.23% on August 20, 2bp lower on the week after touching 5.29% on August 18, while the 2-year rose 2bp to 4.19%, the 10-year 1bp to 4.69%, and 2s10s flattened 1bp to 50bp, the opposite composition of the prior two windows. What moved instead was breakevens, up 7bp to 2.34% as Brent gained 6.39%, so the long real yield fell even with nominals roughly flat. Effective fed funds held at 3.63%, the balance sheet shrank 0.21% to $6.746 trillion (August 19) with overnight reverse repo drained to $0.2 billion, and no-cuts-in-2026 traded at 86.2%, a point higher on $308,000 of weekly volume. Next check: the August 26 core PCE for whether fuel costs are reaching services prices rather than staying in the headline; the 30-year above 5.30% or breakevens above 2.45% re-raises the term-premium read, a clean core cooling lowers it. August 19 update: the long end broke to a 19-year high while everything short of it stayed still. The 30-year touched 5.29% on August 18, the highest since 2007, and closed the week 7bp higher at 5.31%, while the 2-year fell 2bp and then 3bp, effective fed funds held at 3.63%, and 10-year breakevens were unchanged at 2.27%. Bear-steepening with inflation compensation flat isolates term premium as the mover, the second consecutive window in which a quiet front end and cooling data coexisted with a rising long end. Next check: whether the 30-year holds above 5.25% through Wednesday's July FOMC minutes and the August 26 core PCE, or slips back once the oil headlines fade. August 15 update: three cool prints arrived and the long end refused them. July CPI printed 3.54% year over year with core 2.79% and headline +0.1% on the month, PPI fell 0.78%, and retail sales fell 0.58%; the 2-year fell 4bp and the 10-year 2bp while the 30-year rose 2bp to 5.21%, steepening 2s10s 5bp to 51bp. Bear-steepening into cooling data is the cleanest term-premium signature available, because the data itself removed the policy explanation for higher long yields. The $25.0 billion 30-year cleared at 5.216% with a 2.39 bid-to-cover and $31.3 billion accepted, effective fed funds held at 3.63%, breakevens rose 2bp to 2.27%, and the 11-cuts-in-2026 contract sits at 0.1% on $514K liquidity, so the argument is about the long end rather than the path. Speculative 10-year note shorts grew 100,048 to 831,204 (report date August 11), keeping a supply accident and a squeeze hard to separate. Next check: whether the 30-year drops below 5.10% on the next weak print, or back under 5.00% with the curve flattening on strong auction demand. August 12 update: the policy leg cooled and the funding leg did not. June core PCE rose 0.13% month over month, about a 1.6% annualized pace and the mildest since March 2025, with the annual rate still 3.29% and June CPI 3.73%; July payrolls fell 23,000 against forecasts for a gain, June was revised to +20,000, unemployment rose to 4.1% on a 264,000 labor-force decline, and the curve fell across the board to 2Y 4.25%, 10Y 4.69%, and 30Y 5.22% with effective fed funds unchanged at 3.63%. Q2 GDP printed 1.5% annualized versus 2.1% prior. Against that, three of twelve officials voted to hike on July 30, Hammack said a number of increases may be needed to reach 2%, breakevens rose 5bp to 2.27%, and the 30-year reached 5.25%. Two-cuts-in-2026 pricing stayed near 2.8% and no-cut-at-all odds only slipped to 85.9% from 89%, so the market still treats this as inflation-bound rather than growth-bound. Next checks: the July CPI print, and the August 12-13 10- and 30-year auctions, where a 931,252-contract non-commercial 10-year short (down 67,937, CFTC report date August 4) makes a supply accident and a squeeze hard to tell apart within a session. July 10 update: rates stayed data-gated. Warsh taskforces and Fed-governance pressure are institutional signals, not rate-cut evidence. TreasuryDirect showed the long-end reopenings cleared without disorder, while EIA products and July 14 CPI remain the inflation proof path. Keep duration relief tactical until CPI, product prices, services prices, and Fed language confirm softer inflation.

  Evidence phase: bifurcation

  • ·Next check: 2-year holds above 4.80% through October 16.partial

• Iran / Hormuz / Oil

  ◆◆◆

  Tracks whether geopolitical risk around Iran and Hormuz is creating a real oil supply shock or a temporary risk premium.

  highactive

  October 10 update: raised on the week, provisionally. Brent jumped to about $105 on October 8 from a $100.20 close as Hormuz shipping fears returned, then eased to about $104 on October 9 as event-market odds of US crude touching $100 in October fell to 34% from 39%; Brent ended the week up 1.66%. The raise rests on an intraday spike: Brent closed 104.28 on October 8, so the map's below-105 check settled HIT, and US crude barely moved. Breakevens and WTI decoupled over the last 20 sessions (correlation -0.03 versus +0.62 over 60), so the oil move is not passing into inflation pricing. Downstream legs (insurance, freight, refining margins) still have not confirmed. Next check: Brent between 100 and 105 on October 14 and WTI below $95 through October 16 (all open); a Brent close back below $100 is the flip condition that undoes the raise. October 7 update: the weekly reconciliation confirms the lower, Brent fell 3.36% on the week and speculators cut crude bets, on top of the earlier 10.01% five-session drop in natural gas. The two following sessions held rather than extended: Brent read $101.53 on October 5, down 4.04% over seven days, and $101.70 on October 7, back above $100 and firming the dollar but still down 1.81% over seven days. Downstream confirmation (war-risk insurance, freight, refining margins) is still absent in both directions, so the theme sits on price alone. Next check: Brent holds below 105 through October 8 and again through October 14; a close back above $105 is the weekly flip condition that restores the raise, while a slide toward $95 with transit counts normalising would lower it further. October 2 update: lowered as the energy complex gave back part of the supply-shock premium. Natural gas closed $2.97, down 10.01% in five sessions, bigger than all but 5 of the last 120, while Brent's weekly slide and higher Saudi pipeline flows cut the near-term oil inflation argument that had carried the prior raise when Brent settled near $107–$109. Downstream confirmation (war-risk insurance, freight, refining margins) was still missing before this window and remains missing, so the price retracement without those legs met weakens the case rather than proving route normalisation. Next check: Brent holds below 105 through October 8; a reclaim above 105 with transit counts still near the August six-a-day print would restore the raise, while three sessions of normal transits with product cracks easing would lower it further. September 16 update: crude extended hard and the demand side argued against it. WTI rose 13.73% in five sessions to $104.04 and Brent 13.34% to $109.12 on September 14, moves bigger than all but 10 and 13 of the last 120, before Brent settled at $107.19 after a 7.70% week, and the main energy agency now expects global consumption to shrink this year, so the price is being made on the supply side, consistent with the August 30 count of about six Hormuz transits against a normal 85. The standing check that Brent hold above $100 through September 16 is met, and the September 18 line sits $7 below the market. The gap is unchanged and downstream: no war-risk quote, freight rate, or refining margin has moved in either direction, so this is still transit disruption without confirmed cost repricing. Next check: Brent above $100 through September 18, with war-risk insurance and freight rising alongside crude across three sessions the remaining gate; three sessions of normal transits, or Brent back under $95, lowers it. September 12 update: the supply read is now in the composition, not just the level. Brent gained 8.78% on the week to $104.30 after touching $104.70, WTI rose 8.44% to $100.05, and against that natural gas fell 3.41% while prompt gasoline made new lows against the next month, crude accepting higher prices while fuels and gas go the other way points at lost crude supply rather than stronger energy demand, which is consistent with the August 30 count of about six Hormuz transits against a normal 85 and the halted Saudi energy sites. The standing check that Brent hold above $100 through September 16 is met so far. The gap is unchanged and downstream: no war-risk quote, freight rate, or refining margin has moved in either direction, so this remains transit disruption without confirmed cost repricing, and $787,842 across 35 wallets against a US invasion before 2027 scores 33 on observable data, which is attention rather than evidence. Next check: Brent above $100 through September 18, with war-risk insurance and freight rising alongside crude across three sessions the remaining gate; three sessions of normal transits with fuel prices easing, or Brent back under $95, lowers it. September 9 update: the physical evidence this row has demanded since mid-August arrived inside three sessions. About six vessels transited the Strait on August 30 against a normal 85, Saudi Arabia said several energy sites were halted after attacks, and Brent moved $96.28 to $98.01 and then through $100 while WTI held $91.48 after an 8.02% week, the standing check that WTI hold above $88 on September 9 settled HIT. Speculative crude longs at 197,371 net long (report date September 1) are the largest position of the leg and now sit behind a route fact rather than price alone, which is the first change in composition since the seaborne premium argument began. What is still missing is downstream: no war-risk quote, freight rate, or refining margin moved in either direction, so this is transit disruption without confirmed cost repricing, and crude stays the live inflation input against a 2.35% breakeven into September 11 CPI. Next check: Brent holding above $100 through September 16, with war-risk insurance and freight rising alongside crude across three sessions the remaining gate; transits recovering toward 50-plus a day, or Brent back under $95 without product follow-through, lowers it. September 5 update: a seventh window that moved on price and participation with the physical side still blank. WTI rose 8.02% to $91.48 and Brent 7.80% to $96.28 with natural gas +3.48%, the two benchmarks moving together and the seaborne premium at $4.80, while speculative crude longs added 12,970 to 197,371 net long (report date September 1), the largest position of this leg building into the rally. The composition still argues against a route explanation: Brent is not leading WTI, product cracks did not widen, and no transit count, war-risk quote, freight rate, or refining margin moved in either direction for a third straight window, which also leaves crude as the live inflation input against a 2.35% breakeven. Next check: WTI holding above $88 on September 9, and whether Brent starts leading WTI for a full week with cracks widening; crude retracing the 8% without product follow-through lowers this, and only verified transit, insurance, or freight repricing raises it. September 2 update: price fell and took it all back inside one window, with the physical side blank throughout. Brent fell 5.13% to $88.87 as Washington moved to take majority control of Venezuela's output, then the window closed with WTI +10.38% to $89.53 and Brent +8.36% to $94.22, and speculative crude longs sat at 184,401 (report date August 25), still accumulating. The composition points away from the Strait: WTI led Brent for a full window and narrowed the seaborne premium to $4.25, the reverse of what a route disruption produces, and six consecutive windows have moved on crude price alone with nothing touching transits, war-risk insurance, freight, or refining margins. Next check: WTI holding above $88 through September 9, and whether Brent starts leading WTI again or fuel prices and war-risk quotes move at all; only verified transit, insurance, or freight disruption raises this. August 29 update: a fifth consecutive lowering, again entirely on price. Brent fell 4.25% on the week to $89.70 and WTI 3.43% to $83.53, with nothing in the window touching transits, war-risk insurance, freight, or refining margins, and the escalation and enforcement branches producing nothing for a second straight window. The product and participation legs are now quiet rather than supportive, while speculative crude longs still added 4,717 to 184,401 net long (report date August 25), so positioning keeps building against a falling price. Next check: Brent under $85 with product cracks easing completes the retracement; only verified transit, war-risk-insurance, or freight disruption raises it back. August 28 update: a fourth consecutive lowering, entirely on price. Brent fell 6.16% over seven days while credit stayed firm, extending the retracement from $86.57 toward the $85 restore line, and nothing in the window touched transits, war-risk insurance, freight, or refining margins, the escalation and enforcement branches produced no new evidence after both failed last window. Next check: Brent under $85 with product cracks easing, or refining margins normalising while crude holds; verified transit or export disruption remains the only path back up. August 26 update: three consecutive lowerings, price gave back most of the escalation premium while diplomacy softened and sanctions enforcement produced nothing. Brent fell 3.92% to $87.74 and then 5.19% to $86.57 with WTI -3.28% to $82.19, after Iranian civilian leaders publicly pressed for peace and Washington detailed enforcement tools aimed at China's 1.5 million barrels a day of Iranian purchases; traded odds of WTI touching $90 this week collapsed to 2.8% from roughly 51%. The Brent premium over WTI narrowed to $4.92 as Asian refiners lined up to nearly double September purchases of US crude, which is substitution working rather than scarcity. Speculative crude longs still sit at 179,684, up 9,103 (report date August 18), so positioning is now the leg lagging price rather than leading it. Next check: whether the Brent-WTI gap keeps narrowing as September cargoes load; verified transit or export disruption reopens the premium, Brent under $85 with cracks easing lowers further. August 22 update: the last missing leg, positioning, finally fired, while the escalation leg softened. Brent rose 6.39% to $94.39 and WTI 5.49% with natural gas +3.05%, and speculative crude longs grew 9,103 to 179,684 (report date August 18), reversing three straight weeks in which price ran without spec accumulation. That completes the participation condition alongside last week's record refining margins, so the tightness now has price, product and positioning behind it. Against that, Tehran's own public framing walked back escalation, with $1.47 million across 41 converged wallets sitting on No to a US invasion before 2027 at 83.5¢, observable-data score 29, sentiment rather than evidence. The composition therefore changed even as the level rose: less war premium, more accepted physical tightness. Next check: verified transit or export disruption raises the read; Brent under $85 with product cracks easing lowers it. August 19 update: the product leg finally fired. Brent rose 6.51% to $89.00 into August 17 and added 3.61% with WTI 2.51% by August 18, but the decisive evidence was downstream: refining margin on the standard gasoline-and-diesel blend hit an all-time high, worth roughly $93 billion of windfall profit to the majors, which is tightness moving out of headlines and into fuel exactly as the standing check specified. Positioning still refuses to follow, with speculative crude longs down 8,585 to 170,581 (report date August 11), so price has now run without spec accumulation for a third week. Next check: whether pump and diesel prices carry into the energy detail of the next CPI rather than fading with the headlines, and whether war-risk insurance and freight rise with crude across three sessions. August 15 update: the week raised route risk on balances rather than force. Brent rose 5.94% to $88.52 and WTI 5.08% to $82.40 on threatened US economic pressure rather than a military event, and the IEA's August balance cut the projected 2026 surplus to roughly +410 kb/d from +2.46 mb/d, with Hormuz transits at 13 on August 14 against a pre-crisis baseline near 73 a day. The move ran without positioning behind it, speculative crude longs fell 8,585 to 170,581 (report date August 11), and against a heavy build: US commercial crude rose 17.4 million barrels to 424.4 million with refineries at 96.2%, a four-week build of 25.6 million. Crude accepting higher prices for a full week rather than spiking and fading meets the participation condition, so the headline-fade read is invalidated for now. Next checks: gasoline and diesel stocks drawing while crude keeps building, which moves tightness out of headlines and into fuel prices, and 50-plus Hormuz transits a day sustained across three consecutive sessions. August 12 update: the war premium drained, then rebuilt on money terms rather than force. Brent closed -11.8% at $83.54 on August 3 and -6.2% at $79.68 on August 5 as an Iran-Oman paid-passage route took shape and Saudi Arabia routed about 7 million barrels a day west by pipeline; WTI reached $78.18 by August 8. It then re-formed: +3.15% to about $86 after Iran ruled out direct US talks, then +4.0% on the week once a reparations demand was attached to any Strait deal, with traded odds of $90 WTI in August up 13 points to 39%. Physical evidence worsened while price faded: three ADNOC vessels reported attacked, Iranian drafting of vessel bans, transits still at a trickle, against a post-OPEC record 4.1 million barrels a day from the UAE in June and speculators adding 31,227 contracts into a 184,742 net long. Washington now expects roughly 600,000 barrels a day of lost supply to persist through end-2027, with official 2026 gasoline and diesel price forecasts raised 3.7% and 5.4%. Three separate escalation events produced no sustained premium, so headlines no longer count: require published Iran-Oman toll terms, or three consecutive sessions of improving Hormuz transits, war-risk insurance, freight, and delivered fuel prices moving together. July 17 update: reported Hormuz traffic stress and firmer distillate structure raised route and product risk, but oil-on-water and softer crude front structure still reject a broad shortage call. Require current traffic plus insurer, freight, LNG, ship-owner, or product-spread proof. Lower the risk only after three improving sessions.

  Evidence phase: bifurcation

  • ·Next check: Brent holds above 105 through October 16.partial

• Official Currency Defense / Yen-Funded Leverage

  ◆◆◇

  Tracks whether official yen buying and BOJ policy can hold the carry trade, and whether that defense is also a defense of the US Treasury market through Japan's holdings.

  mediumactive

  October 10 update: held on the weekly reconciliation. The dollar rose 0.33% to 158.25 yen, still below the 160 flip line, while speculators added 5,202 contracts to yen bets, a partial rebuild after recent trimming. Japan equities closed one-year highs on October 5 and 6 (to $99.32). No official intervention signal appeared. Next check: the dollar rising above 160 yen is the flip condition that re-tests official defense. October 7 update: held on the weekly reconciliation. The dollar sits near 157.83 yen, up modestly from 157.44 in the prior window, while speculators trimmed yen bets further, extending the unwind of the carry fuel this row tracks rather than rebuilding it. The exporter channel kept dominating price: the Japan equity fund closed one-year highs on October 2, 5 and 6 (to $99.32) as the broad dollar printed a one-year-high 102.17. No official intervention signal appeared. Next check: the dollar rising above 160 yen is the weekly flip condition that would re-test official defense; a renewed rebuild of speculative yen shorts above the 40,000 August level would re-arm the carry leg. September 30 update: held after the prior lower, with the dollar squeeze extending rather than reversing. Speculative yen futures net longs fell another 49,267 contracts, from 120,424 to 71,157 (official latest, CFTC report date September 22), while the US 30-year printed successive one-year highs at 5.47% then 5.59% and the rate gap kept favouring dollars. EUR/USD closed 1.13 on September 29, the lowest close in a year, the dollar index rose about 0.96% on the week, and USD/JPY sat near 157.44, still above the 157 line the deltas flag as confirming retreat from yen strength, with no credit or volatility transmission appearing in the measured block. The carry-unwind fuel this row tracks remains spent; the new evidence is dollar strength and European FX weakness rather than a renewed yen short build. Next check: whether the dollar still buys more than 157 yen after the September 30 Core PCE report; USD/JPY back below 150 with a rebuilt net short would restart the carry case, and a further EUR/USD break with no official defense language keeps the dollar-squeeze read. September 26 update: lowered, because the fuel this row tracks was spent before the event and the rate gap kept widening. The speculative crowd had already closed out the big yen short, sitting 12,575 contracts net long (report date September 8), while US short-term rates kept pulling money toward dollars: the 2-year closed a one-year-high 4.67% on September 15 and 4.87% on September 24, with effective fed funds at 3.88% after the Fed's quarter-point hike. A hawkish September 18 outcome therefore had little forced buying left to trigger, and the carry differential moved against the yen rather than toward it. This merge's evidence carries no read on the decision itself or on USD/JPY a session later, and the measured week of September 20-26 flags no USD/JPY or volatility extreme, so transmission stays absent for a fifth window. Next check: score the delta's condition from stored closes, whether the yen was still stronger a full session after the September 18 decision or handed the move straight back; USD/JPY below 150 with high yield past 290bp still confirms a genuine funding-cost change, and a move back above 158 with the net short rebuilding restarts the carry build. September 16 update: the position flipped outright two days before the decision. The speculative crowd went from 95,781 contracts net short yen to 12,575 net long (report date September 8), which changes what a hawkish September 18 Bank of Japan outcome does, it pays existing length rather than forcing a cover, so the forced-buying risk this row tracks is spent while event risk around the meeting is not. The exporter channel did not object: Japan equities closed $98.56 on September 11, the highest in a year, with USD/JPY at 153.55 after a 1.69% five-session fall, itself bigger than all but 11 of the last 120. Transmission is still absent, with no credit widening in the window. Next check: the VIX holding below 20 through September 18, which keeps this positioning rather than stress; USD/JPY below 150 on a hawkish decision with high yield past 290bp confirms a genuine funding-cost change, and a giveback above 158 with the short rebuilding restarts the carry build. September 12 update: the unwind finished inside the position data rather than in the price of risk. USD/JPY closed 153.27 on September 9, the lowest in 147 sessions, and ended the week at 153.55 after a 1.69% five-session move, while the speculative crowd went from 95,781 contracts net short to 12,575 net long, a full flip, not a trim, which means the accounts that would have been forced to buy yen have already bought. Transmission never arrived: high yield held near 270bp and financial conditions were unchanged at -0.56, and although the VIX rose 3.22 points in five sessions to 17.54, it stayed well under 20, so three consecutive windows of extremes in the largest funding currency produced no credit signature at all. The exporter channel did not object either, with Japan equities closing at a one-year-high $98.56 on September 11. Next check: the September 18 Bank of Japan decision, a hawkish outcome pushing USD/JPY below 150 with high yield past 290bp re-raises this; a giveback above 158 with the short rebuilding restarts the carry build and marks the whole leg a flush. September 9 update: the carry unwind extended for a third straight window and the transmission legs stayed silent. USD/JPY closed 154.46, 153.85, and 153.27 on September 7, 8, and 9, the lowest in 147 sessions, roughly a six-month low, with the September 7 five-session move of -3.31% larger than all but four of the previous 120, and speculative yen net shorts still 95,781 contracts after building 32,989 into the rally (report date September 1), so the position that is losing money has not been cut. Nothing confirmed it: the VIX sat near its 14.03 September 4 close, the lowest in 174 sessions, and high yield held 265bp, so three consecutive extremes in the largest funding currency have produced no credit or volatility signature at all. The 152 confirmation level is now about a point away. Next check: the Bank of Japan's September 18 decision, which either keeps yen borrowing cheap or extends the move; USD/JPY below 152 with high yield past 290bp confirms a genuine funding-cost change, the VIX holding below 16 through September 11 keeps this positioning, and a retrace above 158 with shorts rebuilding marks the whole leg a flush. September 5 update: the clearest conviction move of the week, and it came from the funding channel rather than from Tokyo. USD/JPY fell 2.42% to 156.22, touching 155.60 on September 3, the lowest close in 138 sessions and a five-session move of z -2.6, on a reported carry unwind, with no official action involved, and speculative yen net shorts built 32,989 to 95,781 contracts into the rally (report date September 1), so the position grew as it lost money rather than covering. Transmission did not follow: the VIX fell to 14.53 with a 14.03 close on September 4, the lowest in 174 sessions, and high yield sat at 265bp, so a 2.4% move in the largest funding currency produced no credit or volatility signature at all. Next check: USD/JPY below 152 with high yield past 290bp confirms a genuine funding-cost change, and the VIX holding below 16 through September 11 keeps this a positioning story; a retrace above 158 with shorts rebuilding marks the week as a flush. August 29 update: the level-defense condition fired. USD/JPY rose 0.74% to 160.04 with no official response, against a dollar index +0.85% and USD/CNH +0.16%, while speculative yen shorts were cut 10,316 to 62,792 and euro shorts built 26,729 to 53,872 (report date August 25), the yen weakened as specs covered, which points at rate differentials and real-money flow rather than fresh short-selling. Both channels stayed live in opposite directions and the earnings channel dominated: Japan equities led the majors at +1.16%, the VIX fell to 14.43, and no intervention signature appeared. Next check: USD/JPY through 162, or any official response, flips the dominant channel from exporter tailwind to unwind risk; the monthly cross-border securities data remains the direct read on repatriation. August 22 update: the currency stayed still and the equity channel did the moving. USD/JPY fell 0.18% to 158.94, still under 160 and still without an official response, while Japan equities dropped 3.02%, the worst developed block, and speculative yen shorts were cut 12,436 to 52,476 net short (report date August 18), the first meaningful reduction since the rebuild began. Both channels are live in opposite directions: a firmer yen squeezes carry funding while cutting the translated value of exporters' overseas earnings. With spot barely moving and the index down 3.02%, this window reads as short-covering plus equity de-rating rather than a macro yen move or genuine repatriation. Next check: USD/JPY under 155 with Japanese equities still falling would mark real repatriation; the monthly cross-border securities data remains the direct read. August 19 update: the pressure moved onto the Japanese leg. Japan's 2-year yield reached a 31-year high while USD/JPY was unchanged at 159.56, a rising domestic yield with no currency reaction, which turns this into a flow question rather than an FX-level one. Tokyo again let the drift toward ¥160 run with no official answer, and speculative yen shorts sit at 40,040 net short (report date August 11). The Treasury link is why it matters: Japanese institutions earning a 31-year-high yield at home have less reason to hold long US paper at a 5.31% 30-year. Next check: the next monthly Japanese cross-border securities data, for whether domestic institutions turned net sellers of foreign bonds. August 15 update: the defense lapsed and the leverage started back. USD/JPY rose 0.89% to ¥159.30, back within a point of the level that provoked joint intervention, with no official response, and speculative yen shorts rebuilt 6,911 to 40,040 net short (report date August 11) after collapsing from 171,320. The weaker yen flattered exporters, Japan equities were the week's second-best major line at +1.53%, so the earnings channel dominated price while the funding channel quietly re-accumulated. For the map the leverage channel is the one that matters, because it builds slowly, unwinds violently, and has no self-correcting mechanism. Next check: renewed official buying before ¥160 or an actual BOJ rate move rather than rhetoric; absent either, the July operation prices as a one-off. August 12 update: the first coordinated US-Japan yen buying in nearly three decades moved the level, not the trend. USD/JPY fell 3.78% to 157.40 in the week to August 1 on an estimated $87 billion of BOJ buying, alongside a hawkish 1% hold, an upgraded economic assessment, and one dissent for an immediate hike. It has round-tripped since: 155.23 on August 3, 157.93 on August 4, 157.74 on August 8, 159.24 on August 11, and 159.06 on August 12, giving back roughly half the intervention gain with no fresh official response. The crowding that made this dangerous is largely defused, with speculative yen shorts collapsing from 171,320 to 46,951 contracts (CFTC report date August 4). The theme survives because of the Treasury link: Washington bought a foreign currency partly to stop the largest overseas holder of US government debt from selling those bonds to fund its own defense, which neither the carry theme nor the bond-supply theme covers alone.

  Evidence phase: crystallization

  • ·Next check: Whether the dollar still buys more than 157 yen after Wednesday’s PCE report, which would confirm the retreat from yen strength.partial

• China / Capital Controls / Tech Security

  ◆◆◇

  Tracks how Chinese growth policy, capital controls, and technology restrictions affect global demand and market access.

  mediumactive

  October 10 update: held. The offshore yuan strengthened to 6.69 per dollar on October 9, its strongest close in the 22 sessions on record, and Chinese large caps (the China fund) rose 2.18% on the week, reversing the prior window's -1.88%. Next check: the yuan giving back that gain is the flip condition; a further yuan gain with equities firm would argue the move is no longer only coming through global prices. September 12 update: the decline arrived through global prices again rather than through policy. China equities fell 1.88% while the offshore yuan barely moved, so for a second consecutive window the repricing sits in global risk pricing rather than in licensing, export controls, or capital-flow news, and no domestic policy step landed in either direction. Wallets hold $1.18 million across 18 accounts against a Taiwan invasion this year in a market scoring 29 on observable data, which is calm sentiment rather than proof. Next check: a licensing or export-control action that shows up in disclosed shipments and procurement, or USD/CNH through 6.60, is still what moves this off hold. September 5 update: the block rose and none of it was domestic. China equities gained 0.62% and emerging markets 2.10% while USD/CNH fell only 0.14% to 6.71 into a 0.54% dollar decline, so the move is a dollar effect rather than a repricing of policy or capital-flow risk, and no licensing action, export-control step, or stimulus conversion appeared in the window. Next check: USD/CNH through 6.60, or a fresh licensing or export-control action turning into disclosed shipment and procurement disruption, is what moves this off hold. August 22 update: China outperformed while the tariff pressure landed elsewhere. China equities rose 2.02%, the strongest major block of the week against Europe +0.38% and emerging markets -0.31%, with USD/CNH flat at 6.72, a third consecutive window in which the repricing sits in equity risk premium rather than the currency, and this week it sat on the upside. The structural offset kept building: roughly $60 billion of Chinese factory construction across the global south last year progressively routes output around bilateral tariff lines, which is the same mechanism the transshipment accusations were built to catch. Next check: USD/CNH breaking meaningfully weaker, or licensing delays turning into disclosed shipment and procurement disruption. August 15 update: the repricing sat in equity risk premium, not the currency. China equities fell 3.70% against emerging markets +2.15% while USD/CNH was unchanged at 6.74 (-0.01%). The pressure surface widened on three fronts: US accusations of transshipment through 40-plus countries costing up to $26 billion a year in tariff revenue, AI-assisted customs screening in development, and Japan moving to develop domestic rare-earth deposits, which turns the critical-minerals lane from symbolic controls into capital response. Smart-money convergence of $1.07M across 20 wallets on No to a Taiwan invasion by end-2026 at 0.9615 carries an observable-data score of 29 and 20% narrative-manipulation risk, so it stays a low-resolution war-tail read and nothing more. Next check: USD/CNH breaking meaningfully weaker, or licensing delays turning into disclosed shipment and procurement disruption, either upgrades this from repricing to stress. August 12 update: the evidence rotated from tech security to domestic demand. Manufacturing PMI fell to 49.2 from 50.3, the first contraction in five months, with construction slumping and a Politburo pledge to hasten fiscal stimulus; China was the only negative equity region in the week to August 8 at -0.63%. On the strategic side the coordinated allied tariff wall was confirmed abandoned in practice, leaving US pressure unilateral while Europe hardens separately, and a US agency began reviewing how Chinese AI firms acquire and access Nvidia chips overseas as DeepSeek raised prices. Raise only if the stimulus pledge converts into announced size while PMI stays below 50, or if the overseas-access review produces license denials or shipment disclosures naming specific firms; smart-money Taiwan-invasion odds near 0.96 carry an observable-data score of 29 and remain sentiment, not evidence. Treat this as yen/JGB support potential, not a confirmed flow. Promote only if pension allocation, fund-flow, JGB auction, or FX evidence shows real domestic buying.

  Evidence phase: crystallization

  • ·Next check: Confirm the Chinese tools’ production scale and customer adoption; limited output or poor manufacturing yields would weaken the increase.partial

• Credit / Private Markets

  ◆◆◇

  Tracks stress and liquidity in private credit, leveraged finance, and other opaque funding markets.

  mediumactive

  October 10 update: held after last week's raise. The junk-bond spread rose 5bp on the week to 3.15%, a mild move after the prior week's jump, and the map's HY-above-3.00%-on-October-8 check settled HIT (closed 3.15). The investment-grade gap went the other way, narrowing 4bp in five sessions to 0.82% on October 8 (bigger than all but 6 of the last 120 moves), so the stress is concentrated in lower-quality credit rather than broad. Next check: whether the junk-bond spread keeps widening through next week's big-bank earnings; a close back below 3.00% is the flip condition. October 7 update: the weekly reconciliation confirms the raise already merged daily, the junk-bond yield gap widened 31bp to 3.24% across September 27–October 3, and the map's own HY-above-2.95%-on-October-3 check settled HIT (closed 3.10 on October 2). Into the new week spreads eased from the peak without healing: HY OAS closed 3.10% on October 5, still up 17bp in five sessions (bigger than all but 5 of the last 120), and IG OAS 0.85%, up 4bp. Financial conditions tightened modestly, with the Chicago index up 0.06 to -0.49 in five sessions, bigger than all but 5 of the last 49 such moves, even as equities printed one-year highs, public credit and public equity are still disagreeing. Next check: HY OAS holds above 3.00% through October 8 (open); a close back below 3.00% is the weekly flip condition that lowers the raise, while a renewed climb toward 3.24% with special-servicing data rising again confirms the private channel. October 2 update: raised further as the yield shock kept widening credit spreads into early October. HY OAS printed successive six-month highs at 3.08% on September 29, 3.12% on September 30, and 3.24% on October 1, with five-session rises of 40bp then 39bp then 44bp, each the largest in the stored 120, while IG OAS also marked six-month highs at 0.84%–0.86%. The map's own HY-above-2.90%-through-September-30 check settled HIT (closed 3.12). Soft payrolls cut near-term hike odds without healing the spread leg, so public credit is still pricing stress rather than a one-day rates overshoot. Next check: HY OAS holds above 3.00% through October 8; a snap back under 2.80% with equities firm would lower the raise, while another climb alongside commercial-mortgage special servicing would confirm the private-markets channel. September 30 update: raised as the yield shock moved from rates into credit spreads and commercial-mortgage distress. HY OAS jumped from the prior quieter window to 2.93% mid-week and then 3.02% on September 28, the highest close in about six months (125 sessions) and up 36bp in five sessions, the biggest five-session move in the stored 120, while IG OAS also printed a six-month-high 0.83%. Separately, commercial mortgage loans handed to special servicers rose to 11.42% in August, the highest since February 2013, with office still the leading channel and effective fed funds at a nine-month-high 3.88%. The index-easier read from earlier windows is therefore overwritten by a clear spread widening into Core PCE day rather than a heal. Next check: whether HY OAS holds above 2.90% through September 30 and into early October; a fast snap back under 2.70% with special servicing rolling over would lower the raise, while another monthly climb in special servicing with office leading confirms the private-markets channel. August 29 update: public credit financed another week without complaint and no gate, cap, or valuation reset appeared. High-yield OAS tightened 7bp to 263, investment grade 2bp to 79, and the Chicago financial conditions index was unchanged at -0.57, while the balance sheet shrank 0.22% to $6.73 trillion with the reverse repo facility drained and SOFR still printed 3.64% against effective fed funds at 3.63%, runoff now draws on bank reserves with no funding stress visible in the price of money. Private-credit exit stress stayed quiet at the index level for a third straight window with the single-name evidence unchanged. Next check: a sustained SOFR-to-fed-funds gap, or high yield widening while equities are flat, as the first genuine plumbing signal. August 12 update: public credit financed the whole window without complaint while damage kept resolving inside individual capital structures. High-yield OAS tightened 3bp to 284 and then 14bp to 271, IG reached 78, and the Chicago financial conditions index sat at -0.55 to -0.53 with no new gates, redemption caps, or valuation resets. Against that, creditors took control of a second Strategic Value Partners German holding, wood-panel maker Pfleiderer, with energy costs and interest rates named as the two shocks; direct lenders were reported writing no-growth, no-repayment user-acquisition loans to app developers; Blue Owl slowed in the second quarter. The two readings are consistent rather than contradictory, since rate and energy damage shows up in single names long before index spreads. Promotion needs index-level widening alongside a second sponsor handing over keys, and AI-linked deals breaking below issue are now the most plausible widening trigger. July 4 update: Credit stayed an active watch, not a systemic call.75 and IG OAS at 0.75 on July 2. Keep watching fund exits, collateral marks, BDCs, insurer CDS, alternative-manager equities, and public spreads.

  Evidence phase: crystallization

  • ·Next check: Whether the extra yield on junk bonds keeps widening through next week's big-bank earnings.not_met

• Narrative Rotation / Rolling Bubbles

  ◆◆◇

  Tracks how capital rotates between crowded narratives before broad macro data catches up.

  mediumactive

  October 10 update: held, with the first sign of rotation out of the most crowded trade. Chipmakers fell 4.46% on the week while the S&P 500 fund rose 1.15%, reversing the prior week's 4.18% chip gain; the S&P 500 fund had closed a one-year-high $779.09 on October 6 and the Nasdaq-100 fund one-year highs on October 5 and 6. Crowding sweeps on October 8 and 9 still named long semiconductors as the most one-sided consensus trade, with short long-dated Treasuries second. Next check: chipmakers recovering the week's loss is the flip condition that cancels the rotation read; a second week of chips lagging the broad market would confirm it. October 7 update: held, with leadership still narrow and still ignoring the rate shock. Chipmakers gained 4.18% across September 27–October 3 while small companies gained 0.42%, and the 30-year closing at a one-year high left the crowded chip leg exposed on the same facts the row has tracked since mid-September. Price then extended rather than handing off: the Nasdaq-100 fund printed one-year-high closes on October 2, 5 and 6 (to $759.66) and the S&P 500 fund a one-year-high $779.09 on October 6. The latest crowding sweep names long-only money packed into a narrow group of AI hardware and power stocks as one of the most one-sided bets, alongside leveraged funds short Treasury futures in the basis trade. Next check: small companies outperform chipmakers for a full week (the weekly flip condition); a crowded-leg drawdown that coincides with the 10-year breaking above 5.35% would be the rolling-bubble rotation this row is waiting for. September 30 update: held as the crowded chip bet kept climbing against the yield shock while the broad risk sleeve did not. Chipmakers gained 4.24% on the week while small company stocks fell 1.50%, even as the 2-year rose 20bp to a one-year-high 4.87% and, by September 29, the 10-year and 30-year printed fresh one-year highs at 5.26% and 5.59% with Russell 2000 down 2.85% in five sessions. Rising yields remain the named threat to the most crowded corner, but that corner has not yet handed off, the divergence itself is the rotation signal. Next check: whether chipmakers keep the weekly gain through the September 30 Core PCE print if short-term yields climb again; a chip fade with small caps still weak would mark the crowded corner finally repricing, while another week of chips up and everything else down keeps the held raise. September 16 update: the crowding leg fired while the handoff leg paused. Small caps fell 3.23% in five sessions to $285.14, bigger than all but 20 of the last 120 such moves, with Europe -2.74%, gold miners -5.99%, and chipmakers -5.9% in the same window, and managers named chipmakers the most crowded bet and rising bond yields the biggest threat to it, the first time in this run that the crowded trade and the thing threatening it are the same two facts. The cause is identifiable rather than flow-led: the 2-year rose 28bp to a one-year-high 4.65% into a near-certain September 16 hike, so small caps and chipmakers did travel together, but downward and on data. That is rolling-bubble risk arriving rather than a fifth handoff. Next check: the S&P 500 holding above 745 through September 23; leadership broadening back across the index once the rates repricing settles invalidates, another single-cell handoff with small caps worst keeps it flow-led. September 12 update: the cell moved again and the invalidation stayed unmet. Leadership handed back out of semiconductors after a single week in front, small companies fell hardest, and no macro release explains the switch, the fourth handoff inside the same capital neighbourhood and exactly the condition this row named as keeping it flow-led rather than data-led. Credit again declined to object, with high yield at 270bp and financial conditions unchanged at -0.56, but the price of a shock moved for the first time in this run: the VIX rose 3.22 points in five sessions to 17.54, bigger than all but 24 of the last 120 such moves, off a 14.03 one-year low. Next check: a broad, data-led move in which small caps and chipmakers travel together invalidates the rotation read; a fifth handoff, or the VIX holding above 17 while leadership keeps rotating, keeps it flow-led and starts to price it. September 5 update: breadth stayed narrow and the cell moved again. Semiconductors led at +2.28% and emerging markets at +2.10% against SPY +0.37% and IWM +0.31%, so small caps kept lagging through the September 4 employment report, the standing next check settled against broadening, and leadership handed back from software to semiconductors inside a single week, the third handoff inside the same capital neighbourhood without a macro release explaining it. Nothing objected: the VIX fell 0.71 to 14.53 and high yield sat at 265bp, both near one-year lows. Next check: an equal-weight or small-cap leg outpacing the index for two consecutive weeks invalidates the rotation read; a fourth handoff, or semis giving the 2.28% straight back, keeps it flow-led. September 2 update: the reinstated candidate got its second window and breadth narrowed again. Small caps fell 2.06% while the Nasdaq rose 0.78%, with the gain concentrated in software rather than broadening, and semiconductors fell a further 1.85%, so the semis-to-software handoff held rather than reversing. No stress channel objected, high yield tightened to 263bp and the VIX sat at 15.44, and no macro release explained the split, which keeps this flow-led rather than data-led. Next check: whether small caps stop lagging the Nasdaq in the week after the September 4 employment report, or a third cell ignites in the same capital neighbourhood; leadership broadening back across the index invalidates. August 29 update: the row was retired on August 8 with an explicit condition that it returns if breadth narrows again, and breadth narrowed. Small caps fell 1.27% against the Nasdaq +0.95%, leadership handed from semiconductors to software inside a single session with software singles up 15% and NVDA and IGV up 7%, and the handoff was flow-driven rather than data-led, fund semiconductor exposure at a one-year low, high yield tightening 7bp to 263, and the VIX down 1.47 to 14.43, so no macro data explained the move. Reinstated as a candidate only. Next check: whether the software leg holds its gain for three sessions and a third cell ignites in the same capital neighbourhood, versus leadership broadening back across the index. July 17 update: strong foundry results and positive second-quarter production failed to protect semiconductor prices, moving the rolling-bubble lens from late crowding to live de-grossing. Keep physical AI buildout separate from public-equity flows unless credit, orders, supplier guidance, and earnings conversion weaken together.

  Evidence phase: crystallization

  • ·Next check: Whether chipmakers keep falling through this week's big-bank earnings while the broad index holds.partial

• Real-Asset Split: Monetary Versus Physical Demand

  ◆◇◇

  Tracks whether the monetary real assets, gold, silver, miners, have decoupled from the physically consumed ones, energy, industrial metals, food, so that a single "real assets" basket no longer moves as one and each half answers to a different driver.

  lowactive

  September 16 update: the split held a fifth window and narrowed to a single leg. The monetary half failed once more, the gold fund down 2.38% and miners 5.99% against a 10-year up 19bp in five sessions to 4.97%, while the consumed half kept separating internally: WTI rose 13.73% in five sessions to $104.04 with Brent $109.12 before settling $107.19, and copper fell 4.53% to $6.50, bigger than all but 24 of the last 120 such moves, a week after one-year-high closes. So energy is now the only part of the basket rising, which is a narrower fact than the monetary-versus-consumed divergence this row was opened on and points at the rates repricing marking down everything that is not a supply shock. Next check: the gold fund holding above $385 through September 25, and whether copper's give-back extends while crude holds; the basket recombining on one dollar or real-yield impulse invalidates. September 12 update: the split held a fourth window and the consumed half stopped moving as one. Gold fell into a confirmed Gulf disruption while WTI rose 8.44% to $100.05 and Brent 8.78% to $104.30, but natural gas fell 3.41% and copper gave back 1.78% to $6.47, and the co-movement measures say this is structural rather than a single tape: the 20-session correlation between natural gas and WTI ran -0.38 against a 60-session +0.24 through September 10, with Brent and gas at -0.37 on the same basis. So the monetary half stays pinned away from the complex while the consumed half now separates along a supply map of its own, European gas near EUR 74 per megawatt hour against US gas down 7.30%. Next check: measured cross-correlation between precious and consumed commodities falling further, or consumed strength reaching producer prices, freight, or importer trade balances; the basket recombining on one dollar or real-yield impulse invalidates. September 9 update: the persistence condition is effectively met. Copper closed $6.85 and $6.81 on September 8 and 9, the highest in a year, with London warehouse stocks down for 42 straight days through mid-August, while Brent went through $100 and WTI held $91.48, and against that the monetary half fell, GLD -0.24% into a 0.80% dollar decline with length cut 20,024 to 222,188 (report date September 1). Three consecutive windows, spanning both signs of the dollar, is the stated three-week separation on daily evidence rather than one shock week. The demand mix inside copper points the same way: electric-car sales grew about 2% and solar installations are forecast to fall, so grid construction rather than consumer goods is pulling the metal, which is a physical-demand channel the monetary half has no access to. Next check: measured cross-correlation between precious and consumed commodities falling rather than the divergence resting on weekly percentage moves; the basket recombining on a single dollar or real-yield impulse invalidates. September 5 update: the divergence survived a reversal in the driver that could have explained it away. WTI rose 8.02% and natural gas 3.48% while GLD fell 0.24%, SLV 0.94%, and GDX 0.02%, and unlike last week the dollar fell 0.54%, so the consumed half rallied hard and the monetary half did not respond to an impulse that should have helped it, with the 10-year real yield unchanged. Two consecutive weeks is one short of the three-week persistence trigger, and the composition changed between them, which is the stronger version of the observation. Next check: a third consecutive week of separation, or measured cross-correlation between precious and consumed commodities falling rather than the divergence resting on weekly percentage moves; the basket recombining on a single dollar or real-yield impulse invalidates. August 29 update: the split showed up inside a single week. Gold fell 4.59% and silver 4.44% purely against a 0.85% dollar rally, while natural gas rose 5.71%, copper held roughly flat at -0.47%, and strength was also reported in the food complex. No standing row holds this: the metals rows score gold against the dollar and real yields, the weather row scores crops against El Niño and European drought, and the industrial-demand question was closed when copper reclaimed its January high, but none of them carries the observation that the two halves have stopped co-moving. Carry it as its own row until the basket either recombines or the divergence persists.

  Evidence phase: crack

  • ·Next check: gold ETF holds above $385 through September 25.partial

• Crypto / BTC

  ◆◆◇

  Tracks Bitcoin and crypto as liquidity, policy, and market-structure signals within the wider macro regime.

  mediumactive

  October 10 update: lowered on the week. Ether fell 6.63% and bitcoin 2.25% in the week to October 9, extending the October 7 failure to rally with equities at one-year highs. The latest daily briefs also show crypto pressure spreading into the credit, European sovereign, European industrial and housing rows rather than staying contained. Next check: bitcoin holds above $80,000 through October 14 (open); a close below $80,000 is the flip condition that confirms weakness, while a reclaim led by inflows with the dollar fading would reverse the lower. October 7 update: lowered as bitcoin failed to participate in a risk-on tape. Bitcoin fell 3.63% in the 24 hours to October 7, a day after the S&P 500 fund posted its highest close in a year at $779.09 and the Nasdaq-100 fund a third straight one-year high, failing to rally with equities at highs is weaker evidence than falling in a broad sell-off. The macro legs that explained the earlier run are against it: the dollar index closed a one-year-high 102.17 on October 5 and oil back above $100 firmed the dollar again on October 7, while bitcoin and the VIX moved in opposite directions over the last 20 sessions (correlation -0.53) after moving independently over 60 (+0.09), so the asset is trading as a risk-sensitive line rather than a hedge. Next check: bitcoin holds above $80,000 through October 14; a break below with stablecoin supply contracting would confirm weakness, while a reclaim led by inflows with the dollar fading would reverse the lower. September 12 update: the run ended and the complex split. Bitcoin fell 2.90% to $77,345, including a 2.64% session, after three weeks of record fund buying, while ether rose 3.10% and closed at $2,538.96 on September 12, the highest in about nine months, the first window in this leg where the two moved apart rather than together. The macro leg that had been doing the explaining reversed: the 2-year and 10-year closed at one-year highs of 4.56% and 4.95%, and the VIX rose 3.22 points in five sessions to 17.54, bigger than all but 24 of the last 120 such moves. Stablecoin supply at $312.3 billion keeps the supply-contraction confirmation unarmed, and strong wallets against a bitcoin record by month-end at 0.996 carry almost no information left. Next check: bitcoin holding above $75,000 through September 16; ether holding this level while bitcoin keeps falling makes the split its own theme rather than noise, and accelerating ETF outflows lower this further. September 5 update: a fifth consecutive higher week and nothing in it changed the theme. Bitcoin rose 2.30% to $80,147 and ether 0.40% to $2,473 into a 0.54% dollar decline, with no credit objection at 265bp high yield and no crypto-specific cause replacing the FX explanation that reversed and then returned. Stablecoin supply at $311.5 billion keeps the supply-contraction confirmation unarmed. The one structural note worth carrying is venue mix rather than beta: prediction-market venue volume of $447 million over 24 hours now ranks inside the top decentralised-exchange volume table. Next check: a break below $70,000 or accelerating ETF outflows lowers this; holding the pace through September 11 CPI with a firm dollar raises it. August 29 update: bitcoin rose 1.10% to $78,011 and ether 1.30% to $2,450, so the $75,000 check level held through the inflation print, but the 20%-plus weekly pace stopped the moment the dollar firmed 0.85% to 99.70. The dollar-softening explanation is now clearly reversed and the move continued anyway, which is the more informative half, with still no crypto-specific cause identified. Stablecoin supply reached $310.7 billion, leaving the supply-contraction flip unarmed, and smart wallets are paying for a range on both sides, $343,198 across 12 wallets against a $35,000 dip at 94.5¢ and $253,862 across 13 wallets against a new all-time high this year at 91.2¢, both data-anchored with zero manipulation risk. Next check: a break below $70,000 or accelerating ETF outflows lowers; reclaiming the prior pace with a rising dollar raises. August 28 update: bitcoin cleared $80,000 while the dollar index rose 0.33% to 99.17, so the move that ran two weeks on dollar weakness extended against it. That removes the borrowed FX explanation without supplying a crypto-specific one, which leaves the cause unidentified rather than external. Next check: whether bitcoin holds above $75,000 through September 1 manufacturing data with the dollar firm; below $65,000, or stablecoin supply contracting, reverses the raise. August 26 update: the repricing extended and supply confirmed rather than contradicted it. Bitcoin added 23.7% on the week to $79,432 with ether +31.5%, and total stablecoin supply reached $309.1 billion alongside the move, so growth-into-rally held for a second window and the supply-contraction flip stayed unarmed. The cause is still outside crypto, the same weeks paid gold, silver, and miners, which makes July core PCE firming to 3.34% the first genuine test of a move built on dollar weakness. Next check: whether bitcoin holds above $75,000 through the week after core PCE; below $65,000, or stablecoin supply contracting, reverses the raise. August 22 update: the non-participation read reversed outright. Bitcoin rose 22.70% on the week to about $77,200, 10.3% in the August 20 session alone to $71,655, then 7.97% more to $77,337, with ether +28.70%, against the S&P 500 down 1.35% to 1.77% depending on the window, so crypto led while the leaders lagged. The cause sits outside crypto: the dollar index fell 0.86% to 98.80 and paid bitcoin, gold and crude in the same week. Confirmation came from the dated markets rather than spot alone, the No leg on a $55,000 dip rose 77.65% over seven days to 75.5¢, traded odds of $100,000 this year moved to roughly 26% from about 14% in a day, and stablecoin supply grew to $309.5 billion with DEX volume at roughly 1.8x its seven-day average, which kills the supply-contraction confirmation the map was carrying. Converged wallets still fade a new all-time high at 92¢, so this is a range repricing higher, not a breakout call. Next check: whether bitcoin holds above $70,000 through the August 26 core PCE release; back below $65,000, or stablecoin supply contracting, reverses the raise. August 15 update: crypto failed to participate in a risk-on week, which is the weaker kind of weakness. Bitcoin fell 3.10% to $63,036 and ether 2.10% to $1,882.81 while equities rose broadly and the VIX fell 1.15 to 14.25, not selling off with risk, just not showing up. Traded odds of a $55,000 touch before year-end rose to roughly 57%, while smart money sits against both tails, 0.9555 against a new all-time high and 0.9675 against a $25,000 dip, describing a range rather than a breakdown. Stablecoin supply kept climbing above $307.6 billion, which cuts directly against the supply-contraction confirmation the map was carrying. Next check: a second consecutive risk-on week with crypto down confirms a structural de-rating; one full week of re-coupling to equity beta restores the theme. August 12 update: crypto participated without leading. BTC fell 2.0% to $62,766 into August 1, then rose 3.40% to $65,079 into August 8, with ETH at $1,922.97 and total stablecoin supply essentially flat at $306.8B then $307.2B (USDT $183.1B, USDC $72.3B). The tension worth naming is that spot moved about 2% while the medium-downside distribution moved almost ten points: the No side of a dip to $55,000 fell 9.9 points to 0.455 before No on $45,000 settled at 0.805 on only $54,283 across nine wallets, which is too small to lean on. Positioning is repricing faster than price. Confirm weakness on stablecoin supply contracting alongside spot; invalidate if supply keeps growing and downside odds mean-revert. July 6 update: BTC/ETH bounced with risk after payrolls, but the cleaner signal stayed in rails and market plumbing. Wintermute framed the move as relief rather than independent crypto strength, while Delphi and Castle kept value-accrual and options-market structure in Context. Keep BTC as liquidity-sensitive and focus proof on ETF flows, MSTR funding quality, and whether rails publish real usage.

  Evidence phase: crack

  • ·Next check: bitcoin holds above $80,000 through October 14.not_met

• Housing / Residential Construction

  ◆◆◇

  Tracks whether higher financing costs and a changed ownership rule set are pushing residential activity from existing-home turnover into new construction, and where the cost pressure shows up when they do.

  mediumactive

  October 10 update: raised on the weekly reconciliation. Long-term yields stayed near one-year highs, with the 30-year at a one-year-high 5.67% on October 7 before easing to 5.60%, and consumer sentiment showed record-low current conditions, so the financing-cost and demand legs both pressed on residential activity. Next check: the 30-year closing below 5.50% is the flip condition that would mark this as a rates spike rather than a sustained squeeze. October 7 update: raised on the weekly reconciliation, with the strain now visible in the housing data itself rather than only in Treasury yields. Mortgage rates sit near 7.03%, past the toward-7% path flagged in late September and up from 6.76% on September 10, and builder confidence is at 32. The financing channel kept pressing: the map's own 30-year-above-5.40%-on-October-3 check settled HIT (closed 5.63 on October 2), and the 30-year closed 5.66% on October 5, a fresh one-year high. The October 6 bull steepening lowered short-term yields but not the long rates that set mortgage costs. Next check: the 30-year stays above 5.50% through October 14; a close back below 5.50% is the weekly flip condition, while starts and permits data showing new construction holding up as existing-home turnover falls would confirm the redirection this row is built on. October 2 update: raised as the long-rate channel that sets mortgage costs held at fresh extremes even after soft payrolls. The 30-year closed 5.64% on September 30, the highest close in a year and up 24bp in five sessions, the biggest such move in the stored 120, then 5.61% and 5.63% into October 2, so financing-cost pressure on residential activity did not ease with the front-end hike repricing. July starts already down 13.5% year over year leave little cushion if mortgage rates stay near the prior 6.76% toward-7% path. Next check: 30-year holds above 5.50% through October 3 and stays clear of a sustained break under 5.45% into October 8; a mortgage-rate rollover with starts stabilizing would mark this as a rates spike rather than a sustained housing squeeze. September 30 update: raised as the long-rate channel that sets mortgage costs printed fresh extremes. The 30-year Treasury closed 5.49% on September 25 and 5.59% on September 29, both the highest closes in a year, after a 30bp five-session rise, the largest in the stored 120, the same move the deltas flag as pushing US mortgage rates toward 7% against the prior 6.76% thirty-year mortgage versus 6.35% a year earlier, with July starts already down 13.5% year over year. Financing-cost pressure on residential activity is therefore intensifying rather than stabilizing into October. Next check: 30-year holds above 5.40% through October 3; a break back under 5.40% with mortgage rates rolling over would mark this as a rates spike rather than a sustained housing squeeze, while another leg up in mortgage rates with starts still contracting confirms the raise. September 12 update: the row revives after 53 quiet days because three legs arrived together. Borrowing costs are higher than a year ago, with the thirty-year mortgage at 6.76% on September 10 against 6.35% a year earlier, builders answered by cutting output with July starts down 13.5% year over year, and a new law bars large investors from buying additional existing houses while leaving new construction untouched, which redirects that capital toward building rather than out of housing. The three fit together rather than conflicting: a rate-constrained market with restricted existing-home demand from the largest buyers points investment at the supply side, and it does so with the 10-year at a one-year-high 4.95%, so the financing cost is still rising underneath the redirection. What is unproven is whether the redirected money shows up as starts rather than as land banking. Next check: whether pricing power appears in inputs such as cement and lumber rather than in finished homes, which is what would confirm the construction channel.

  Evidence phase: crystallization

  • ·Next check: 30-year holds above 5.50% through October 3.partial

• Industrial Metal Demand Gap

  ◆◆◆

  Tracks whether copper is signalling industrial demand weakness that the AI-buildout and precious-metal narratives are currently masking.

  highactive

  September 2 update: the settled check went against the retirement premise. Copper closed at $6.5395 on September 1, a miss on the $6.76 level whose reclaim was the stated reason for retiring this row on August 26, so the January high turned back a fourth approach after all. The reopen condition is still not met, because copper sits above $6.37 and precious metals fell hard rather than rising, with gold -5.41%, silver -4.24%, and miners -5.55% in the same week, so the industrial-versus-monetary divergence is not what is happening. Next check: copper breaking $6.37 while precious metals rise, or ISM Manufacturing new orders weakening alongside the price. August 19 update: the two metals split hard in the same week. Copper fell 2.16% to $6.46 a pound and made a new monthly low after three failed runs at January's $6.765 high and two failed London auctions this month, while gold miners rose 3.20%. Copper is the metal that should be tightest if data-center electrification, grid capex, and turbine buildout were converting into physical demand at the pace the AI rows assume, so a monthly low is a direct contradiction to the buildout narrative rather than a currency or real-rate signal. Nothing is settled yet: failed auctions can be a liquidity or settlement artifact rather than demand. Next check: whether copper defends $6.37 and then $6.00, the most heavily traded price of the past year.

  Evidence phase: bifurcation

  • ·Next check: Whether shrinking warehouse stocks start showing up in manufacturers’ input costs rather than only in exchange prices.partial

• AI Component-Cost Pass-Through

  ◆◇◇

  Tracks whether the AI buildout imposes measurable input costs on firms that are not participating in it, through memory, packaging, and power.

  lowactive

  September 2 update: the cost source finally has a size on it. Prepaid supply and capacity commitments at the largest AI chip buyer rose from $119 billion to $279 billion as of July 26, primarily for memory and manufacturing facilities, the buildout contracting years ahead for the same memory and packaging capacity that non-AI hardware buyers price off, which makes the Apple guidance cut look less like inventory timing. The birth condition is still unmet, because no second or third buyer has attributed margin or guidance pressure to the same inputs in this window. Next check: whether memory and packaging costs show up in hardware makers' gross margins next quarter rather than only in prepayments; memory contract prices rolling over invalidates. August 12 update: the first clean instance arrived. Apple cut forward guidance solely because of an AI-driven memory and chip shortage, from a $109 billion quarter growing 16%, with management framing it as a "100-year flood" rather than a normal cycle. No standing theme holds this: the AI/semis row is about leadership and flows, and the China/trade row is about export controls. This is someone else's capex cycle setting the cost base for firms outside it. The birth condition is deliberately strict, because one guidance cut can be inventory timing.

  Evidence phase: crystallization

  • ·Next check: Whether a filing, contract disclosure or customer statement independently confirms the 20% renewal premium within the next two weeks.partial

• Cross-Asset Crowded Calm

  ◆◆◇

  Tracks whether separate-looking positions across volatility, rates, metals, and equity futures are in fact one stacked bet on continued quiet, so that a single shock unwinds several markets on the same day.

  mediumactive

  September 16 update: the shared-shock condition fired. Long borrowing costs cleared 5% intraday and the crowded equity leg began unwinding in the same sessions, with the gold fund down 2.38%, miners 5.99%, small caps 3.23% over five sessions and chipmakers 5.9% in a session, bonds, gold, and equities marked down together, which is the correlation this row was opened to catch rather than another leg moving alone. Two of the four positions have now unwound: yen shorts flipped to a 12,575-contract net long and gold length was cut to 222,188. The rates leg went the other way and got more extreme, the 2-year rising 29bp and then 28bp in five sessions, the two largest such moves in the stored 120, with the 2s10s-to-semiconductor correlation at +0.51 against a 60-session -0.15. Next check: whether the VIX, last at 17.54, breaks above 20 while the 10-year net short is still near 897,236 contracts, which converts a correlated markdown into a funding event. September 12 update: one leg unwound completely and the price of a shock finally moved. The speculative yen position went from 95,781 contracts net short to 12,575 net long, a full flip rather than a trim, following gold length being cut 20,024 to 222,188 the week before, two of the four legs have now reduced in consecutive windows, which meets the stated trigger in substance even though they reduced separately rather than in a shared shock. The VIX rose 3.22 points in five sessions to 17.54, bigger than all but 24 of the last 120 such moves, off a 14.03 one-year low, while high yield stayed near 270bp and financial conditions were unchanged at -0.56, so the calm is fraying in volatility before credit. The rates leg went the other way and got more extreme, with the 2-year up 17bp in five sessions to a one-year-high 4.56%, 2s10s 10bp flatter, and the 10-year net short at 897,236 contracts, while the equity-rates correlation held near +0.51. Next check: one session in which bonds, gold, and equities fall together, or high yield through 290bp with the VIX above 18. September 9 update: one leg is now moving hard and none of the others has answered. The 20-session correlation between 2s10s and semiconductors rose to +0.52 through September 8 against a 60-session -0.15, up from +0.47, extending the co-movement this row was opened to catch, while USD/JPY fell to 153.27, a 147-session low, with yen shorts still 95,781 contracts (report date September 1). The price of a shock did not move with it: the VIX stayed near its 14.03 September 4 close and high yield held 265bp, so a 3.3% five-session move in the largest funding currency raised nothing anywhere else. Gold length was cut 20,024 to 222,188, making it the first of the four legs to reduce, and it reduced on its own rather than in a shared unwind. Next check: a second leg unwinding in the same window, one session in which bonds, gold, and equities fall together, or high yield through 290bp with the VIX above 18. September 5 update: raised, because the calm is now measured rather than inferred. The crowded-calm check fired on September 3, 4, and 5 with the VIX in the 1st-to-10th percentile of the past year, high-yield OAS in the 2nd, and SPY in the 93rd-to-96th, the VIX closed 14.03 on September 4, the lowest in 174 sessions, and high yield sat at 265bp. The legs kept moving separately and against their own prices: the 10-year net short grew 90,452 to a fresh extreme of 897,236, yen shorts built 32,989 to 95,781 into a 2.42% yen rally, and gold length was cut 20,024 to 222,188 (all report date September 1). The correlation trigger stayed part-way there, with the 20-session 2s10s-to-semiconductors correlation at +0.47 against a 60-session -0.17. Next check: one session in which bonds, gold, and equities fall together, or high yield through 290bp with the VIX above 18, which converts crowded calm into an actual stress read. September 2 update: the calm held a third week and the first correlation actually moved. The 20-session correlation between 2s10s and semiconductors ran +0.54 through September 1 against -0.14 over 60 sessions, rates and equity leadership beginning to move together, which is the co-movement this row was opened to catch, while the VIX stayed under 16 at 15.44 and high yield tightened to 263bp, so nothing in the price of risk has repriced. The metals leg broke without the position responding: GLD fell 7.32% over five sessions, a 1.8 standard-deviation move, with gold length still 242,212 (report date August 25). Next check: one session in which bonds, gold, and equities fall together, or the VIX above 20 with the rates short still near record size. August 29 update: the invalidation path advanced a second week and the positions still did not unwind. The VIX fell 1.47 to 14.43, staying under 16 through the core PCE window, while gold length grew 22,717 to 242,212, yen shorts were cut 10,316 to 62,792, and euro shorts built 26,729 to 53,872 (all report date August 25), legs moving separately and in opposite directions, with no shared shock and no rise in cross-asset correlation. Gold length building while gold fell 4.59% is the sharpest version of the observation: the position is not responding to its own price. Next check: one session in which bonds, gold, and equities fall together, or the VIX above 20 with the rates short still near record size. August 22 update: partially confirmed on the calm, not yet on the correlation. The VIX sat at 15.13 through a 5.61% semiconductor drawdown, which is the cleanest evidence yet that volatility is not pricing single-market stress; the 10-year net short stayed extreme at 875,800 contracts although 44,596 smaller, and gold length grew to 219,495 (both report date August 18). So one leg trimmed, one leg grew, and neither moved because of a shared shock. The birth condition is unchanged and remains correlation rather than level: one session in which bonds, gold and equities fall together. Next check: the August 26 core PCE release and Nvidia's results, either of which can move more than one of these markets at once. August 19 update: four apparently independent positions are leaning on the same condition. The VIX closed at 15.02, the lowest of the year, while the CFTC report dated August 11 showed a record 831,204 net short in 10-year Treasury futures, up 100,048 on the week, and gold length up 24,208 to 214,856; leveraged funds sat net short Nasdaq-100 futures. No standing row holds this, because the rates row reads the short as a supply question, the gold row reads length as a debasement question, and the AI row reads equity positioning as crowding. The observation here is that all of them are short volatility in different clothes, and nothing in the price of risk is paying for a shock, with the 30-year at 5.31% and high-yield OAS at 271bp. The birth condition is correlation, not level.

  Evidence phase: compression

• Insurance Balance-Sheet Divergence

  ◆◇◇

  Tracks whether softer property pricing is coinciding with rising casualty-reserve and private-credit risk inside insurer balance sheets, while public bond spreads stay calm and price none of it.

  lowactive

  September 4 update: the pricing half of the divergence now has numbers. Property catastrophe reinsurance rates are down 16% for 2026 after the June and July renewals, against 12% at January 1, so the softening accelerated through the year rather than stabilising at renewal, and a below-normal hurricane forecast removes the event that normally turns the cycle. That leaves underwriting margin compressing into a book still carrying casualty-reserve and private-credit risk, with public spreads pricing none of it at 265bp high yield. The birth condition is unchanged and company-level: a reserve charge or a redemption-term change, not a soft-pricing headline. Next check: insurer reserve top-ups, and January renewal quotes coming in below current guidance rather than in line with it. August 28 update: the candidate opens on a divergence rather than a loss. Property pricing is softening at the same time casualty reserves and private-credit holdings are getting riskier, and public bond spreads register nothing, high yield tightened to 267 in the same window. No standing row holds this: [[credit-private-markets]] treats insurers as an exposure channel into private credit, not as underwriters whose own pricing cycle is turning, and the calm in public spreads is exactly what would delay the signal. Nothing here is confirmed; the birth condition is a reserve charge or a redemption-term change, not a soft-pricing headline.

  Evidence phase: crystallization

  • ·Next check: Watch for insurer reserve top-ups and January renewal quotes coming in below current guidance rather than in line with it.partial

• Weather-Driven Food And Crop Supply

  ◆◇◇

  Tracks whether El Niño intensity and European heat and drought turn into a 2027 food-price and emerging-market fiscal shock, with agricultural output as the transmission channel rather than chokepoints, tariffs, or energy.

  lowactive

  August 15 update: two independent signals landed in the same week that no standing row can hold. African nations are preparing for what is being described as a one-of-a-kind El Niño shock, and Europe's heat and drought are threatening next year's crops. The channel is agricultural output and EM fiscal stress, not the energy, trade, or supply-chain rows, and the price effect is a 2027 problem rather than a 2026 one, which is exactly why it is cheap to watch now, before grains and softs reprice. Nothing here is tradable yet: the birth condition is a futures move that lines up with confirmed planting or harvest shortfalls, not another forecast headline.

  Evidence phase: crystallization

  • ·Next check: Whether the dollar keeps sliding after the August 26 inflation release, or a hot print pulls it back and strands the miners' three-to-one run.fired

• Allied Tariff Generalization

  ◆◆◇

  Tracks whether tariff escalation is generalizing from strategic rivals to treaty allies, which transmits through freight, autos, and cross-border capex rather than through the China-exposure and production-relocation channels the standing trade row covers.

  mediumactive

  August 26 update: the strictest birth condition is met. Canada answered the 50% US levy on roughly $20 billion of goods with $20 billion of counter-tariffs mirroring it, so retaliation is implemented policy rather than a threat and the negotiating-posture reading no longer explains the sequence. What remains open is transmission: freight rerouting, auto content rules, and deferred cross-border capex are the channels that would make this more than a headline exchange, and none has produced evidence yet. Next check: whether the counter-tariff list takes effect on schedule rather than being suspended for talks, and whether a second treaty ally receives comparable treatment. August 22 update: the instrument was turned on the closest US partner. A 50% tariff on roughly $20 billion of Canadian goods took effect after negotiations collapsed at the eleventh hour, with Ottawa threatening dollar-for-dollar retaliation. No standing row holds this: the supply-chain row is built around China exposure and production relocation, and neither mechanism applies to an integrated North American manufacturing base where tariffs hit the same vehicle several times across the border. This is a fragmentation story with different second-order effects, freight rerouting, auto content rules, and deferred cross-border capex, and it lands beside the Canada Section 338 duties already scheduled for August 19. The birth condition is deliberately strict, because one bilateral breakdown after failed talks can be a negotiating posture rather than a policy turn.

  Evidence phase: crystallization

  • ·Next check: Whether the counter-tariff list takes effect on schedule rather than being suspended for talks.partial

• Labor Supply Versus Labor Demand

  ◆◇◇

  Tracks whether weakening payrolls reflect workers leaving the labor force or employers cutting demand, since the two imply opposite inflation and policy paths.

  lowactive

  September 5 update: the month that was supposed to identify this produced neither signal. August payrolls rose 162,000 with unemployment unchanged at 4.1% and July job openings at 7.3 million, so the July pairing of a 264,000 labor-force decline against a negative payroll print did not repeat, and JOLTS stayed inside the little-changed range that has held all summer. The market read it the same way: the 2-year ended the week exactly where it began at 4.34% after touching a 252-session-high 4.39%, so nothing in the front end treated the release as a demand break. Next check: a payroll print under 75,000 with the 2-year below 4.20% raises this; two further months where labor-force change and payroll change point the same way, with prime-age participation confirming, is still the only clean identification. August 12 update: July produced both signals in the same month, a 264,000 labor-force decline beside a -23,000 payroll print, unemployment at 4.1%, and June JOLTS openings, hires, quits, and layoffs all little changed. If the marginal worker is leaving, payroll weakness is not disinflationary and wages stay firm; if demand is breaking, the front-end rally to 2Y 4.25% is justified. No standing theme resolves this cleanly: the rates row treats labor as an inflation input and the demographic material treats participation as a 2030s fiscal problem, while the live question is identification.

  Evidence phase: compression

• Stablecoins / Tokenized Cash / Collateral

  ◆◆◇

  Tracks whether stablecoins and tokenized cash are becoming durable financial infrastructure, with focus on reserves, redemption, custody, and collateral.

  mediumactive

  October 10 update: held and reconciled. After 93 days with no new evidence on settlement adoption, the weekly review checked the size legs: stablecoin supply stands at $312.0 billion and tokenized funds at $35.1 billion, so the row stays on the map rather than being retired. Next check: fresh settlement or collateral adoption evidence; continued silence alongside flat supply would put the row back up for retirement. September 12 update: the rails grew again and the new evidence was mechanical rather than commercial. Total stablecoin supply reached $312.3 billion with USDT at $183.5 billion and issuer fee income still climbing, while the concrete change in the window was governance work on collateral vaults, including a penalty adjustment on an RLUSD vault adapter with execution planned for September 14, parameters being tuned on existing collateral, not settlement volume being won, and no issuer terms, custody, redemption, or named-counterparty evidence moved. Next check: growth sourced from real settlement volume rather than from crypto funding rates, with exchange or collateral acceptance broadening and tokenized money-market use in visible size the promotion path; supply contraction remains the flip. September 5 update: the rails grew again and the quality of the growth got easier to see. Total stablecoin supply reached $311.5 billion with USDT at $183.4 billion and USDC at $74.7 billion, Tether generated $16.10 million and Circle $6.58 million of 24-hour fees, and governance buyback votes spread, Ethena's fee switch passed on 100% of 17.8 million votes to route protocol revenue into programmatic buybacks that scale with USDe supply. The caveat is the source: that revenue is funding-rate dependent, so the buyback stream is a claim on crypto carry, not evidence of settlement usage, and no issuer terms, custody, redemption, or named-counterparty evidence moved. Next check: exchange or collateral acceptance broadening, or tokenized money-market use in visible size; supply contraction remains the flip. August 22 update: the rails grew with the crypto move rather than independently of it. Total stablecoin supply reached $309.5 billion with USDT at $183.2 billion and USDC at $73.7 billion, Tether generated $15.90 million and Circle $6.37 million of 24-hour fees, and DEX volume ran at roughly 1.8x its seven-day average alongside bitcoin's 22.70% week. Growing supply into a rally is the ordinary direction, so this is a modest raise on issuer economics being visible and measurable rather than on new settlement proof, no issuer terms, custody, redemption, or named-counterparty evidence moved. Next check: whether supply holds if spot gives the week back, and whether activity survives netting out incentives; supply contraction is the flip. July 10 update: tokenized rails gained a bank-connector proof point but stayed below macro-driver status. Crypto is Macro Now says SWIFT's 17-bank deposit-token platform runs on a permissioned EVM chain, while Coinbase and Sky added stablecoin usage and issuer-side data. The proof path remains real settlement volume, named counterparties, issuer terms, custody, redemption, reserves, and activity net of incentives.

  Evidence phase: crystallization

  • ·The proof path is reserves, redemption, custody, and real settlement volume.not_met

• Red Sea / Bab al-Mandab Export Chokepoint

  ◇◇◇

  Tracks whether Bab al-Mandab and Saudi Red Sea loading infrastructure become a chokepoint separate from the Strait of Hormuz.

  unratedactive

  August 12 update: live shipping risk started migrating off Hormuz. Houthi forces launched a large-scale attack on Yemen's Saudi-backed government, reportedly killing at least 30 government troops and injuring 17 civilians in shelling, in the same week Brent fell 7.98% on Hormuz reopening talk. Saudi Arabia is reported to be preparing an offensive to break the chokehold on its exports, and a Turkey-Saudi-Pakistan defense pact was signed in Mecca. The mechanical reason this matters now is that Saudi Arabia is routing roughly 7 million barrels a day west by pipeline to avoid the Strait, concentrating volume on the same water. Promotion requires a loadings or insurance number, not another attack headline.

  Evidence phase: compression

Candidates 6

• European Sovereign Debt Stress

  ◇◇◇

  Tracks whether government-debt strain inside the euro area, starting with France's fiscal and political position, becomes a standing driver of the euro, European rates and the dollar rather than a one-week headline.

  lowcandidate

  October 10 update: held as a candidate, with the spread confirmation it lacked now arriving. The gap between French and German 10-year borrowing costs reached its widest since 2011, with Italy, Greece and Belgium also under pressure, so the strain is no longer confined to one country. EUR/USD closed one-year lows at 1.12 on October 5 and 7 as French debt worries kept weighing on the euro, and the map's own dollar-index-above-101.50-on-October-9 check settled HIT (closed 102.21). European Central Bank officials say no intervention is needed yet, and the bank's bond-buying backstop is the tool markets are watching. Next check: the euro closing back above $1.14 is the flip condition that would mark the move as dollar strength; further widening in Italian or other high-debt spreads, or a rating action or failed budget vote, would raise conviction. October 7 update: the candidate opens because a euro-area sovereign-debt strain is moving a major currency and no standing row holds it. The euro fell to $1.116, a 17-month low, on French debt and election worries; EUR/USD closed a one-year-low 1.12 and the dollar index a one-year-high 102.17 on October 5. The European industrial cost row had been claiming the euro leg, but its own lower-conviction list names currency channels as the alternative. Unproven: breadth beyond France and spread confirmation against German debt. Next check: the dollar index holds above 101.50 through October 9.

  Evidence phase: ignition

  • ·Next check: dollar index holds above 101.50 through October 9.partial

• European Industrial Cost Compression

  ◇◇◇

  Tracks whether European manufacturers are being squeezed from both sides at once, structural labour shedding against a rising energy input bill, so that European industrial equity underperformance becomes a cost story rather than a demand or tariff story.

  lowcandidate

  October 10 update: held as a candidate. European stocks (the Europe fund) were flat on the week, up 0.07%, while natural gas rose 6.66%, reopening the energy-bill gap that last week's 10% gas drop had closed. The dollar-index-above-101.50 check settled HIT (102.21 on October 9), but the euro leg continues to trace to French sovereign-debt worries, now carried by its own row. Next check: a second structural European industrial job cut, and whether higher gas prices show up in European equities underperforming. October 7 update: the weekly reconciliation raised the row, European stocks fell 1.96% on the week, the Europe fund dropped 2.93% in five sessions to $85.49 by October 1 (bigger than all but 16 of the last 120 such moves), and the euro hit a one-year low. The attribution weakened in the following sessions, though: the euro's slide to $1.116, a 17-month low, with EUR/USD closing a one-year-low 1.12 and the dollar index a one-year-high 102.17 on October 5, was tied to French debt and election worries, a currency and sovereign channel that this row's own lower-conviction list names as the alternative explanation. Natural gas falling 10.01% in five sessions also eased the US side of the energy-bill gap rather than widening the European one. The sovereign leg is now tracked separately as a candidate row. Next check: dollar index holds above 101.50 through October 9 (open); a second European industrial employer announcing structural cuts, or producer prices following the energy bill, is what this row still needs to separate itself from the currency story. September 30 update: raised on plant-level cost evidence rather than another broad Europe-versus-world tape. An 84-year nylon-precursor plant in Saxony-Anhalt is under a state-ordered wind-down after insolvency, in a chemical-heavy state where the AfD just took 43.8% of the vote, a concrete feedstock and energy-cost casualty rather than only weak final demand. The prior window's Europe -2.74% five-session decline had looked broad because small caps fell alongside; this plant closure separates the industrial-cost channel. Dollar strength and EUR/USD at a one-year-low 1.13 tighten the imported-energy squeeze into Core PCE week. Next check: whether the next round of European factory guidance names energy or feedstock costs, not only weak demand, as the reason margins are shrinking; a guidance wave that cites only demand with energy bills easing would lower the raise back to watch. September 16 update: the equity leg moved and this row cannot claim it. Europe fell 2.74% in five sessions to $89.23, bigger than all but 22 of the last 120 such moves, while crude rose above $104 and Brent reached $109.12, but small caps fell 3.23% in the same window with miners and chipmakers alongside, so the decline was a broad rates-led markdown rather than European underperformance on input costs. No second European industrial employer announced structural cuts after Volkswagen's 50,000, so the labour leg stands at one instance and the birth condition stays half met. Next check: European factory output or company guidance naming energy cost rather than weak demand as the margin driver, and Europe underperforming a rising index rather than falling with a falling one. September 12 update: the input-cost leg now has a geography rather than only a barrel price. European gas sits near EUR 74 per megawatt hour after doubling this year while US natural gas fell 7.30%, and the divergence is measurable in co-movement: Brent and natural gas ran a -0.37 20-session correlation against +0.24 over 60 sessions through September 10, with WTI and gas at -0.38 on the same basis, so the same crude shock raises the European bill while the American one falls. The labour leg did not extend, no second structural announcement followed Volkswagen's 50,000 cuts, so the birth condition stays half met on evidence that is now firmer on cost than on employment. Next check: whether European factory output and company guidance name energy cost rather than weak demand as the reason margins are shrinking, and whether European equities underperform for a third consecutive week while gas holds these levels. September 5 update: three pieces arrived in the same week and no standing row holds the combination. Volkswagen approved a further 50,000 job cuts, Europe was the only equity block lower on the week at -0.39%, and the input bill rose underneath both, with crude up 8.02% and natural gas 3.48%. The trade and tariff rows are about goods exposure and allied escalation, the energy rows are about route risk and crude price, and neither carries the case where a labour-shedding manufacturing base meets a rising energy cost while its equities lag a rising world. Nothing is confirmed: one employer restructuring inside one bad week is the normal state of European industry, which is why the birth condition is a second structural announcement or a third week of underperformance. Next check: a second European industrial employer announcing structural cuts, or European equities underperforming for a third consecutive week while energy holds these levels.

  Evidence phase: ignition

  • ·Next check: dollar index holds above 101.50 through October 9.not_met

• Defense Industrial Cycle

  ◇◇◇

  Tracks whether defense budget growth is converting into appropriated dollars and supplier order books, making the sector a durable industrial demand channel rather than a headline budget request.

  lowcandidate

  October 7 update: held as a candidate on the weekly reconciliation. Contract awards are large, but government funding is running on a stopgap budget, which delays new-start work and so blocks the budget-to-order-book conversion this row needs, the same appropriation gap its lower-conviction list names. Nothing in the window moved the fiscal 2027 request itself. Next check: a full-year appropriation near the request, or prime and subsystem suppliers reporting backlog growth tied to missile defense; continued stopgap funding through the next reporting cycle with flat backlogs would argue for retiring the row. September 12 update: the row revives on a budget fact after lying dormant since July. The Space Force is asking for roughly 80% more money for fiscal 2027, with $17.5 billion directed at missile defense, a scale of increase large enough to matter to suppliers if it survives appropriation, and one that lands beside the same critical-minerals and grid-hardware constraints the buildout rows already track. A request is not a contract, which is the whole limitation here: nothing in this window shows appropriated dollars, awarded programmes, or supplier backlog moving. Next check: appropriated figures rather than requested ones, and supplier order books or backlog disclosures naming the missile-defense line.

  Evidence phase: compression

• Event-Market Open Interest As Macro Gauge

  ◇◇◇

  Tracks whether open interest on regulated event exchanges, rather than volume, is a usable read on how concentrated the macro bet has become, since holding into a dated event and churning a market produce opposite signatures.

  lowcandidate

  September 12 update: the candidate opens on a divergence inside one exchange's own monthly data rather than on an odds level. August cash trading on the largest regulated event exchange fell 7.1%, ending twelve consecutive monthly increases, while open positions grew 18.8% over the same month, falling turnover with rising open interest is the signature of participants holding a view into a dated event rather than traders recycling risk. The concentration is what makes it macro rather than venue trivia: open interest in economics contracts rose 129.2% and Fed-decision contracts supplied 88% of that increase, landing directly on the September 16 meeting the rates row is built around. No standing row holds this, because the prediction-market board scores odds moves against news and traded prices and says nothing about position accumulation. What is unproven is durability: this may simply be event plumbing around one unusually contested meeting. Next check: whether open interest stays elevated after the September meeting rather than collapsing back, which is what separates a standing gauge from a one-event artifact.

  Evidence phase: compression

• Copper Tightness / Grid Demand Pull

  ◇◇◇

  Tracks whether copper's inventory drawdown and price extreme are the physical confirmation that electrification and data-center grid buildout are converting into metal demand, separate from the monetary metals and from the equipment order books.

  lowcandidate

  September 16 update: the price leg gave back the extreme. Copper fell 4.53% in five sessions to $6.50, bigger than all but 24 of the last 120 such moves, a week after one-year-high closes of $6.85 and $6.81, and nothing new arrived on warehouse stocks, equipment lead times, or input-cost language, so the 42-day London drawdown through mid-August is still the entire case. The metal is still above the $6.37 line whose break would restore the demand-gap read, and it fell in the same window that marked down chips, small caps, and miners, which points at the rates repricing rather than at physical demand turning. Next check: stocks continuing to draw while the price stabilises above $6.37, and turbine, transformer, or cable makers reporting longer lead times rather than only larger order books. September 12 update: the price leg gave some back and nothing new arrived underneath it. Copper fell 1.78% to $6.47 after closing $6.85 and $6.81 on September 8 and 9, the highest in a year, and no fresh evidence landed on warehouse stocks or equipment waiting times, so the 42-day London drawdown through mid-August is still the entire inventory case and the grid-construction attribution still rests on the demand mix rather than on capex data. Next check: stocks continuing to draw while the price holds, and turbine, transformer, or cable makers reporting longer lead times rather than only larger order books; inventories rebuilding with a break back below $6.37 restores the demand-gap read. September 9 update: the candidate opens on an inventory series and a price extreme rather than on a forecast. Copper closed $6.85 on September 8 and $6.81 on September 9, the highest closes in a year, with London warehouse stocks drawing for 42 consecutive days through mid-August, while electric-car sales grew about 2% and solar installations are forecast to fall, so the consumer and renewables end-markets are not the source of the pull, which leaves grid and data-center construction. No standing row holds this: [[grid-hardware-bottleneck]] scores the equipment order book at GE Vernova and its peers, the AI rows score chips and financing, and the retired industrial-metal-demand-gap row was built on the opposite claim, that copper was contradicting the buildout, this is that row's premise inverted, not its reopen condition. What is unproven is transmission: a drawn-down warehouse and a high price are not yet a cost in anyone's accounts. Next check: whether stocks keep drawing and the tightness appears in manufacturers' input costs or in turbine and transformer lead times rather than only in the exchange price.

  Evidence phase: compression

• Grid Hardware Bottleneck

  ◇◇◇

  Tracks whether forecast AI electricity demand is converting into committed orders for turbines, transformers, and grid hardware, making power equipment the observable pacing item for the buildout rather than chips or memory.

  lowcandidate

  September 3 update: the candidate opens on a verified order book rather than a forecast. GE Vernova reported $24.2 billion of quarterly orders against a $176 billion backlog, which ties projected electricity demand to commercial commitments that a customer has already signed for, and it is the channel through which this window's entire power-constraint evidence arrived. No standing row holds it: the AI momentum rows score chip leadership and flows, [[ai-buildout-financing]] scores how the capex is funded, and neither carries whether the physical equipment exists to energise it. What the order book does not yet prove is delivery or revenue, since backlog can stretch out as easily as it converts. Next check: equipment deliveries, data-center energisation, and actual electricity sales confirming that announced demand becomes revenue rather than a longer queue.

  Evidence phase: compression

  • ·Next check: Equipment deliveries, data-center energization and actual electricity sales confirm whether announced demand becomes revenue.partial

Retired 8

• Dollar Plumbing Versus Gold

  ◇◇◇

  Tracks whether Federal Reserve balance-sheet runoff and a drained reverse-repo facility are removing the buffer that has absorbed reserve drain, and whether a falling dollar with a persistent metals bid is the first market read on that.

  unratedretired

  August 29 update: the invalidation completed. The dollar index rose 0.85% to 99.70, a second consecutive week of recovery, and this time gold did not hold, down 4.59% with silver 4.44% and miners 4.16%, which is the row's own stated close condition. The plumbing itself produced nothing again: the balance sheet shrank 0.22% to $6.73 trillion with the reverse repo facility drained, yet SOFR printed 3.64% against effective fed funds at 3.63% and financial conditions were unchanged at -0.57, so reserve drain is still invisible in the price of money. Next check: a sustained SOFR-to-fed-funds gap, reserve balances falling further with the facility at zero, or a quarter-end with visible repo pressure keeps this open; absent any of those, this retires next window. August 28 update: half the invalidation path ran and the other half did not. The dollar index rose 0.33% to 99.17, a recovery from 98.80, while gold rose 2.72%, silver 3.01%, and miners 7.31%, so the dollar rebuilt without gold stalling. No plumbing measure moved in the window: no reserve-balance, effective-funds, or repo evidence surfaced, so the drained facility still produces nothing in the price of money. Next check: effective fed funds drifting up relative to interest on reserves, or reserve balances falling further with reverse repo at zero; a second week of dollar recovery with gold stalling closes this row. August 22 update: three plumbing measures moved the same way in one week. The Fed balance sheet shrank 0.21% to $6.746 trillion on August 19 while overnight reverse repo drained to $0.2 billion, effectively zero, meaning further runoff now comes out of bank reserves rather than out of the facility that has cushioned it, and the dollar index fell 0.86% to 98.80 with EUR/USD +0.89% and gold +5.22%. Neither the gold row nor the rates row holds this: the gold row reads the metal against real yields and official demand, and the rates row reads the long end as a term-premium and supply question, while the observation here is that the shock absorber between them is gone. Nothing is settled, a $0.2 billion facility can sit there for months without funding stress appearing, which is why this is worth watching before repo rates move rather than after.

  Evidence phase: downdraft

  • ·Next check: gold ETF holds below 390 through September 30.partial

• Gold / Miners / Real Assets

  ◆◇◇

  Tracks gold, silver, and miners as a policy-credibility, real-rate, and official-demand signal rather than as a simple dollar mirror.

  lowretired

  October 2 update: lowered further as the map's own gold check settled against the metal and the dollar printed a fresh extreme. GLD closed 380.84 on September 30, a HIT on the below-390-through-September-30 condition, while DXY closed a one-year-high 102.10 on October 1 with EUR/USD still at a one-year-low 1.13 and the 10-year and 30-year near their year highs. Gold is still trading off the dollar and real rates rather than a separate safe-haven or official-demand story, extending the prior dollar-up, metals-down invalidation. Next check: gold ETF holds below 380 through October 8; a reclaim above 400 with the dollar fading would be the first contrary read, while another five-session metals drop alongside a further dollar rise keeps the lower in force. September 30 update: lowered again as the dollar-up, metals-down pattern repeated into fresh long-yield highs rather than healing. Across September 26-30 the dollar rose sharply over five-session windows while gold ETFs fell, DXY +0.75% with GLD -4.54%, then DXY +0.77% with GLD -4.29%, the measured block's own dollar-versus-metals signature, as the 10-year and 30-year printed one-year highs at 5.26% and 5.59%. Gold is trading off the dollar and real rates, not a separate safe-haven or official-demand story, and the prior miners-through-supply-shock invalidation still stands. Next check: gold ETF holds below 390 through September 30 and into early October; a reclaim above 400 with the dollar fading would be the first contrary read, while another five-session metals drop alongside a further dollar rise keeps the lower in force. September 16 update: a sixth consecutive lowering, and this one had every excuse available. The gold fund fell 2.38% on the week and the miners 5.99% while crude ran more than 13% in five sessions and the 10-year rose 19bp to a one-year-high 4.97%, the metal declined through a confirmed supply shock and a rising nominal yield at once, which is this row's own stated invalidation of rising yields with gold fading, now fired twice in two windows. Speculative length stands at 222,188 after the 20,024-contract cut, so the positioning reset is still incomplete. Next check: the gold fund holding above $390 through September 23; length under 200,000 contracts marks the reset complete, and only gold rising while the inflation-adjusted yield also rises restores this. September 12 update: the row failed its best available setup. A verified Gulf supply disruption, Brent at $104.30, and a third consecutive window in which the usual supports arrived produced no bid, with gold lower on the week after the September 9 pattern where a 0.80% dollar decline bought nothing. The rate leg went actively against it: the 10-year closed 4.95% and the 30-year 5.37% on September 10, both one-year highs, while inflation compensation moved about 1bp, so the market read the oil shock as a tax on growth rather than as debasement and the policy-credibility hedge did not price. Next check: gold rising while the inflation-adjusted yield also rises is the only clean restore; length under 200,000 contracts from 222,188 marks the positioning reset complete, and a fresh break below $390 on GLD with the dollar soft confirms the de-rating. September 9 update: the dollar handed gold the impulse the row was waiting for and gold did not take it. The dollar index fell 0.80% on the week while GLD slipped a further 0.24%, so for a second consecutive window the causal legs pointed the right way and price did not respond; the row's own September 4 continuation check settled HIT on a 99.093 dollar close, and the re-raise condition of GLD reclaiming $415 with a flat-to-lower dollar settled MISS. Speculative length was cut 20,024 to 222,188 (report date September 1), the first reduction since the break, which reads as a positioning reset working through rather than a live thesis. The same window paid the consumed real assets instead, with copper closing $6.85 for a one-year high and Brent through $100, so the money is not leaving real assets, only this half of them. Next check: GLD holding above $400 through September 11 CPI; length under 200,000 contracts marks the reset complete, and only a reclaim of $415 with the dollar still soft restores this. September 5 update: the lowering continues on the quality of the evidence rather than the size of the move. GLD fell 0.24%, SLV 0.94%, and GDX 0.02% into a 0.54% dollar decline with the 10-year real yield unchanged, the stronger-dollar invalidation that fired the past two windows no longer held, and price still did nothing, which says the dollar leg was doing less work in this row than the trigger assumed. The row's own next check settled HIT, with the dollar index closing 99.093 on September 4, so the firm-dollar continuation was confirmed on the day even as the broad dollar eased on the week and only the yen cross actually moved. Speculative gold length was finally cut 20,024 to 222,188 (report date September 1), the first reduction since the break, after GLD's $396.75 close marked a 7.32% five-session fall at z -1.8. Next check: GLD reclaiming $415 with the dollar flat to lower restores this; length under 200,000 contracts marks a completed positioning reset rather than a live thesis. September 2 update: the lowering extended across both weekly windows and positioning still refused to follow price. Gold fell 5.45% and then 5.41%, silver 4.49% and 4.24%, and miners 6.50% and 5.55%, against a dollar index up 0.71% to 99.56 and then 0.58% to 99.57, with inflation compensation slipping to 2.31%, so the dollar leg of the invalidation is confirmed twice over and the real-rate support that survived last window is gone. GLD closed at $396.75, a 7.32% five-session fall and a 1.8 standard-deviation move against the prior 120 sessions, already through the $400 line, while speculative gold length grew 22,717 to 242,212 net long (report date August 25), the largest position of the year against the sharpest break of the year. Next check: the dollar index holding above 99 through the September 4 jobs report, and whether GLD reclaims $400 into September 11 CPI or the next positioning report finally shows those longs being cut. August 29 update: the dollar-rebound invalidation fired and the real-yield invalidation did not, so half the lowering condition is met on the leg that mattered. Gold fell 4.59%, silver 4.44%, and miners 4.16% against a dollar index up 0.85% to 99.70 and the euro down 0.83%, while the 10-year real yield still eased roughly 4bp, supportive cause, price fell anyway, which is this row's own leveraged-flow reclassification rather than a real-rate failure. Speculative gold length rose 22,717 to 242,212 net long into the decline (report date August 25), so positioning got longer as price fell. Next check: the dollar index back under 99 with gold reclaiming its starting level near $428 on GLD re-raises; spec length breaking down alongside price lowers again. August 28 update: the run extended and the dollar leg reversed, which is the condition this row already names as the leveraged-flow tell. Gold rose 2.72%, silver 3.01%, and miners 7.31% while the dollar index rose 0.33% to 99.17, with speculative gold length unchanged at 219,495 contracts (report date August 18), miners outrunning the metal roughly three to one with the dollar firmer is price without either causal leg this week, after two windows in which both confirmed. Next check: whether miners keep outrunning the metal through the September 1 ISM release without any help from a falling dollar, which would mark the leg as flow rather than real-rate or official demand. August 26 update: the run extended past a completed setup, so crowding is now the whole question. Miners gained 13.21% into August 24 and 10.21% in the latest week with silver +4.95%, on top of a window in which the dollar and real-yield legs both finally confirmed. July core PCE firming to 3.34% from 3.29% is the first credible threat to those legs, because a dollar rebound and rising real yields would strand speculative length of 219,495 contracts (report date August 18) at nine-month highs with no causal support. Next check: the dollar holding below 99 and real yields still falling in the week after the print. August 22 update: the missing leg arrived. The dollar index fell 0.86% to 98.80 with EUR/USD +0.89%, alongside gold +5.22%, silver +6.31%, and miners +13.07%, the roughly three-to-one miner beta the map has been treating as leveraged flow now has both causal legs under it, because breakevens rose 7bp to 2.34% against a 30-year 2bp lower, meaning the long real yield fell for a second consecutive week. Speculative gold length grew 4,639 to 219,495 (report date August 18), so the position keeps building into a cause that is now established rather than half-established. The risk shifts accordingly: what was previously a long into an unproven thesis is now a long into a proven but crowded one. Next check: whether the dollar stays below 99 after the August 26 core PCE; a hot print that rebuilds the dollar and pushes real yields back up strands the miners' run. August 19 update: the real-yield leg confirmed for the first time in this leg. Gold rose 1.63%, silver 2.72%, and miners 3.07% into August 17, with miners adding 3.20% by August 19, while the dollar index slipped 0.14%, the 10-year fell 2bp, and breakevens rose 3bp to 2.27%, nominal down with compensation up is an inflation-adjusted yield actually falling rather than price moving alone. The dollar leg is still absent at -0.14%, which is noise rather than a downtrend, and speculative length stays at 214,856 contracts (report date August 11), so the position is long into a cause that is only half-established. Next check: whether the August 26 core PCE keeps real yields falling, or a firmer dollar takes the support back. August 15 update: the move decelerated sharply while length kept building. Gold rose 0.95%, silver 1.18%, and miners 1.20%, against 7.17%, 12.52%, and 18.27% the prior week, while speculative gold longs grew 24,208 to 214,856 (report date August 11). The causal legs are still missing: the real 10-year eased only about 4bp to roughly 2.36% and the dollar index rose 0.08%, so a cool CPI week produced neither a real-yield decline nor dollar weakness, which after July CPI is itself the information. Next check: nominals falling while breakevens hold or rise, or a clear dollar downtrend; rising length into fading momentum is the risk if neither arrives. August 12 update: metals led while the causal legs only half confirmed. Gold rose 7.67% and 7.17% across the two most recent weekly windows, including a 4.3% session to about $4,290, its best since February, with silver up 12.20%-12.52% and miners 18.27%-20.97%; speculative gold length reached 190,648 contracts, up 16,517 (CFTC report date August 4). The dollar leg did not fire (DXY -0.09% and +0.11%) and real yields moved only about 3bp in the supporting direction, while breakevens fell 5bp to 2.22% on August 6 and then rose 5bp to 2.27% on August 12. The July 31 week is the useful negative: the dollar fell 1.37% while gold went nowhere at +0.08%, which argues against reading dollar weakness as a structural debasement bid. Miners holding this level with soft breakevens supports the real-rate and official-demand read; a firmer dollar with miners still leading reclassifies the move as leveraged flow.

  Evidence phase: downdraft

  • ·Next check: gold ETF holds below 390 through September 30.partial

• EM / Political Stress

  ◆◆◆

  Tracks where emerging-market political or fiscal instability could spill into currencies, credit, commodities, or regional growth.

  highretired

  September 5 update: emerging-market equities rose 2.10%, the second-strongest block, but into a 0.54% dollar decline rather than against a dollar rally, so the live test this row is built on, EM risk absorbing dollar strength, was not run this week, and no fresh monthly flow report landed to extend the July reversal. Next check: three consecutive positive flow reports with spreads intact raises further; a renewed outflow month, 50bp of widening, or a failed or materially costlier frontier issuance lowers. August 29 update: the raise is narrow and rests on flow. Emerging-market equities rose 1.11% into a 0.85% dollar rally, alongside a July flow reversal that followed -$26.6 billion in May and -$17.8 billion in June, with an investor base now roughly 80% nonbank, convergent across two independent feeds. That is a partial live test passed, a dollar shock that did not crack EM risk, on one data point rather than three, and the nonbank composition means the flow can reverse faster than bank funding would. Next check: three consecutive positive flow reports with spreads intact raises further; a renewed outflow month, 50bp of widening, or a failed or materially costlier frontier issuance lowers. Higher prices and tighter fiscal room are pressuring weaker political systems.

  Evidence phase: downdraft

• Demographic fiscal drag

  ◇◇◇

  No theme summary yet.

  unratedretired

  Evidence phase: compression

• Geopolitics

  ◇◇◇

  No theme summary yet.

  unratedretired

  Evidence phase: compression

• Nuclear fuel bottleneck

  ◇◇◇

  No theme summary yet.

  unratedretired

  Evidence phase: compression

• Japan repatriation

  ◇◇◇

  No theme summary yet.

  unratedretired

  Evidence phase: compression

• Long-end funding pressure

  ◇◇◇

  Tracks term premium and long-end Treasury supply as a funding-stress signal.

  unratedretired

  Evidence phase: compression