Dated, falsifiable predictions — ours and the panel's — with their windows and status, and the calendar of the releases and events they resolve on.
| resolved | HV-011a · p=0.75 · CPI headline 3m SAAR >= 3.0 on the September-2026 data month (from 0.18 on August) CPIAUCSL saar3m >= 3.0 · touch | INDEPENDENT — the June base effect plus the oil impulse | 2026-09-01 · verified 2026-09-01 · brier 0.0625 |
| resolved | HV-011c · p=0.40 · CPI core 3m SAAR >= 2.5 on the September-2026 data month CPILFESL saar3m >= 2.5 · touch | INDEPENDENT | 2026-09-01 · verified 2026-09-01 · brier 0.36 |
| resolved | HV-001 · p=0.80 · The 2026-09-16 FOMC raises the target range (DFF steps to >= 3.80 within three sessions) DFF >= 3.8 · touch | DERIVED FROM Wang / Bianco / Rosenberg (direction); independ | 2026-09-21 · verified 2026-09-17 · brier 0.04 |
| resolved | HV-008 · p=0.45 · SOFR > IORB on three consecutive sessions between 2026-09-14 and 2026-10-31 SOFR − IORB > 0.0 · touch x3 | INDEPENDENT — the plumbing seat (Wang) holds the mechanism, | 2026-10-31 · verified 2026-10-27 · brier 0.3025 |
| resolved | HV-013a · p=0.15 · The 2026-10-28 FOMC hikes: DFF >= 4.05 on some session 2026-10-29..11-02 DFF >= 4.05 · touch | INDEPENDENT on the number; market 0.18 (rate_path.meeting_da | 2026-11-02 · falsified 2026-11-04 · brier 0.0225 |
| resolved | HV-016 · p=0.40 · SP500 posts a daily close >= 8,000 on some session 2026-10-12..11-02 (before the midterm) — the house side of owner_spx_8000_before_midterm SP500 >= 8000.0 · touch | INDEPENDENT — measured base rate 35% near the high, 49% all | 2026-11-02 · falsified 2026-11-04 · brier 0.16 |
| resolved | HV-002a · p=0.70 · DGS10 posts a daily close >= 5.00 between 2026-09-14 and 2026-11-04 DGS10 >= 5.0 · touch | INDEPENDENT | 2026-11-04 · verified 2026-09-15 · brier 0.09 |
| resolved | HV-002b · p=0.45 · DGS10 posts five consecutive closes >= 5.00 inside the same window DGS10 >= 5.0 · touch x5 | INDEPENDENT | 2026-11-04 · verified 2026-09-29 · brier 0.3025 |
| resolved | HV-002c · p=0.25 · DGS10 posts a daily close >= 5.25 inside the same window DGS10 >= 5.25 · touch | INDEPENDENT | 2026-11-04 · verified 2026-09-29 · brier 0.5625 |
| resolved | HV-003b · p=0.70 · DGS10 posts NO daily close >= 5.17 on or before 2026-12-04 (not even his weakened band) DGS10 >= 5.17 · never | INDEPENDENT | 2026-12-04 · falsified 2026-09-24 · brier 0.49 |
| resolved | HV-005c · p=0.70 · SP500 closes above 7,798.99 (a new all-time high) by 2026-12-31 SP500 > 7798.99 · touch | INDEPENDENT — 20 points UNDER a measured 90% base rate | 2026-12-31 · verified 2026-10-06 · brier 0.09 |
| resolved | HV-006a · p=0.30 · MD_XAUUSD closes <= 4,000 by 2026-12-31 MD_XAUUSD <= 4000.0 · touch | INDEPENDENT — scores AGAINST the house thesis if it lands | 2026-12-31 · verified 2026-10-30 · brier 0.49 |
| resolved | HV-014 · p=0.40 · DGS10 posts a daily close <= 5.00 on some session 2026-10-12..12-31 DGS10 <= 5.0 · touch | INDEPENDENT — measured base rate 41-51% | 2026-12-31 · verified 2026-10-22 · brier 0.36 |
| open | HV-005a · p=0.60 · SP500 closes below 7,489.72 (Dale's 07-31 mark) by 2026-11-30 SP500 < 7489.72 · touch | the LINE is Dale's (dale_20260731_tradeable_low); the p | 2026-11-30 |
| open | HV-010 · p=0.97 · PCEPILFE YoY does NOT print <= 2.50 for two consecutive 2026 data months PCEPILFE yoy <= 2.5 · never x2 | INDEPENDENT (arithmetic) | 2026-12-01 |
| open | HV-003a · p=0.88 · DGS10 posts NO daily close >= 5.52 on or before 2026-12-04 (no Gromen rupture) DGS10 >= 5.52 · never | INDEPENDENT — the mirror of gromen_bond_rupture_q4 | 2026-12-04 |
| open | HV-018 · p=0.55 · SP500 posts a daily close >= 7,800 on some session 2026-11-03..12-04 (the post-midterm bounce, before the December FOMC) SP500 >= 7800.0 · touch | INDEPENDENT — measured base rate 40-42% all starts, 53-55% f | 2026-12-04 |
| open | HV-004a · p=0.40 · BZ_F closes >= 120.00 on some session by 2026-12-31 BZ_F >= 120.0 · touch | INDEPENDENT on the number; framing shared with Bianco (oil i | 2026-12-31 |
| open | HV-004b · p=0.25 · BZ_F closes <= 85.00 on some session by 2026-12-31 BZ_F <= 85.0 · touch | INDEPENDENT | 2026-12-31 |
| open | HV-005b · p=0.25 · SP500 closes <= 7,019.09 (-10% from the 08-13 peak) by 2026-12-31 SP500 <= 7019.09 · touch | INDEPENDENT | 2026-12-31 |
| open | HV-006b · p=0.40 · MD_XAUUSD closes above 4,658.64 (the 08-25 close; a new high) by 2026-12-31 MD_XAUUSD > 4658.64 · touch | INDEPENDENT | 2026-12-31 |
| open | HV-007a · p=0.80 · BAMLH0A0HYM2 posts NO daily close >= 4.00 through 2026-12-31 BAMLH0A0HYM2 >= 4.0 · never | INDEPENDENT | 2026-12-31 |
| open | HV-012 · p=0.35 · A second 2026 hike: DFF posts a daily close >= 4.05 on some session through 2026-12-31 DFF >= 4.05 · touch | INDEPENDENT on the number; direction DERIVED FROM Howell (fi | 2026-12-31 |
| open | HV-013b · p=0.80 · DFF posts a daily close >= 4.05 on some session through 2026-12-31 (supersedes HV-012's number) DFF >= 4.05 · touch | INDEPENDENT on the number; market 0.86 for 12-09, contaminat | 2026-12-31 |
| open | HV-015 · p=0.80 · SP500 posts a daily close >= 8,000 on some session 2026-10-12..12-31 SP500 >= 8000.0 · touch | INDEPENDENT — measured base rate 82-83% | 2026-12-31 |
| open | HV-017 · p=0.29 · BZ_F closes <= 85.00 on some session 2026-11-03..12-31 (supersedes HV-004b's number, not its verdict) BZ_F <= 85.0 · touch | INDEPENDENT — measured base rate 15% all starts, 29% after a | 2026-12-31 |
| open | HV-019 · p=0.40 · SP500 posts a daily close >= 8,000 on some session 2026-11-03..12-31 (supersedes HV-015's number, not its verdict) SP500 >= 8000.0 · touch | INDEPENDENT — measured base rate 29-32% all starts, 38-45% f | 2026-12-31 |
| open | HV-009 · p=0.25 · SAHMREALTIME >= 0.50 on any data month through 2027-09 (a recession begins by 2027-09-30) SAHMREALTIME >= 0.5 · touch | INDEPENDENT — LOW-BY-CATEGORY; an NBER announcement override | 2027-09-01 |
| OVERDUE | Treasury's stated TGA rebuild to ~$950bn at end-September and a projected ~$1.05tn (±$50bn) peak in late October drains bank reserves roughly commensurately, because the RRP buffer that absorbed the 2023 rebuild is gone. Expect measurable reserve drawdown, widening SOFR−IORB, and materially higher SRF usage through October. MECHANISM: in the 2023 rebuild, MMFs rotated out of RRP into bills and reserves were largely insulated. That channel is exhausted — RRP is $2.251bn (2026-08-04), ~0.09% of its 2023 peak — so bill issuance to rebuild the TGA is funded out of reserves. Implied magnitude ~$140bn from the served 2026-08-04 level to the stated late-October peak. SERVED BASELINE (verified §0.1a, run 2026-08-05T00:30Z): TGA 910.8bn (94.3rd pct), reserves 2984.6bn (77.2nd), net_liquidity 5824.8bn (79.5th), all as_of 2026-07-29; RRP 2.251bn; SOFR−IORB 0.0bp (SOFR 3.65 / IORB 3.65, aligned 2026-08-03); SRF 0.001bn with 9 of the last 20 sessions nonzero. CONFIRMS: TGA rises toward 1.0-1.05tn through October WHILE reserves fall roughly commensurately, SOFR−IORB widens above 0, and SRF usage rises MATERIALLY above its current $1mn/9-of-20-sessions baseline. Note the SRF condition is deliberately NOT 'rises off zero' — it is already off zero, so a binary would score itself true on arrival. FALSIFIES: TGA builds toward target while reserves hold flat or rise and SOFR−IORB stays pinned. Something would then be absorbing the drain this read does not account for (RRP rebuilding, Fed balance-sheet action, foreign/official flows) — the 'no buffer left' mechanism is wrong and gets rebuilt, not defended. ALSO FALSIFIED IF: Treasury abandons the target. The 950bn/1.05tn figures are STATED INTENTIONS, not commitments; a revision at the 2026-11-04 QRA voids the setup rather than proving anything either way. CONTAMINATION RISK — read this as a liquidity FACT with a policy-agnostic cause. The TGA path is driven by seasonal tax and outflow mechanics, not by a liquidity intention. Do NOT read the drain as a policy signal. ANALYTICAL NOTE: it cuts against a 'conditions are being managed to look pleasing into the midterms' reading — Treasury is publicly planning to withdraw ~$140bn of liquidity into November, the opposite of what a pure suppression thesis predicts. Whether that is deliberate, indifferent or mechanical is NOT something the data can answer. FIRST MEANINGFUL READ: weekly H.4.1 prints through September. Observable: WTREGEN, WRESBAL, RRPONTSYD (confirm it stays ~0; a rebound restores the buffer), RPONTSYD, SOFR/IORB via derived.sofr_iorb, derived.net_liquidity. All already collected — no new build required. | read 2026-08-05 (QRA_BASELINE.md, Treasury sb0590) | 2026-11-04 · -2d |
| DUE SOON | CPI upward jumps in spring and October 2026 Spring leg: May CPI 4.2% headline is consistent — score after full-year path. | Bianco, Dec-2025 | 2026-11-30 · 24d |
| DUE SOON | The late-July 2026 tradeable low in equities does not prove durable over a 3-4 month horizon (stated 2026-07-31 with SPX at 7,489.72). id: dale_20260731_tradeable_low. RATIONALE FOR SEEDING (rule 3): this is Dale's cleanest UNCONDITIONAL dated call and it resolves on a series already collected, which the existing dale_1998_correction entry does not - that one is CONDITIONAL on the Sept/Oct FOMC flipping the policy cycle. Seeded so a skate-through on the conditional cannot be read as a win on the unconditional, and so the panel has at least one Dale line that cannot resolve into 'the branch never activated'. CONFIRMS: SP500 trades below 7,489.72 at any point through 2026-11-30. FALSIFIES: SP500 holds above 7,489.72 for the whole window. Context at seed (served, build 8b79e3cf21ca): SP500 7,641.16 (2026-08-20, 99.5th pct of 10.1y), peak 7,798.99 on 08-13, so the call is currently +2.0% against him; MARGIN_DEBT 1,417,225 (2026-07) after a +99.7% three-year run and a first -5.6% m/m decline off the June peak; positioning.leverage 94.6th pct 'crowded'; cot_crowding ES crowded long. His stated mechanism is the R-star/capital-competition channel (Framework 9), not a growth scare - score the price, not the mechanism. Cross-references: dale_1998_correction (same analyst, different structure - DO NOT DOUBLE-COUNT), howell_2026_next_move_hike, read_20260818_real_led_supply. | Dale (42 Macro) - Yahoo Finance Opening Bid roundtable, 2026-07-31 (primary, captions) | 2026-11-30 · 24d |
| DUE SOON | A momentary bond-market rupture within two to three months — the 10-year running from 4.8 through 5, 5.5, 6, 7. 'It's not a tornado watch, it's a tornado warning.' id: gromen_bond_rupture_q4. LEAN: fiscal dominance / nonlinearity. COLUMN [s2, s4] as a flow: the rupture itself is S2's signature (yields up, fast-intervention branch), while HIS mechanism — Bessent has already intervened via buybacks and it fails anyway — is S4. Seeding it to one column would lose half the claim. OBSERVABLE: DGS10 daily close. BASELINE 4.77, as_of 2026-09-03 (build 9f9176311288). CONFIRMS: DGS10 closes at or above 5.52 (+75bp from baseline) at any point on or before 2026-12-04. WEAKENED: a high of +40bp to +75bp — direction his, magnitude not. FALSIFIED IF: DGS10 never exceeds +40bp from baseline inside the window. The +75bp bar is deliberately BELOW his illustrative 7%: the claim is "a rupture" and the 7% was his illustration of one. Holding him to 7% would be a harsher test than he stated. *** BINDING CROSS-REFERENCE — READ BEFORE SCORING EITHER. *** This entry resolves AGAINST his own Framework 2 ceiling line ("Bessent is never going to let the 10y over 4.7–4.8", scored on five consecutive DGS10 closes >= 4.80 with no new Treasury or Fed action). If this entry CONFIRMS, the ceiling line FAILS on its own pre-written rule. NEITHER OUTCOME MAY BE REPORTED AS A WIN ON BOTH. He revised the ceiling from an intervention thesis into a failure thesis in this same session — he is entitled to have revised it, and his profile records that he did, but not to hold both versions at once. Status at seeding: DGS10 4.75 / 4.79 / 4.79 / 4.77 across 2026-08-31..2026-09-03 — no close at or above 4.80, so the ceiling's falsifier has NOT fired. SCORES A LIVE PANEL DISAGREEMENT, which is the best reason to hold it: Alden, in the same room, called the degradation ORDERLY — no MOVE event, decent liquidity — and Bessent's intervention PREMATURE. One entry, both sides. ERROR ACCEPTED: type-2 — a rupture driven by something other than his private-credit / insurer mechanism scores as a win on direction. Accepted because that mechanism is unscoreable here (insurer asset allocation is not collected — check file §6) and the alternative is no entry at all.
| Gromen (FFTT) — BTC Sessions joint session, published 2026-09-04 (primary; checks in data/panel/checks/2026-09-07-gromen-alden-btc-sessions-joint.md) | 2026-12-04 · 28d |
| DUE SOON | The 2026-11-04 refunding announcement, not the buyback programme, is the lever that brings the 10y down (the November-2023 analogue). id: rosenberg_nov_refunding. LEAN: plumbing — the supply-composition school rather than the Fed-path school. MECHANISM: as in Nov-2023, a refunding that shifts issuance toward bills or holds coupon sizes below expectation relieves duration supply and the long end rallies on the announcement itself. OBSERVABLE: issuance_vs_guidance and AUCT_*_OFFER_BN (both built 2026-09-02, so this scores MECHANICALLY the day it lands); DGS10 daily_close. CONFIRMS: DGS10 falls at least 25bp from its 2026-11-03 close within 30 calendar days AND the refunding is the identifiable lever — issuance_vs_guidance registers a departure, or coupon OFFER sizes come in below the same cycle position one quarter back. FALSIFIED IF: DGS10 is unchanged-to-higher 30 days after 11-04 (a fall of 10bp or less counts as unchanged). ALSO FALSIFIED IF the 10y does fall 25bp while the refunding held sizes flat and the move traces to the buyback programme or the Fed path — that outcome refutes his MECHANISM, which is the whole content of the claim, and it must not be allowed to score as a win on direction alone. ERROR ACCEPTED: type-1 — a 25bp rally on a macro shock inside the window will look like confirmation. The issuance leg is what guards against it; score both or score neither. SIZE THE GUIDANCE QUESTION ON OFFER, NEVER ON SIZE: AUCT_*_SIZE_BN stores totalAccepted = offeringAmount + somaAccepted, and the SOMA add-on ran 10.62bn of a 52.62bn 10Y total on 2026-08-12. Use AUCT_*_OFFER_BN and compare the SAME CYCLE POSITION one quarter back — a reopening is smaller than its own new issue by design. CONTEXT AT SEED (served, build 78f16b8c334e): DGS10 4.75 (2026-08-31, 58.9th pct of 40.1y); all seven tenors read as_guided at the 2026-09-02 build. CROSS-REFERENCES: read_20260820_long_end_repression (resolves at the same 2026-11-04 event, OPPOSITE lever — it scores Treasury's SUPPORT escalation, this one scores the ISSUANCE decision; a single 11-04 action can confirm one and not the other). DO NOT DOUBLE-COUNT.
| Rosenberg (Rosenberg Research) — David Lin session, 2026-09-01 (owner-supplied draft, processed 2026-09-02) | 2026-12-04 · 28d |
| open | If the market keeps a 10-28 hike priced and the Fed declines: 'Fine, then I'll take the 10-year to 550. I will just abandon the bond market… we don't raise rates and bond yields go to the moon.' id: bianco_no_oct_hike_10y_550. disclosed-interest (he bought bonds 09-28..~10-01, by his own account). SUPERSEDES the proposed-optional bianco_october_skip_lifts_long_end (09-24): same event, same branch structure; this one carries HIS magnitude (5.50) where that one carried ours (+25bp). Seed one, not both. THE WINDOW IS OURS (RULINGS section 30 class — a windowless conditional keyed to a dated event); the magnitude is his, so the 09-24 note's objection that section 30 does not cover supplying a magnitude no longer applies. LEAN: policy-tightening, from the bond-vigilante seat; on the skip branch it lands beside Gromen. ACTIVATION: a HOLD or CUT at the 2026-10-28 FOMC (DFF unchanged or lower through 2026-10-30) WHILE rate_path.meeting_dated priced the hike at p_move >= 0.5 on the 2026-10-27 close. His own escape hatch: "their bail-out is weak data goes under 50%… and then they've given the opening to not do it" — a skip with the hike NOT priced is NOT ACTIVATED, never wrong. A hike is NOT ACTIVATED for this entry (the hike branch stays in the 09-24 seed's form if the owner prefers to keep it; it is not restated here). OBSERVABLE: DGS10 daily close (cash series — section 21, never the stitched future); rate_path.meeting_dated and DFF for activation. CONFIRMS: once activated, any DGS10 close at or above 5.50 through 2026-12-08. FALSIFIED IF: activated and no close at or above 5.50 by 2026-12-08. PARTIAL: activated and a close at or above the 2026-10-27 base + 25bp but below 5.50 (the 09-24 form) — report as partial, not as a confirm. RULINGS section 28 GUARD: does not amend, offset or re-open bianco_hike_caps_long_end (resolved FALSIFIED 09-16) or bianco_10y_near_peak (a 5.50 close confirms this entry and does NOT falsify that one — its ceiling is 5.75; a close above 5.75 falsifies that one and confirms this one: report both in one breath). ERROR ACCEPTED: type-1 — a rise to 5.50 for fiscal or energy reasons after a skip scores as a win though his mechanism is Fed credibility. CONTEXT AT SEED (served, build 24b6f00ce734): DGS10 5.28 (2026-10-02), highest close to date 5.29 (2026-09-30); rate_path.meeting_dated p(hike, 10-28) = 0.20 (2026-10-02) — THE ACTIVATION CONDITION IS FAR FROM MET at seed; December p = 0.86. DFF 3.88 (2026-10-01). CROSS-REFERENCES: gromen_bond_rupture_q4 (DGS10 >= 5.52 by 2026-12-04 — one breach can confirm both; do not double-count); bianco_10y_near_peak (above); rosenberg_nov_refunding and rosenberg_gridlock_yields_down (the counter-claims — a hold followed by a rally scores for them and against this entry); alden_no_failed_auction. CHECKPOINTS: 2026-10-27 close (base and activation read); 2026-10-28 FOMC; 2026-11-04 refunding; 2026-12-08 close.
| Bianco (Bianco Research) — Rational Dissent ep. 010, recorded 2026-09-25 (R3DoJATEV7k, primary, his own podcast); the same conditional without a magnitude ~09-17 and 09-21 (checks data/panel/checks/2026-09-24-bianco-fomc-week.md §5); checks in data/panel/checks/2026-10-05-bianco-master-investor-bond-pivot.md §7 | 2026-12-08 |
| open | The Fed does not move before the midterm — it holds on 2026-10-28 — and 'they're really going to hike rates in December' (2026-12-09). id: rosenberg_oct_hold_dec_hike. disclosed-interest (section 28 conduct trigger — the session opens on ROSY and its NYSE listing). LEAN: policy-tightening timing, from the cyclical-counter seat — his READING of the committee (the doves "doubled up"), not his normative view, which is that every hike is a mistake. Score the reading. OBSERVABLE: the FOMC decisions of 2026-10-28 and 2026-12-09; DFF daily_close as the served footprint (DFF 3.88 since 09-17; a +25bp hike prints DFF >= 4.05 within days of the decision). CONFIRMS: no change on 2026-10-28 AND the target range is raised on 2026-12-09 (first DFF print >= 4.05 falls after 12-09, with no print >= 4.05 between 10-28 and 12-08). FALSIFIED IF: a hike on 2026-10-28 (the timing clause is the content — "not before the midterm"); OR a hold or cut on 2026-12-09 after an October hold. No PARTIAL: he stated both legs as one sequence. An intermeeting move resolves on the same clauses by date. ERROR ACCEPTED: type-1 — a December hike confirms him whether or not the midterm calendar was the reason, and whether the doves he named were the swing votes. CONTEXT AT SEED (served, build 24b6f00ce734, rate_path.meeting_dated contracts as of 2026-10-02, collector 2026-10-05T13:20Z): DFF 3.88 (10-01); 10-28 hold priced, p(hike) 0.20 (X26 implied 3.93); 12-09 hike priced, p 0.86 (Z26 day-weighted 4.13, F27 4.145). CONFIRMS is priced, not true, at seed — the same shape as the September-hike entries. CROSS-REFERENCES: yardeni_2026_fed_second_hike (DFF >= 4.05 by 12-31 — a December hike confirms both; report as ONE hike, never two wins); HV-012 (house p 0.35 on a second hike — the other side); rosenberg_2027_disinflation (its "hikes AND the labour market does not crack" clause is where his normative view is scored); rosenberg_sep_hike (resolved — the reversal sentence travels with any report of this seat's hike calls). CHECKPOINTS: FOMC 2026-10-28, 2026-12-09.
| Rosenberg (Rosenberg Research) — RiskReversal Media, recorded 2026-09-16 after the FOMC press conference, uploaded 2026-09-17 (primary, ZQupKFnSEEI; checks in data/panel/checks/2026-10-05-rosenberg-talking-markets-and-riskreversal.md §9) | 2026-12-09 |
| open | Forward guidance producing a durably negative real floor funds rate | Dale, Jan-2026 deck | 2026-12-31 |
| open | Q4-2026 food/fertilizer shock | Johnson, May-2026 | 2026-12-31 |
| open | SPR flow reversal ~Aug-2026 puts a floor under crude into Q4 | Owner, seed | 2026-12-31 |
| open | Core PCE (PCEPILFE) prints <= 2.5% YoY for two consecutive months in 2026 Falsifies the sticky-core leg. Observable: PCEPILFE YoY (calendar-indexed, F-01). | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | No 10y/30y auction tails in the extreme band through 2026 REINSTATED live 2026-07-15 (v2.2 §3.3, T2): auction_internals is now built and served — per-tenor tail = high yield minus the same-day CMT close. Scored against this CMT-close PROXY tail until the owner pastes real when-issued tails via data/manual/overrides.yaml#auction_tail_manual (which supersedes and relabels it per auction). Deterministic + free from TreasuryDirect; the spec's 'tails are qualitative' call is disputed (review I-8) — for a fiscal-dominance framework these ARE the direct observable of the mechanism. | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | Broad dollar (DTWEXBGS) holds its 2025-26 range through 2026 — no reserve-currency debasement break The dollar-bull call, previously untracked. Observable: DTWEXBGS range vs 2025-26. | Johnson counter / dollar-bull | 2026-12-31 |
| open | Real GDP positive every quarter of 2026 WITHOUT a Fed balance-sheet re-expansion > 5% SAAR Falsifies 'the regime REQUIRES a liquidity-driven melt-up'. Observable: GDP + policy_flow SAAR. | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | Productivity (OPHNFB) YoY stays > 1.5%, keeping unit labor costs contained The soft-landing MECHANISM made falsifiable. Observable: OPHNFB (wired via the benign co-anchor). | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | Gold/oil ratio mean-reverts below 45 and holds a quarter Falsifies a STRUCTURAL debasement bid. Observable: gold_oil ratio + band. | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | Bank credit (TOTBKCR) YoY positive with SLOOS standards easing and no credit event, through 2026 Healthy credit creation. Observable: TOTBKCR (wired via co-anchor) + HY OAS + DRTSCILM (SLOOS net % tightening C&I, collected since v2.3 — easing = <=0; +8.1 i.e. mild tightening at 2026-Q2, so the 'standards easing' half is checkable and currently NOT met). | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| open | Establishment/consensus: US soft landing holds — no NBER recession in 2026 and core PCE glides toward ~2.5% Adds a non-debasement voice — the panel was all debasement-sympathetic. Observable: NBER + PCEPILFE. | Establishment / consensus (counter-panel) | 2026-12-31 |
| open | Private-credit / PE-insurer stress transmits to PUBLIC markets in 2026 — HY OAS breaks its calm band and/or bank-NBFI lending rolls over, as BDC non-accruals and the pasted PC default index climb. COUNTER (currently winning): stress stays idiosyncratic and gated. HY OAS ~12th pctile (no public leak); Chicago Fed / IMF read 'contained'; default estimates disperse ~2-6% with ~65% distressed restructurings; insurers ~80% IG on CLO holdings and well-capitalized. FALSIFIER (systemic side): HY OAS holds sub-~4% through 2026 with no gate-driven forced-selling event visible in public spreads. Observable: BAMLH0A0HYM2 + credit_transmission.leak_state. | Owner, seed (credit-transmission gauge v2.4) | 2026-12-31 |
| open | Under Warsh, the Fed restarts/extends balance-sheet shrinkage (QT) even after the Dec-2025 bill-buying expansion, leaning on heavier bill issuance to cushion it — net effect is curve-steepening and tighter conditions through the balance sheet even if the policy rate stays put. LEAN: plumbing / roughly neutral (the one non-debasement add). FALSIFIER: WALCL keeps rising (expansion continues) rather than flat/declining, or reserves climb with no drain. CURRENT READ: not yet visible — WALCL hasn't shrunk ($6.74tn), net_liquidity still ~85th pctile, SAAR +2.2%; forward call the balance sheet confirms or denies. Observable: WALCL, WRESBAL, bill_share / FISCAL_MKT_BILLS_BN, SOFR-IORB, T10Y2Y. His separate 'market underpricing Trump cuts' view is NOT seeded (likely predates Warsh's hawkish turn — re-date before use). | Wang (Fed Guy / Monetary Macro) | 2026-12-31 |
| open | The ~65-month global liquidity cycle peaked around Aug-2025 and is rolling down into 2027; the 2026-28 refinancing wall (~$40tn rollovers by 2027) absorbs liquidity, making 2026 tough for risk assets and favoring commodities / gold / long-duration bonds. LEAN: caution / late-cycle. FALSIFIER: SP500 keeps making new highs through 2026 with US net_liquidity holding >70th pctile -> falsifies the liquidity-driven risk-asset downturn (for the US). CURRENT READ: split scorecard — equity call currently WRONG (SPX ATH, US net_liquidity ~85th pctile, positive SAAR); commodity/gold rotation RIGHT (gold ~95th pctile, CPER ~99.8th, gold_oil elevated); log both halves. SCOPE GAP: Howell's real signal is GLOBAL liquidity; TripWire net_liquidity is US-only (WALCL-TGA-RRP) — the divergence between his global GLI (down) and US net_liquidity (up) is itself the read; global-liquidity proxy is a panel_gap candidate. Observable: net_liquidity, MD_XAUUSD, CPER, gold_oil, SP500. | Howell (CrossBorder Capital, GLI) | 2026-12-31 |
| open | The ~2.5-3% 2026 growth priced into markets is too high; the oil shock (Brent >$110 earlier in 2026) feeds inflation and drags real growth, and both stocks AND bonds are mispricing it — both can struggle as yields rise. LEAN: caution (NOT a benign counterweight — an earlier characterization of Elliott as data-first counterweight was stale and corrected 2026-07-22). FALSIFIER: growth_fundamentals holds expansion AND SP500 holds gains through 2026 with no oil-driven inflation re-accel -> the 'both assets struggle' call fails. CURRENT READ: challenged — growth_fundamentals 10/10 expansion, claims ~3rd pctile, equities ATH all cut against 'growth won't hold'. His oil-> inflation leg partially aligns with the existing July-CPI re-acceleration entry (window 2026-08-12) — CROSS-REFERENCE, do not double-count. Observable: growth_fundamentals vs regime.growth_norm, SP500, DCOILWTICO/CL_F, inflation_fundamentals, DGS10. | Elliott (Unlimited Funds) | 2026-12-31 |
| open | IF the Sept or Oct FOMC flips the monetary-policy cycle to a headwind (hike + balance-sheet tightening), a 1998-style 15-20% SPX correction follows over ~2-4 months, uptrend resuming after. id: dale_1998_correction. CONDITIONAL on the flip occurring — his own unprompted falsifier: Fed doesn't tighten AND the tape trends higher => 'skate through and have a fantastic year.' Checkpoints: FOMC 2026-09-15/16 primary. Context at seed: distance_to_flip fragility false, SP500 +8.67 sigma from its 7,163 flip level; no funding stress (sofr_iorb 0bp, SRF $0); futures ~2/3 no-move — HIS FALSIFIER BRANCH IS THE ONE CURRENTLY GAINING PROBABILITY. Score the conditional honestly: no flip => the correction leg never activates, and that outcome routes to skate-through, not to 'wrong'. | Dale (42 Macro) — Thoughtful Money session 2026-08-05 (primary); promotes dale-darius-42macro.md open item #1 as previously framed | 2026-12-31 |
| open | The S&P 500 closes at or above 8,400 by year-end 2026 — 2027 consensus EPS ~$415 x a 20.24 forward P/E; 'FEMO not FOMO' id: yardeni_2026_spx_8400. LEAN: benign / consensus — the counter that costs the debasement school points if it lands. MECHANISM: earnings momentum ("FEMO") with a falling forward P/E; bond yields capped at 4-5% by bill issuance and a strong economy. OBSERVABLE: SP500 daily_close. CONFIRMS: any SP500 close >= 8,400.00 on or before 2026-12-31. FALSIFIED IF: no such close. ERROR ACCEPTED: type-1 — a September/October sell-off he calls "a buying opportunity" does not void it. CONTEXT AT SEED (served, build 416950ac101f, 2026-09-11): SP500 7,656.98 (99.1st pct of 10.2y), ATH 7,798.99 on 08-13; base rate for +9.7% in 76 sessions 29% full / 32% since 2022 (measured 2026-09-13). SEEDED 2026-09-15 on owner OK; the mark that day was 7,619.98 (2026-09-14 close) — recorded because the entry landed two sessions after the block was written, NOT because the bar moved: the bar is an absolute level and never moves. CROSS-REFERENCES: HV-005c (new ATH by 12-31, 70%), HV-005b (-10%, 25%), dale_20260731_tradeable_low, dale_bubble_peak_2027h2. His own record: the 2025 target moved five times and was scored "Mission Accomplished" at 6,901 against a 7,000 target — score this one on the number. REVERSAL RECORDED 2026-09-24 (RULINGS section 28 - the entry is NOT amended): on 2026-09-15 (QuickTakes 'Proceed With Caution', the same day this entry went live) Yardeni moved the 8,400 target 'to mid-2027' and set a new year-end 2026 target of 7,900 ($425 x 18.6 forward P/E), restated on Bloomberg TV 09-17 ('more likely by the middle of next year'). This entry still resolves on any SP500 close >= 8,400.00 by 2026-12-31. Report the pair as one reversal: a miss here is a miss on the 08-31 call he withdrew on 09-15, and a hit is a hit on a call he no longer held. The 7,900 view is deliberately NOT seeded (owner, 2026-09-24). Check: data/panel/checks/2026-09-24-yardeni-sessions03-06-target-cut.md.
| Yardeni (Yardeni Research) — weekly briefing recorded 2026-08-31 ('Warsh World', pub. 09-04); the 8,400 target first set 2026-08-11 per the profile | 2026-12-31 |
| open | Gold reaches $5,000 by year-end 2026 — central banks buy the dips; a retest of 4,000 doubted id: yardeni_2026_gold_5000. LEAN: benign-seat gold bull — unusual, and useful: a gold call from the consensus/productivity seat rather than the debasement school. MECHANISM: central-bank buying on dips outweighs the drag from rising short rates. OBSERVABLE: MD_XAUUSD daily_close (never live_overlay — gold_falsifier_2026 states why). CONFIRMS: any close >= 5,000.00 on or before 2026-12-31. FALSIFIED IF: no such close. SECONDARY (his "doubt a retest of 4,000"): a close <= 4,000.00 inside the window is recorded against him qualitatively. ERROR ACCEPTED: type-1. CONTEXT AT SEED (served, build 416950ac101f, 2026-09-11): MD_XAUUSD 4,348.87 (96.7th pct of 18.9y), 08-25 high 4,658.64, DFII10 2.55 (98.5th pct), COT GC crowded long 98.7th pct; base rate for +15% in 76 sessions 17% full / 28% since 2022 (measured 2026-09-13). SEEDED 2026-09-15 on owner OK; the mark that day was 4,298.35 (2026-09-14 close). Absolute bar, unmoved. CROSS-REFERENCES: gold_falsifier_2026 (different marks, 2027-03-01), HV-006a/b, Gromen's "$100, $200, $300 days" (undated). Do not double count with the house view: HV-006b resolves at 4,658.64, this at 5,000.
| Yardeni (Yardeni Research) — weekly briefing recorded 2026-08-31 ('Warsh World', pub. 09-04): 'we're still using 5,000 by the end of the year... doesn't come with a money back guarantee' | 2026-12-31 |
| open | The Fed is not one-and-done: at least one more hike in 2026 after September ('likely another one or two increases this year') id: yardeni_2026_fed_second_hike. LEAN: caution / policy-tightening, from the benign seat: he reads the hikes as a strong economy plus higher-for-longer oil, not as a policy error. MECHANISM: oil above $100 seeping into core ('second and third order effects') keeps Warsh hiking. OBSERVABLE: DFF daily (served). CONFIRMS: any DFF print >= 4.05 on or before 2026-12-31 (the HV-012 bar). SECONDARY (his 'or two'): a DFF print >= 4.30 is recorded for him qualitatively. FALSIFIED IF: no DFF print >= 4.05 through 2026-12-31. ERROR ACCEPTED: type-1 — a hike that slips to January 2027 scores FALSIFIED; 'this year' is his framing. CONTEXT AT SEED (served, build 95444956b014, rate_path.meeting_dated as_of 2026-09-23, collector 2026-09-24T13:22Z): DFF 3.88 (09-22); 10-28 hike priced p 0.70 (X26 implied 4.055); 12-09 p 0.788, Z26 day-weighted 4.25. CONFIRMS is not already true at seed. CROSS-REFERENCES: HV-012 (same bar, house p 0.35 stated 09-12 — this seat is the other side of the house view; do not double-count them as one call); howell_2026_next_move_hike (the NEXT move, resolved 09-16); wang/rosenberg/bianco sep-hike entries (09-16 only). CHECKPOINTS: FOMC 2026-10-28, 2026-12-09. Check: data/panel/checks/2026-09-24-yardeni-sessions03-06-target-cut.md.
| Yardeni (Yardeni Research) — Bloomberg TV 2026-09-17 (primary): 'we're likely to have another one or two increases this year'; also his webcast 09-14 and Thoughtful Money ~09-17 | 2026-12-31 |
| open | The 10-year has topped — the stall at the highs, then the rally off yesterday's high, was the turn; DGS10 does not close above the 5.31 high again. id: owner_10y_yield_topped_531. LEAN: benign / bond-bull — the growth-scare side; against the fiscal-dominance column. MECHANISM (the owner's, by price action): a stall at the highs, then a reversal the day after CPI. Context given in session: the AI-financing credit scare (CoreWeave 2032s below 78, 5y CDS 1,118bp; KOSPI -7.6%) overrode a hot September CPI and PPI (+1.3% m/m, 6.3% y/y) — press, not served. The mechanism is not scored; a top for any reason confirms. THE WINDOW IS OURS, NOT HIS — the owner stated no date (RULINGS section 30 class). 12-31 is ~53 sessions, spanning the 10-28 FOMC, the 11-02 QRA / 11-04 refunding, the midterm and the 12-09 FOMC. OBSERVABLE: DGS10 daily_close (H.15). The live ^TNX overlay is not a close; the 10-14 intraday ~5.33 was a cash mark and is not the level. CONFIRMS: no DGS10 close above 5.31 from 2026-10-15 through 2026-12-31. FALSIFIED IF: any DGS10 close above 5.31. REPORT BESIDE THE VERDICT: the lowest close in the window (a top that drifts to 5.20 and a top that runs to 4.80 both confirm, and they are different results). BASE RATE (measured 2026-10-15 on FRED DGS10 1962-01 -> 2026-10-13; trigger = a close >= 20bp under a 120-session high, 53-session forward window, de-clustered): the yield closed back above the high in 29/54 (54%); in the rising-rate eras (1962-81, 2021-26) 15/25 (60%); 4/6 since 2021. Confirmation runs ~40-46% before any view. No house probability is stated (owner's call). ERROR ACCEPTED: type-1 — a Treasury/Fed lever or a recession that caps yields scores for the owner though the reason differs. CONTEXT AT SEED (served, build 5e89d1234b64): DGS10 5.19 (10-14 close); highest close 5.31 (10-05); live ^TNX 5.085 (10-15 18:58Z, flagged stale); DGS30 5.52 (10-14); 20-session decomposition to 10-13 +24bp nominal, real +28 / breakeven -4 (real-led); DFII10 2.88; rate_path 10-28 P(hike) 0.34 (10-14); HY OAS 3.16 (+40bp/20s, 10-13); CCC-BB 1,076bp (fired, top of a 3.1y window). CROSS-REFERENCES: bianco_10y_near_peak (same side, looser bar 5.75 through 2027-03-31 — this entry's falsification does NOT falsify his); gromen_bond_rupture_q4 and HV-003 (>= 5.52 by 12-04, opposite side — a close above 5.52 falsifies this entry too; score the number once); dale_10y_fair_value_6 (opposite side); owner_private_credit_transmits (the mechanism). CHECKPOINTS: the 10-15 close and the 10-16 decomposition (real-led vs breakeven-led decline); 10-28 FOMC; 11-04 refunding; HY OAS +100bp/20s (the credit_transmission gate). RESTATED BESIDE, NOT AMENDING (RULINGS section 28), 2026-10-16 in chat: "I don't have strong feelings about the 10 yr bottom." The entry stands and scores on its terms; the owner may withdraw it. Same message, a counter-view the owner relayed, verified against the source: Jason Shapiro (Crowded Market Report, "Stocks, Gold, Silver & Bitcoin Rally While Bonds Continue To Crash!", youtube ygzh9dGzkYg, recorded 2026-10-05 — the day of the 5.31 close, BEFORE the rally): "while everybody thinks the contrarian trade is to buy them, it can't be the contrarian trade if everybody thinks it's going to buy them ... if this ends up being a market turn in the bond market, it would be the first time ever that I saw so many people participating in a market turn and having it work. So, I don't think this is the low for bonds ... Doesn't mean bonds can't have a rebound here for a few weeks." No level, no window. His expression was the Russell (crowded short stock index, confirmed 10-02/10-05), not a bond short. Not seeded: no window of his own, and any DGS10 close above 5.31 already scores his side here — counting it twice would double-count one close. Served check of his premise (cot_crowding, 2026-10-06): ZN large-spec net -11.9% of OI, 58th pct 1y / 60th 3y — not crowded long by COT; his "everybody" is sentiment, which is not collected. Re-read on the 10-13 report. RE-READ 2026-10-17 (cot_crowding.by_trader, build e23a5fefd6e9, report 2026-10-13): the large-spec number was the wrong crowd (basis-trade shorts). ZN ASSET MANAGERS are net long 49.81% of OI, the 100th pct of 20.3y and a record, +0.81 over 4 weeks. His premise (everyone is long) HOLDS on positioning. His inference (crowded, so not the low) has not held in this data: asset-manager robust longs were followed by bonds falling 43% of the time vs a 51% base at 13w (docs/evidence, the trader-split measurement of 2026-10-16). RESTATED AGAIN 2026-10-18 in chat: "i'm low conviction on this call. but i'll own it for fun. jason shapiros comments are ringing in my ear. when everyone is in concensus and very vocal about the pivot of a trending asset, it seems like a great place for everyone to wrong again" — the owner now leans toward Shapiro's side and KEEPS the entry. Base rate (analyst, 2026-10-15, DGS10 1962-): after a >= 20bp drop off a 120-session high, the yield retook the high ~54% of the time, so the entry is about a coin flip that leans slightly against it.
| Owner, seed — analyst session 2026-10-15 (in chat): 'it looked like the 10 yr was bottoming with the stall at the lows and then the rally from the bottom yesterday' | 2026-12-31 |
| open | The bond market forces a Fed-Treasury merger into explicit yield curve control - the 'Fedsury' - probably this year, probably in the lame-duck Congress between the midterms and January. id: dale_fedsury_lameduck. LEAN: debasement / plumbing — adds to the same side as the standing Durable YCC entry, at a much nearer date. MECHANISM: the gap between spot 10y and his 5.83% fair value is a debt the market collects; collecting it pressures Bessent and Warsh until yield management is formalised. WHAT COUNTS — an ACTION, announced: (a) an FOMC statement or Fed announcement naming an explicit yield target or yield RANGE for Treasury securities, or (b) a formal Treasury-Fed joint statement or accord on yield management. Logged in data/events/events.jsonl as a named_policy_response. WHAT EXPLICITLY DOES NOT COUNT, and this is the load-bearing half: continued reserve-management purchases, an enlarged buyback programme, a bill-heavy issuance mix, or any further arrival of the SUBSTANCE of coordination without the announcement. That is the "already YCC" reading of Session 08, it is a live definitional dispute in his own profile, and letting it resolve this entry would make the entry unfalsifiable — the exact defect that put Paradigm D beyond scoring. FALSIFIED IF: no such announcement by 2027-01-31. HIS HEDGE, RECORDED SO SCORING STAYS FAIR: he said it "probably won't be [imminent]" and "if they don't, they don't, but I think it's a risk" in the same breath as the window. Score the TIMING claim only. A FAILED here does not touch the structural when-not-if claim, which lives in the standing Durable YCC entry and runs to 2029-12-31. ERROR ACCEPTED: type-1 — an announcement-only test can miss a de-facto peg that is never announced. Deliberate: an unannounced peg is unfalsifiable by construction, and his own Framework 13 ambiguity doctrine predicts substance-without-announcement, so the two entries cover the two branches between them. CONTEXT AT SEED (served, build 940a8c83bc33, 2026-09-03T20:19Z): DGS10 4.79 (2026-09-01, 59.6th pct of 40.1y) — note the rates lane was a session behind at seed; DGS30 5.27 (09-01). His fair value 5.83% and his quoted 112bp gap are HIS numbers, unscoreable by construction (five proprietary models, rule 4). CROSS-REFERENCES: "Durable YCC" (window 2029-12-31) and "Fed transition from funds-rate to repo-rate targeting" (2028-12-31) — this is a NEARER, more specific version of the same construct, NOT a second event. DO NOT DOUBLE-COUNT: if the announcement lands, it resolves this entry and advances those; if it does not, only this one fails.
| Dale (42 Macro) — Macro Minute excerpt, YouTube cC2abmDDIqQ, uploaded 2026-09-03 (primary but a PARTIAL EXCERPT; capture in the profile, Session 11) | 2027-01-31 |
| open | Gold's bid is monetary and fiscal rather than rate-driven, and it will therefore hold its level through a sustained real-yield state at or above the 95th percentile — the condition under which the rate-driven reading predicts reversion. id: gold_falsifier_2026. LEAN: debasement — this is the entry that costs the base case points if gold is a rate asset. MECHANISM: if gold is a claim on the debt arithmetic rather than on the real rate, a sustained 95th-percentile real yield should not force reversion. If it is a rate asset, six months at that level is the condition under which it gives the divergence back. OBSERVABLE: MD_XAUUSD daily_close; DFII10 daily_close. NOT live_overlay for gold — the overlay advances its value without advancing as_of, so a resolution read off it is not reproducible. HOSTILE-RATE GATE: DFII10 closes >= 2.14 (p95 of the aligned span from 2007-11-02) on at least 80% of sessions in the window. CONFIRMS: gate held AND MD_XAUUSD >= 4448.80 at 2027-03-01. FALSIFIED: gate held AND MD_XAUUSD <= 3559 (-20% from seed) at 2027-03-01. Calibrated to the -23.7%/126-session in-sample precedent of Feb-Jul 2026, not chosen. PARTIAL: gate held, gold between -20% and 0% — direction the counter-thesis's, magnitude not. VOID (reported as proving nothing, never as a win): DFII10 under 2.14 on more than 20% of sessions, or under 1.96 (p90) at resolution. Gold rallying into FALLING real yields is the ordinary mechanism and is evidence for neither school. Named in advance so hindsight cannot convert it into a confirmation. ERROR ACCEPTED: type-2 — the 80% gate and the -20% magnitude will let a genuine slow reversion resolve PARTIAL rather than FALSIFIED. Deliberate: a falsifier that fires easily against the owner's largest conviction will be argued with, and one that gets argued with has already failed. DOES NOT SETTLE: the debasement thesis (this is one entailment over six months); whether the level is right (§4 refused to supply a valuation); anything about a fiat cycle — the aligned span starts 2007-11-02, contains one gold bear market and no 1970s. SEED MARKS (served, build fc73b902d1d2, snapshot 2026-09-01T00:46:24Z): MD_XAUUSD 4448.80 (2026-08-31), DFII10 2.42 (2026-08-28). AT RATIFICATION (served, build a57d4f012ca1, 2026-09-02): MD_XAUUSD 4328.58 (2026-09-01), DFII10 2.44 (2026-08-31) — already 2.7% below the seed and inside the PARTIAL band, and the marks were kept anyway.
| Owner analyst session 2026-08-31 (framing) + this repository's calibration (data/panel/GOLD_FALSIFIER_2026-08-31.md); owner-ratified 2026-09-02 | 2027-03-01 |
| open | The 10-year is close to its peak: 'I don't think they need to go much higher than they are right now… they could probably go to 550 or 575' — and he is buying above 5.25. id: bianco_10y_near_peak. disclosed-interest (he says he bought within the week). THE WINDOW IS OURS, NOT HIS — he gave a ceiling, not a date (RULINGS section 30 class; accepted by the owner 2026-10-05). Six months spans the 10-28 and 12-09 FOMCs, the 11-04 refunding and the midterm result. LEAN: counter to the fiscal-dominance side — the first bond-bull call from the seat that carried the bear case since 2020. OBSERVABLE: DGS10 daily close. CONFIRMS: no DGS10 close above 5.75 through 2027-03-31. FALSIFIED IF: any close above 5.75. PARTIAL is not offered — 5.75 is the top of his own stated range. Report the highest close beside the verdict (a peak at 5.30 and a peak at 5.74 both confirm, and are different results). NOT TESTED HERE: his 'yields slow down in the next week or two' (a wish, not a call); his plan to add toward six (a plan cannot fail on the yield). ERROR ACCEPTED: type-1 — a recession or a Treasury/Fed lever that caps yields scores for him though his reason is value at strong growth. CONTEXT AT SEED (served, build c8f9c7419052): DGS10 5.24 (2026-10-01), 5.28 (2026-10-02); highest close to date 5.29 (2026-09-30); live cash mark 5.347 (2026-10-05 18:24Z, ^TNX — not the H.15 series). CROSS-REFERENCES: gromen_bond_rupture_q4 (DGS10 at or above 5.52 by 2026-12-04 — can confirm WITHOUT falsifying this entry: his range reaches 5.75); bianco_hike_caps_long_end (resolved FALSIFIED 09-16 — section 28: this is a new call at a new level, it does not offset that miss); rosenberg_nov_refunding; rosenberg_gridlock_yields_down.
| Bianco (Bianco Research) — David Lin, 2026-09-28 ('5.2… may be close to a peak' — 'Yes, I do'); Bloomberg TV interview ~2026-10-01, re-posted 4kdVLLaNkEI (primary; checks in data/panel/checks/2026-10-05-bianco-master-investor-bond-pivot.md §6) | 2027-03-31 |
| open | The 10-year has 'another like 80-90 basis points to go just to get back to fair value' — around 6.0–6.15% — and 'we wouldn't touch bonds with a 10-ft pole until we at least get back to fair value.' id: dale_10y_fair_value_6. THE WINDOW IS OURS, NOT HIS — he gave a level and a direction with no date (RULINGS section 30 class, as accepted for bianco_10y_near_peak on 2026-10-05). 2027-03-31 is chosen to MATCH bianco_10y_near_peak so the two seats resolve on one observable over one window. THE LEVEL IS PINNED TO WHAT HE SAID ON 09-30, NOT TO THE MODEL: his fair value has risen 5.77 -> ~6.0-6.15 in five weeks while spot rose 65bp (two of his five models take market yields as inputs). A later, higher fair value does not move this bar. LEAN: fiscal dominance / debasement — the yields-go-higher side; the mirror of Bianco's peak call and of Alden's "where they are now or a little higher." MECHANISM (his): the geopolitically driven supply/demand imbalance in the Treasury market (Framework 9) forces the 10y to his five-model fair value, and "there's plenty of time series history of the value being above fair value." OBSERVABLE: DGS10 daily close (cash, never ZN_F — section 21). CONFIRMS: DGS10 closes at or above 6.00 on any session through 2027-03-31. FALSIFIED IF: no DGS10 close at or above 6.00 through 2027-03-31. HIS OWN ESCAPE HATCH, RECORDED SO SCORING STAYS FAIR: "unless they start doing yield curve control well before that" (Session 08) and the "Accord 2.0" he now expects sooner. If dale_fedsury_lameduck RESOLVES (an announced target or formal accord) before DGS10 reaches 6.00, this entry is marked NOT ACTIVATED — resolved by intervention — never falsified; the intervention is scored on its own entries. ERROR ACCEPTED: type-1 — a 6.00 print inside the window that the 10y then gives back scores for him on the level while the stance ("don't touch bonds") loses; a stance is not what is seeded. CONTEXT AT SEED (served, build 24b6f00ce734): DGS10 5.28 (2026-10-02), 5.29 on 09-30 — the highest closes in the served window; DGS30 5.63; DGS2 4.83; DFF 3.88; ACM TP 1.020 (09-25, 63.1st pct of 36.7y); rate_path 12-09 p(hike) 0.86. 72bp to the bar. CROSS-REFERENCES: bianco_10y_near_peak (same observable, same window, opposite sign — the good case, two seats scored against each other); gromen_bond_rupture_q4 (6 inside a rupture by 12-04 would confirm both — do NOT double-count the panel); alden_no_failed_auction; read_20260820_long_end_repression and dale_fedsury_lameduck (the escape hatch). CHECKPOINTS: 10-28 FOMC; 11-04 refunding/QRA; 12-09 FOMC; the 10Y auctions.
| Dale (42 Macro) — Bitcoin Magazine interview, recorded ~2026-09-30/10-01, uploaded 2026-10-01 (primary, 9KYH0yY6Zcw; fair value 6.04% stated on Fox Business ~09-25, Hdf2jTncA4E; checks in data/panel/checks/2026-10-05-dale-fed-treasury-accord-2.md) | 2027-03-31 |
| open | The S&P 500 reaches 8,400 by mid-2027 — the year-end-2026 target 'moved to mid-2027', on 2027 EPS $425; 'it's going to take a little longer' id: yardeni_2027_spx_8400_mid. LEAN: benign / consensus — the same side as yardeni_2026_spx_8400, a new window on the same number, NOT a second side of one event. MECHANISM: FEMO — earnings momentum on a forward P/E that has already compressed; the delay is the multiple (oil, two more hikes, the AI story "pushed out"), not E. OBSERVABLE: SP500 daily_close. CONFIRMS: any SP500 close >= 8,400.00 on or before 2027-06-30. FALSIFIED IF: no such close. ALSO RECORD, never score: a further written move of the 8,400 date beyond mid-2027 (a restatement inside the window does not amend this entry — section 28). ERROR ACCEPTED: type-1 — a correction he calls a buying opportunity does not void it. DOUBLE-COUNT RULE: if yardeni_2026_spx_8400 confirms (a close >= 8,400 by 2026-12-31) this entry confirms on the same close; score the pair as ONE hit on the number and one reversal on the timing, never two hits. CONTEXT AT SEED (served, build 24b6f00ce734, 2026-10-02): SP500 7,722.72, +8.8% to the bar; closing high 7,798.99 (2026-08-13); DGS10 5.28; BZ_F 102.25; rate_path 12-09 hike p 0.86. CROSS-REFERENCES: yardeni_2026_spx_8400 (same number, 12-31 window); HV-005c (new ATH by 12-31); dale_bubble_peak_2027h2 (the opposite timing on the same tape); HV-005b (-10%). CHECKPOINTS: 2026-12-31 (the first entry resolves), Q4-26 and Q1-27 earnings seasons.
| Yardeni (Yardeni Research) — QuickTakes 'Proceed With Caution' 2026-09-15 (written); restated Bloomberg TV 09-17, Thoughtful Money ~09-17, Excess Returns recorded 2026-10-02 (primary, eEF1YMHMcZs; check data/panel/checks/2026-10-05-yardeni-excess-returns-normal-for-longer.md) | 2027-06-30 |
| open | I've called it the gradual print … generally flat to up, but it's not a gigantic print … the next year I don't see a big print. id: alden_no_big_print_12m. disclosed-interest (section 28 conduct trigger). THE 1% MARGIN ON LEG 2 IS OURS — she said "flat to up"; a $5bn dip is noise, a sustained reduction is Warsh running QT, which is her own stated falsifier 4 of the gradual print. LEAN: benign within the debasement school — the counter to every big-print call on the panel. MECHANISM (hers): the things breaking "are not really the types that require a really big print"; responses are "numerous and smaller"; standing facilities absorb stress first. OBSERVABLE: WALCL weekly (H.4.1). CONFIRMS: no print above 8,743,031 AND the window-end print not more than 1% under 6,743,031. FALSIFIED IF: either leg fails. FALSIFIED IF: any weekly WALCL print above 8,743,031 ($mn; 6,743,031 at seed + her $2tn big-print threshold) through 2027-09-30. The flat-to-up leg is scored by hand (note). Leg 1 is a near-certain confirm and is not the reason to seed; LEG 2 IS — WALCL has been flat since June (+11.1bn 06-17 -> 10-07), so "up" is not already true. ERROR ACCEPTED: type-2 on leg 1 (a $1.9tn crisis expansion confirms her letter while breaking her spirit — report the peak beside the verdict). CONTEXT AT SEED (served, build e98331bb906b): WALCL 6,743,031 (2026-09-30), 6,747,560 (2026-10-07); TREAST 4,564,161 / 4,566,384; +211.8bn from the 2025-12-03 trough. CROSS-REFERENCES: benign_gdp_positive_no_qe (balance-sheet re-expansion > 5% SAAR, window 2026-12-31 — overlaps, does not duplicate); alden_no_failed_auction.
| Alden (Lyn Alden Investment Strategy) — Market Disruptors (Mark Moss), published 2026-09-30 (primary, fDo2Yy_0k_0; recorded ~late September, inferred); the >$2tn threshold is hers, Document 04 (lynalden.com, 2026-03-22); check data/panel/checks/2026-10-10-alden-hindsight-crisis.md §h | 2027-09-30 |
| open | Fundamental disinflationary environment between now and end-2027: demand running below potential (four-quarter real GDP heading to ~1.4% vs 2.0% potential), trimmed-mean PCE decelerating, unit labor costs ~0.5% — so the Fed should NOT hike. id: rosenberg_2027_disinflation. LEAN: benign / cyclical counter — the seat that makes the debasement case earn it. FALSIFIED IF: core PCE (PCEPILFE) 3m SAAR prints above 3.0% on two consecutive monthly vintages inside the window — the forward-looking measure he himself argues from, not the YoY. ALSO WEAKENED IF the Fed hikes AND the labor market does not crack within two quarters (his cost-of-error argument fails). CONFIRMS: core PCE 3m SAAR <= 2.5% sustained with YoY converging toward it through 2027. SEEDED AT (served, build 8b79e3cf21ca): PCEPILFE 3m SAAR 2.89 / 6m 3.76 / YoY 3.29, impulse decelerating (as_of 2026-06-01); CPI core 1.64 / 2.42 / 2.47 (as_of 2026-07-01). SCORING SCOPE, CORRECTED 2026-09-07 — this note had gone STALE IN THE GENEROUS DIRECTION, which is the dangerous one: it said four cited inputs were unserved and that the entry 'scores only the observable half', and an audit of every open entry that day found ALL FOUR are now available. GDPPOT and PCETRIM12M159SFRBDAL had been collected since the seed and nobody re-checked; ULCNFB and DSPIC96 were genuinely missing and were collected 2026-09-07. The entry therefore scores its WHOLE claim from here, not half of it. The lesson is the standing one: re-check whether an old 'unscoreable' is still true before resting a verdict on it. The FALSIFIED IF / CONFIRMS rules above are UNCHANGED — they were always core-PCE based, so widening the evidence does not renegotiate the test. SERVED 2026-09-08 (build 2c08f711f271, run-2026-09-08T00:30:03Z), and it raises a question rather than settling one: ULCNFB now serves yoy 1.40 / 6m 1.24 / 3m 1.17, against the 'unit labor costs ~0.5%' the claim above attributes to him. FLAGGED, NOT SCORED on 2026-09-07 pending his basis. RESOLVED 2026-09-08, and the note above it was WRONG on one point, corrected here beside it: he DID name the measure — YoY, recorded in his profile since 2026-08-10. On the FRED vintage available the day he spoke (~2026-08-05) Q1-2026 ULC read +0.49% YoY; the Q2 print that took it to +1.43% was released 2026-08-06, the day after, and 52% of that jump is the soft Q2-2025 base. THE LEG PASSES ON ITS VINTAGE and is stale as a current description — both true, neither touches this entry's core-PCE rule. data/panel/checks/2026-09-08-rosenberg-ulc-vintage.md. Cross-reference: the benign-counter line 'Core PCE <= 2.5% YoY two consecutive months in 2026' (2026-07-09) — related, different measure and window; do not double-count. ERROR ACCEPTED: type-2 on the hike half — a late-2027 re-acceleration after a 2026 hike scores against him only if the two-print rule trips. | Rosenberg (Rosenberg Research) — Excess Returns, ~2026-08-05 (primary) | 2027-12-31 |
| open | Demand is running below potential — a negative and widening output gap. The load-bearing premise under his entire cyclical-counter seat, restated in every session since 2026-08-05. id: rosenberg_output_gap_negative. disclosed-interest (§28: first promotional session). LEAN: benign / cyclical counter — the seat that makes the debasement case earn it. WHY THIS ENTRY EXISTS: this premise was UNSCOREABLE for the entire life of his profile — the 2026-08-10 check listed GDPPOT as the top build gap and the 2026-08-21 update still had it open. It became scoreable this week, and the first measurement disagrees with him. It is the cleanest single test of the cyclical counter-thesis the panel has. OBSERVABLE: (GDPC1 / GDPPOT - 1), both served, quarterly. CONFIRMS: the gap turns negative and stays negative for two consecutive quarters inside the window. FALSIFIED IF: the gap is still positive at Q4-2027. SEEDED AT (served, build 9f9176311288, as_of 2026-04-01): the gap is +0.83% — POSITIVE — having narrowed every quarter from +1.46% at Q3-2025: +1.46, +1.03, +0.98, +0.83. *** SCORE THE TWO AXES SEPARATELY AND NEVER COLLAPSE THEM. *** His sentence is about a LEVEL, so the level answers it and the level currently says no. The IMPULSE is his: slack is building on every print. The repository has already made the mirror error once (check file 2026-08-10 §1 — citing a level-based label to rebut a rate-of-change claim), and this entry is the place the symmetric error would be made. ERROR ACCEPTED: type-2 — CBO revises GDPPOT, and a revision can close the gap without the economy doing anything. Record the vintage at every reading.
| Rosenberg (Rosenberg Research) — Framework 1, restated Thoughtful Money 2026-09-06 (primary; checks in data/panel/checks/2026-09-07-rosenberg-thoughtful-money-session03.md) | 2027-12-31 |
| open | Credit turns roughly one year ahead of the AI-driven equity peak. id: rosenberg_credit_leads_equity. LEAN: caution. THE BEST-SPECIFIED OF THE SEVEN, and the reason is worth copying: it names UNTESTED explicitly and refuses to let a non-firing mechanism count as confirmation, which is the property most seeded entries lack. It also resolves the standing open item in his profile. MECHANISM: credit markets price the funding stress of an investment boom before the equity market prices its earnings disappointment. OBSERVABLE: t=0 is the FIRST session on which credit_transmission.hy_band_break == true. The equity peak is the SP500 all-time close that is subsequently not exceeded for 180 calendar days. CONFIRMS: the equity peak follows t=0 by 270-450 days. WEAKENED: lag of 450-550 days — direction his, calibration not. FALSIFIED IF: the equity peak PRECEDES t=0, or the lag exceeds 550 days. UNTESTED — the state that matters most here: if hy_band_break never fires inside the window, this entry resolves NOT ACTIVATED. A mechanism that did not fire is not a mechanism that was right, and the absence of a credit break must never be reported as his call working. ERROR ACCEPTED: type-2 — the 180-day confirmation lag means the equity peak is only identifiable half a year after the fact, so this entry will resolve late by construction. Accepted deliberately: a peak called in real time is not a peak, it is a drawdown. CONTEXT AT SEED (served, build 78f16b8c334e, 2026-09-02T21:04Z): hy_band_break FALSE, leak_state contained; HY OAS 2.63 (2026-08-31) at the 2.1st pct and IG OAS 0.80 at the 28.4th — BOTH short_window, n=825 over 3.1y, so the percentile is cited with its span and is NOT a historic low; delta_bp_20s -15 on HY, i.e. still TIGHTENING. bizd is the one elevated upstream input (13.35, 16.8% off its 1y high). SP500 7,631.47 (2026-09-01), 99.2nd pct of 10.1y, peak on record 7,798.99 (2026-08-13). CROSS-REFERENCES: hunt_2026_reversal, dale_20260731_tradeable_low.
| Rosenberg (Rosenberg Research) — David Lin session, 2026-09-01 (owner-supplied draft, processed 2026-09-02) | 2027-12-31 |
| open | They're not going to have failed bond auctions this year. They're not going to have failed bond auctions next year. id: alden_no_failed_auction. disclosed-interest (section 28 conduct trigger). THE DEFINITION OF "FAILED" IS OURS, NOT HERS — she gave none. A Treasury auction fails technically when accepted bids do not cover the offering. LEAN: benign counter inside the debasement school — "a leaky bucket", not a break; the other side of gromen_bond_rupture_q4. MECHANISM (hers): above ~100% debt/GDP the outcome is emerging-market characteristics (persistent above-target inflation, a constrained central bank), not a funding stop; "someone has to buy those bonds", and the fourth buyer is the central bank or commercial banks. OBSERVABLE: AUCT_2Y_BTC / 3Y / 5Y / 7Y / 10Y / 20Y / 30Y (bid-to-cover, every coupon auction through 2027-12-31); auction_internals.absorption beside it. CONFIRMS: no coupon auction prints a bid-to-cover below 1.00 through 2027-12-31. FALSIFIED IF: any does. ALSO RECORD, never score: a cancelled or postponed coupon auction, or a Fed purchase programme announced as an auction backstop — that is her own "fourth buyer" sequel, and it would mean the failure was prevented rather than absent. ERROR ACCEPTED: type-2 — the bar is the literal one, so a disorderly auction that still covers (a large tail, a record dealer takedown) scores for her. That is the cost of not sharpening a claim she did not sharpen; the absorption read carries the quality question. CONTEXT AT SEED (served, build c8f9c7419052): lowest bid-to-cover 2.21 (5Y, 2026-09-23); 7Y 2.42 (09-24); 10Y 2.71 (09-09); 30Y 2.61 (09-10); absorption "weak flow / capacity_exhausted stock" (20Y). DGS10 5.28, DGS30 5.63 (2026-10-02). CROSS-REFERENCES: gromen_bond_rupture_q4 (a rupture is a yield event, this is an auction event — both can be true or false independently); read_20260820_long_end_repression (a backstop inside the window is its evidence).
| Alden (Lyn Alden Investment Strategy) — David Lin Report, recorded 2026-10-02, published 2026-10-04 (primary, NQYo5_KMm6g; checks in data/panel/checks/2026-10-05-alden-david-lin-flat-curve.md) | 2027-12-31 |
| open | I think we've most likely seen the bottom. I think the sentiment bottom and probably the price bottom are in … call it 3 years from now this will still be viewed as an attractive time to have bought. id: alden_btc_bottom_in. disclosed-interest (section 28 conduct trigger — Orange Juice's Bitcoin treasury; partner, Ego Death Capital). THE WINDOW IS OURS, NOT HERS — she gave none for the bottom; her three-year line is recorded below and NOT scored here (RULINGS section 30 class). "Bottom" is read as the cycle's price low: a lower low falsifies. LEAN: debasement / Bitcoin-bull side. MECHANISM (hers): fast money and leverage are out ("went to AI"), sellers are exhausted, so "it only takes minor good news or just less bad news"; the sequence is chart traders, then momentum traders, then retail. OBSERVABLE: MD_BTCUSD daily close (Twelve Data spot; trades seven days). CONFIRMS: no MD_BTCUSD daily close below 58,624.71 (the 2026-06-30 low) through 2027-12-31. FALSIFIED IF: any daily close below it. ALSO RECORD, never score: a close below 50,000 — her Session 08 "would surprise me" level — is a second, harsher failure inside the same event. ERROR ACCEPTED: type-1 — a liquidity-driven rally confirms it whatever the reason. CONTEXT AT SEED (served, build e98331bb906b): MD_BTCUSD 82,635.55 (2026-10-09) — a 29.1% fall to the falsifier; peak since the low 86,620 (09-21); all-time high 124,658.54 (2025-10-06). CROSS-REFERENCES: none on MD_BTCUSD. The Session 07 Bitcoin/gold-ratio seed (two-to-three-year view) remains PROPOSED, unseeded — do not seed both off one bullish stance without the owner choosing. Counter-claim: no benign/bear Bitcoin entry exists; this entry's falsifier is the bear outcome and is reported as such — the owner may want a named bear-side seat beside it.
| Alden (Lyn Alden Investment Strategy) — VRIC Media, published 2026-10-10 (primary, M8HnA245plQ; recorded on or before, inferred >= mid-September); restates Kitco 2026-09-28 ('exhaustion of sellers … I think we've crossed now') and Session 08 ('my guess is the bottom's in'); check data/panel/checks/2026-10-10-alden-hindsight-crisis.md §h | 2027-12-31 |
| open | Prime (Fannie/Freddie) delinquency, which began at the early stage in summer 2026, takes 'about 18 months before it is at a very material level.' id: wright_2028_gse_sdq. THE THRESHOLD AND WINDOW ARE OURS, NOT HERS — "a very material level" has no number in her words; the window is her "about 18 months" from summer 2026, read to the 2028-04 monthly summary (published about a month later, hence the 2028-05-31 window). LEAN: housing distress / forced supply — her two-wave mechanism (Framework 6): FHA, non-prime and investors first; the prime, fixed-rate GSE borrower second, on costs (taxes, insurance, HOA, energy) rather than job losses. OBSERVABLE: FNMA_SF_SDQ_RATE and FHLMC_SF_SDQ_RATE (serious delinquency, 90+ days or in foreclosure, % of each GSE's single-family book; monthly). A LAGGING PROXY: her object is early-stage (30-59 day) delinquency, which is not served. CONFIRMS: either series prints >= 1.00% in any month through the 2028-04 summary. FALSIFIED IF: both stay below 1.00% on every print through 2028-04. GUARD: a print driven by a declared disaster forbearance (the 2017 hurricane class) is recorded but confirms only if the rate is still >= 1.00% three months later. ERROR ACCEPTED: false negative — her wave could be real but slower, or sit in early-stage buckets that never roll to 90+ inside the window. BASE RATE: against. In the served history (from 2009-11 / 2012-12, so no crisis onset in sample), outside the 2017 hurricane and COVID, neither series has risen 40bp in 18 months. CONTEXT AT SEED (served, build 2b9e8963bbfb -> 39c28edddb17, collector run 2026-11-02T13:00Z): Fannie 0.62, Freddie 0.61 (2026-09); 2025-05 lows 0.53 / 0.55; 2019 range 0.66-0.76 / 0.61-0.70. Fannie last printed >= 1.00 in 2022-03 (1.01), as the forbearance overhang ran off. CROSS-REFERENCES: wright_2026q4_price_drag (proposed 2026-09-13, never seeded; a price observable, would not double-count); DRSFRMACBS (all-bank SF delinquency, quarterly) is context.
| Wright (M3 Insights) — Triangle Investor 2026-09-28 and Commodity Culture ~2026-09-22/23 (primary; checks in data/panel/checks/2026-11-02-wright-gse-delinquency-rescore.md) | 2028-05-31 |
| open | If the 2026-11-03 midterms end one-party control, the 10-year Treasury yield is lower on presidential election day 2028 than on midterm day 2026 — '80% of the time bond yields go down in that two-year period.' id: rosenberg_gridlock_yields_down. disclosed-interest (section 28 conduct trigger). DATE CORRECTED 2026-10-05 (owner OK): date_made moved from 2026-09-30 to 2026-09-28: he first stated it on the Maggie Lake live stream (rNqtnTz4e5Q, 2026-09-28; check data/panel/checks/2026-10-05-rosenberg-talking-markets-and-riskreversal.md). Scoring unchanged. CONDITION IS HIS: "when a one-party power… swings to a two-party system with gridlock." VOID, never scored, if Republicans hold both chambers of the 120th Congress on the certified result. LEAN: political-cycle / fiscal-drag — no new stimulus after 11-03, demand cools, yields follow. OBSERVABLE: DGS10 daily_close on 2026-11-03 and 2028-11-07 (the election-day closes he named). CONFIRMS: DGS10 on 2028-11-07 below its 2026-11-03 close. FALSIFIED IF: at or above it. No PARTIAL — he stated a direction, not a size. ERROR ACCEPTED: type-1 — a recession, a crisis or a refunding-led rally anywhere in two years scores for him whether or not gridlock is the cause; his own base rate carries the same weakness (three of four historical wins ran through a recession or crisis). CONTEXT AT SEED: his base rate measured on served DGS10 — 4 of 5 since 1994 (1994 −173bp, 2006 −85, 2010 −85, 2018 −232, 2022 +12); the only case after the 1981–2020 secular bond bull went the other way. DGS10 5.28 (2026-10-02, build c8f9c7419052) — the baseline is the 2026-11-03 close, not this. data/panel/checks/2026-10-01-rosenberg-macrovoices-552.md §3. CROSS-REFERENCES: rosenberg_recession_long_end_down (a recession trigger, 126 sessions); rosenberg_nov_refunding (a 30-day lever). DO NOT DOUBLE-COUNT — each resolves on its own clause; a rally may confirm all three, and must be reported as one rally.
| Rosenberg (Rosenberg Research) — MacroVoices #552 'Navigating The Noise', recorded ~2026-09-30, published 2026-10-01 (primary, DsUSyYrfUYM; checks in data/panel/checks/2026-10-01-rosenberg-macrovoices-552.md) | 2028-11-07 |
| open | Fed transition from funds-rate to repo-rate targeting | Dale, Jan-2026 deck | 2028-12-31 |
| open | The risk-asset bubble peaks in the second half of 2027 or the first half of 2028. id: dale_bubble_peak_2027h2. LEAN: caution on timing, bullish until then — the dated end of a thesis that is otherwise open-ended, which is why it is worth having. OBSERVABLE, AND IT IS DELIBERATELY BORROWED: the peak is the SP500 all-time close that is subsequently not exceeded for 180 calendar days — THE SAME DEFINITION as rosenberg_credit_leads_equity. That is the point of seeding it this way: two panel seats now resolve on one observable with one definition, so their timing models can be compared directly instead of argued. This is the first such pair in the file. CONFIRMS: the peak so defined falls between 2027-07-01 and 2028-06-30. FALSIFIED IF: the peak precedes 2027-07-01, or SP500 is still setting new all-time closes after 2028-06-30 (a bubble that has not peaked by then missed his window, whatever happens later). RESOLUTION LAG IS STRUCTURAL: the 180-day confirmation means a peak inside the window is only identifiable ~6 months after the fact, hence the 2028-12-31 window against a 2028-06-30 claim. Accepted deliberately — a peak called in real time is not a peak. ERROR ACCEPTED: type-2 — the 180-day rule will not distinguish a true top from a long, deep consolidation that later resolves higher. Same trade-off the Rosenberg entry accepts, and taking it identically is what makes the two comparable. CONTEXT AT SEED (served, build 940a8c83bc33): SP500 7,666.60 (2026-09-02), 99.3rd pct of 10.2y and a fresh all-time close — so the peak has NOT yet been set and the clock has not started. Percentile span cited because 10.2y is the collected window, not the index's history. CROSS-REFERENCES: rosenberg_credit_leads_equity (same peak definition, different mechanism and different timing model — do NOT double-count; they can both be right, or one right on the date and the other on the cause); dale_20260731_tradeable_low (a near-term drawdown call that is NOT inconsistent with a 2027-28 peak).
| Dale (42 Macro) — Macro Minute excerpt, YouTube cC2abmDDIqQ, uploaded 2026-09-03 (primary but a PARTIAL EXCERPT); stated as the current base case | 2028-12-31 |
| open | Durable YCC | Dale, Jan-2026 deck | 2029-12-31 |
| open | >50% odds of US fiscal crisis or total war within 5–7 years Probabilistic claim — scoring rule is qualitative. | Dale, Jan-2026 deck | 2031-12-31 |
| standing | Yen 200, then 300 | Johnson, May-2026 | long (unspecified) |
| standing | Gold/oil ratio >100 over the cycle | Gromen, structural | cycle |
| standing | MOVE >130 always answered ('Alamo') Carry with the RMP-era caveat from gauge 7: standing purchases make the discrete 130 trigger the tail case. | Gromen, structural | standing |
| standing | Oil >= 85: 'once it hits 85, it starts to dysfunction' — the operative Alamo substitute for the oil leg INSTRUMENT (ruled 2026-08-14): BZ_F daily close >= 85 fires; a SINGLE close suffices (threshold observation, not a regime flip). DCOILBRENTEU is fallback ONLY if BZ_F is missing 3+ sessions; WTI (CL_F/DCOILWTICO) NEVER satisfies a Brent-benchmark claim. STATE AT BINDING: FIRED — BZ_F 88.98 (2026-08-12), >= 85 for three consecutive closes since re-crossing at 87.72 on 2026-08-10; under this instrument it had also fired 2026-07-29..07-31 (90.74 peak) before dropping to 79.36 (08-04). The oscillation around the line is itself a reading: 'dysfunction begins at 85' is being tested repeatedly, and what 'fired' obligates is watching the DYSFUNCTION observables (auction tails, MOVE, sofr_iorb), not calling a price top. Related gap on record: FRED Brent spot printed $4.35 above BZ_F on the shared 08-11 date — the systematic spot-vs-future divergence that forced this ruling. | Gromen, structural (twice-stated across two sessions) | standing |
| standing | Multi-year 'funk,' not crash | Bianco, Dec-2025 | multi-year |
| standing | S2 intervention timeline: verbal in days; Treasury-side ≤3 weeks or at QRA; balance-sheet action only on confirmed dysfunction | Owner, seed | conditional (activates on S2 trigger) |
| standing | If the Fed eases (or is perceived too easy relative to fiscal/inflation reality), the long end sells off and term premium rises — tightening financial conditions while intending to ease. The Fed risks losing control of the long end -> term-premium-led steepening ('can the Fed out-dove the market?'). LEAN: fiscal / vigilante mechanism (reinforces the fiscal-dominance lean, but as a MECHANISM with a falsifier). FALSIFIER: term premium stays mid-range and the long end does NOT lead even if the Fed eases; or the curve FLATTENS on easing rather than steepening. CURRENT READ: dormant trigger — Warsh is hawkish, so the 'Fed eases' precondition isn't active; this is the S2 vigilante mechanism (line 104) with a named owner, fires if the 7/29 hike-risk resolves dovish. Observable: THREEFYTP10 (ACM term premium, ~52nd pctile), DGS30, DGS20, T10Y2Y. | Constan (Damped Spring) | conditional (activates on a Fed-easing signal) |
| standing | Foreign/EM central banks keep rotating reserves from US Treasuries toward gold and hard assets ('outside money'); gold's monetary role rises and the dollar's reserve role erodes over the cycle, with a structural inflationary bias in the West (Bretton Woods III). STATUS 2026-08-21: DORMANT — roster vetting (data/panel/ROSTER_VETTING_2026-08-21.md) found no public 2026 appearance or publication (press page ends 2025-03); the entry stays standing but is not refreshed and its source cannot currently be re-dated. LEAN: debasement / structural (overlaps the existing panel's debasement axis) — seed as a STANDING entry, not a live trade signal; low current output, mixed multi-year scoreboard, score qualitatively like the Gromen/Johnson standing entries. FALSIFIER/TENSION: dollar reserve role and DTWEXBGS strengthen durably -> the dollar-weakness leg fails. CURRENT READ: split — gold leg RIGHT (ripping, ~95th pctile, CB gold buying real); dollar-weakness leg WRONG so far (DTWEXBGS ~86th pctile, firm). Directly contests Johnson's milkshake on the dollar axis — that disagreement is the useful part, mapped to DTWEXBGS which resolves it. SCOPE GAP: the cleanest BW3 observables (official-sector gold buying, TIC foreign Treasury holdings) are NOT collected — track on gold + broad-dollar for now; flag TIC / CB-gold as panel_gap candidates. Observable: MD_XAUUSD, DTWEXBGS. | Pozsar (Ex Uno Plures / BW3) | structural |
| standing | The canonical debt-deflation voices have vacated the thesis: Hoisington's Q2-2026 letter (2026-07-16) migrates its equilibrium-inflation range to 3.5-4.5% with 'significant risk of episodes above 5%' and cuts duration from ~21y to under 1y; Albert Edwards (Odd Lots 2026-05-15) declares the Ice Age over. Logged as a CROWDING signal about the debasement consensus, not as vindication of it. id: hunt_2026_reversal. This entry is about the PANEL, not the market: when the best-credentialed opponents capitulate, the counter-thesis seat is empty and the consensus is crowded (BIAS_REVIEW §2). Scoring is qualitative. It 'pays off' for the debasement case only if the variable Hunt reversed ON — money velocity rising with inflation above 3% — actually persists; it scores AGAINST the consensus if disinflation resumes after the two loudest bears left (the late-capitulation shape). Observable today: the core PCE ladder (served). M2 velocity (M2V) was named here as NOT collected and as the cleanest scoreboard between Hunt's new view and the standing disinflation seat (Shvets, provisional). CORRECTED 2026-09-07: M2V IS collected and has been for some time — this note was stale. The scoreboard it asks for is available now, so the entry no longer waits on a build. Keep ingesting Hoisington's free quarterly letters as a bond-bear data point. Cross-reference: rosenberg_2027_disinflation (opposite sign on the same ladder). | Hunt (Hoisington) Q2-2026 letter 2026-07-16; Edwards (SocGen) Odd Lots 2026-05-15 — recorded 2026-08-21 | standing |
| standing | The yield backup is a Fed-expectation reset under a new chair, not a fiscal or term-premium story: breakevens are flat and the move is in the real rate. id: rosenberg_yield_attribution. LEAN: counter to fiscal dominance — this is the entry that costs the debasement case points if the long end is being repriced by the Fed path rather than by the debt arithmetic. THE FALSIFIER IS OURS, NOT HIS. He did not state one; the repository supplies it, and that is recorded so nobody later credits him with a test he never accepted. OBSERVABLE: decomposition.term_premium (ACM split) and the decomposition windows. NEVER the 2s10s spread alone — a spread is not an attribution, and §21's instrument discipline applies. CONFIRMS: the decomposition windows stay real-led with tp_share_of_nominal low, and the 2s10s does not steepen materially while breakevens stay pinned. WEAKENED IF: 2s10s steepens 25bp or more from +41bp while T10YIE stays below 2.40 AND decomposition.term_premium rises through its 70th percentile. That combination is the fiscal signature, and it is the one reading his attribution cannot absorb. DOES NOT SETTLE: the term-premium question at the long end. Over 250 sessions the 10y rose 52bp against breakevens -10bp — a ~62bp real move — which is exactly where a term-premium story lives. His claim is about the MARGINAL move being front-end led; it does not hand him the exclusive reading, and the check file says so. ERROR ACCEPTED: type-2 — a standing window with no resolution date can drift indefinitely. Reviewed at each FOMC; if the fiscal signature has not appeared by 2027-06-30, close it as unresolved rather than letting silence accrue to him. CONTEXT AT SEED (served, build 78f16b8c334e, 2026-09-02T21:04Z): ACM term premium 0.8751 (2026-08-28), 56.8th pct of 36.7y — MID-RANGE, so neither school is currently vindicated; T10YIE 2.35 (2026-09-01), 73.3rd pct of 23.7y, which is 5bp under the 2.40 pin; 2s10s +41bp; ytd decomposition real-led (+56bp nominal on +50bp real, +6bp breakeven); since_trough_252s real-led (+78bp on +76bp real). CHECKPOINTS: FOMC 2026-09-16, 2026-10-28, 2026-12-16. CROSS-REFERENCES: read_20260818_real_led_supply, read_20260820_long_end_repression.
| Rosenberg (Rosenberg Research) — David Lin session, 2026-09-01 (owner-supplied draft, processed 2026-09-02) | standing |
| standing | In the next US recession, long-end Treasury yields rise rather than fall — 'maybe they'll go down for like one little second, but they're going to scream higher.' id: gromen_recession_long_end_up. THE POST-ACTIVATION HORIZON IS OURS, NOT HIS — he gave none; an activation-dated conditional accepted under RULINGS section 30 by owner ratification 2026-09-24. Accepting it costs a type-2 error if his 'one little second' dip outlasts 126 sessions. LEAN: fiscal dominance / debt gamma — the counter-consensus half of his school; the textbook recession trade is the opposite. MECHANISM (his): a stronger dollar forces foreigners short $13-14tn of dollar debt to sell Treasuries; banks, pensions and levered basis funds de-gross Treasuries in a slowdown; a recession adds 600-1,000bp of GDP to the deficit. OBSERVABLE: DGS30 (primary), DGS10 (reported beside it); SAHMREALTIME for activation, first vintage at or above 0.50. CONFIRMS: the DGS30 close on the 126th session after the activating print's release is ABOVE its close on the release day. FALSIFIED IF: that close is 25bp or more BELOW. PARTIAL: between. NO ESCAPE HATCH for intervention: a long-end purchase programme inside the window is his own predicted sequel, but the entry scores the yield, not the response. ERROR ACCEPTED: type-1 — a recession yield rise driven by an oil/inflation shock rather than his NIIP/de-grossing mechanism scores as a win; that mechanism is unserved (no TIC). CONTEXT AT SEED (served, build 95444956b014): SAHMREALTIME -0.07 (2026-08); DGS10 5.11, DGS30 5.40 (2026-09-23); DFF 3.88 (2026-09-22). CROSS-REFERENCES: gromen_bond_rupture_q4 (unconditional, dated; do not double-count a rise that lands inside both windows). COUNTER: the FALSIFIED leg is the textbook flight-to-quality.
| Gromen (FFTT) — Macroscopic Podcast (GoldRepublic), recorded ~2026-09-21, published 2026-09-23 (primary; checks in data/panel/checks/2026-09-24-gromen-goldrepublic-macroscopic.md) | conditional (activates on the first SAHMREALTIME print >= 0.50) |
| standing | In the next US recession, Treasury yields fall — risk appetite and private credit demand drop, banks rebuild liquidity in Treasuries and MBS; true even in the 1970s. id: rosenberg_recession_long_end_down. disclosed-interest (section 28 conduct trigger). THE POST-ACTIVATION HORIZON IS OURS, NOT HIS — he said only that the trade "could last a year"; mirrored on gromen_recession_long_end_up so both seats resolve on one clock (RULINGS section 30). LEAN: cyclical / textbook flight to quality — the other side of Gromen and of Gundlach. OBSERVABLE: DGS30 (primary, the same leg Gromen's entry scores), DGS10 beside it; SAHMREALTIME for activation, first vintage at or above 0.50. CONFIRMS: the DGS30 close on the 126th session after the activating print's release is 25bp or more BELOW its close on the release day. FALSIFIED IF: that close is ABOVE. PARTIAL: between. NO ESCAPE HATCH for intervention: a Fed or Treasury long-end programme that pulls yields down scores as a win, and he has said the 2023 refunding template is the lever — record the lever beside the verdict. ERROR ACCEPTED: type-1 — a policy-driven rally scores for him even if his bank/risk-appetite mechanism is not the cause. CONTEXT AT SEED (served, build c8f9c7419052): SAHMREALTIME 0.00 (2026-09) — NOT ACTIVATED; DGS10 5.28, DGS30 5.63 (2026-10-02). CROSS-REFERENCES: gromen_recession_long_end_up (mirror — scored against each other, never both won); rosenberg_nov_refunding (a refunding-led rally inside the window is its evidence, not this entry's, unless the recession has activated it); rosenberg_gridlock_yields_down.
| Rosenberg (Rosenberg Research) — Webcast with Dave featuring Jeffrey Gundlach, recorded ~2026-09-29, published 2026-09-30 (primary; checks in data/panel/checks/2026-09-30-rosenberg-gundlach-webcast-sep29.md) | conditional (activates on the first SAHMREALTIME print >= 0.50) |
| standing | Fade the bond rally: unless Treasury or the Fed steps up support (larger long-end buybacks, a shift of issuance to bills, or Fed coupon buying), the 10-year's fall from 5.31 reverses after the 2026-11-04 refunding. id: gromen_fade_bond_rally_post_qra. THE ACTIVATION RULE, BASE AND WINDOW ARE OURS, NOT HIS — he gave no level or date (RULINGS section 30 class). The exception list is his, verbatim: "the one way I would not fade is if the Fed was involved or if Treasury was doing buybacks or they shifted a bunch to the short end." HE MADE THE CALL WITH THE ENLARGED BUYBACKS ALREADY RUNNING ($6bn caps since 09-10) and called $4-6bn operations "tinkering" in Session 09 — so the status quo is NOT his exception; only a new step is. LEAN: fiscal dominance / supply — the same side as gromen_bond_rupture_q4 and his Session 11 "bear market rally". MECHANISM (his): a growth or AI scare lowers long yields "for a second on sentiment"; the next QRA then shows the larger borrowing need and "the long end will go". ACTIVATION (pre-written; read on the first long-end buyback operation dated after 2026-11-04 or on 2026-11-20, whichever is first, and on the November refunding auctions): ACTIVATED unless one of his exceptions is engaged. EXCEPTION ENGAGED (=> NOT ACTIVATED, never wrong) if ANY of: (i) BUYBACK_LS_10Y20Y_CAP_BN or BUYBACK_LS_20Y30Y_CAP_BN on the first post-11-04 operation is ABOVE 6.0; (ii) AUCT_10Y_OFFER_BN on the November new-issue 10Y is BELOW 42.0, or AUCT_30Y_OFFER_BN on the November 30Y is BELOW 25.0 (a shift toward bills); (iii) the FOMC announces outright purchases of coupon securities beyond reinvestment on or before 2026-11-20. Caps renewed at 6.0 (or cut) with coupon sizes unchanged (or raised) = ACTIVATED. OBSERVABLE: DGS10 daily close (cash series — section 21). BASE = the DGS10 close on 2026-11-04. CONFIRMS: once activated, any DGS10 close at or above BASE + 25bp, 2026-11-05..2026-12-31. FALSIFIED IF: activated, no such close, and the 2026-12-31 close is at or below BASE. PARTIAL: activated, no +25bp close, and the 2026-12-31 close above BASE. BASE RATE (measured 2026-10-30): an unconditional +25bp DGS10 touch within 39 sessions occurred from 37.2% of start dates (7,961 starts, 1994-11..2026-10). A CONFIRM is evidence only to the extent it beats that; report it beside the verdict. NO ESCAPE HATCH after activation: a Treasury or Fed response inside the window is his predicted sequel; the entry scores the yield. ERROR ACCEPTED: type-1 — a rise driven by the Fed's December hike or by oil, not by supply, scores for him. THREEFYTP10 is reported beside the verdict, not scored. CONTEXT AT SEED (served, build 85f1e5ba3d3d, collector 2026-10-30T00:39:45Z): DGS10 4.98 (10-29), 5.31 peak (10-05); THREEFYTP10 0.94 (10-23) from 1.10 (10-05); long-end buyback caps 6.0 (last operations 10-15, 10-27); refunding-month offers 10Y 42.0 / 30Y 25.0 (August). CROSS-REFERENCES: owner_10y_yield_topped_531 (opposite side; never report one 10y path as a win for both); gromen_bond_rupture_q4 (5.52 by 12-04 — same direction, do not double-count); gromen_recession_long_end_up (SAHMREALTIME-activated recession half); bianco_no_oct_hike_10y_550. COUNTER: owner_10y_yield_topped_531 is the standing counter; the NOT-ACTIVATED branch records a support step and scores nothing. CHECKPOINTS: 2026-11-02 financing estimates; 2026-11-04 QRA and tentative buyback schedule; the November 10Y and 30Y auctions; 2026-12-09 FOMC; 2026-12-31.
| Gromen (FFTT) — MacroVoices #556, recorded 2026-10-27, published 2026-10-29 (primary, FvBf9HBT_F8; checks in data/panel/checks/2026-10-30-gromen-macrovoices556-fade-the-bond-rally.md) | conditional (activation read at the 2026-11-04 refunding; resolves 2026-12-31) |
| resolved | SLR reduction VERIFIED via public regulatory record (not Dale's deck): OCC/Fed/FDIC final rule modifying the eSLR adopted 2025-11-25, effective 2026-04-01 (early adoption from 2026-01-01). Note: finalized before the Jan-2026 deck — decision largely already made when the call was logged. | Dale, Jan-2026 deck | 2026-03-31 |
| resolved | June-2026 CPI headline cools while core stays sticky PARTIAL — half-falsified (SPEC_PATCH v2.1.1 §P6.1). Headline cooled ✓ (4.2 -> 3.5 YoY; -0.4 MoM SA, largest drop since Apr-2020; energy -5.7 MoM). Core sticky ✗ — core 0.0 MoM, 2.9 -> 2.6 YoY; shelter +0.1 MoM (smallest since Jan-2021). The miss is the informative half — see Seed B below. | Owner, seed | 2026-07-21 |
| resolved | July CPI headline re-accelerates on energy reversal (ceasefire collapsed Jul-8; June's -9.7% gasoline unwinds) FALSIFIED on its own line (YoY <= 3.5 with crude holding >= current). July print (released 2026-08-12, in the 2026-08-13 collector run): headline YoY 3.5 -> 3.3; MoM +0.07 (332.813 / 332.568, served vintage). Crude held ABOVE the 07-14 mark the entire window - DCOILWTICO 80.44 (07-14) -> 84.77 (08-11), touching 93.08 (07-23) - so the energy reversal HAPPENED and headline disinflated anyway: the gasoline unwind was absorbed. Honest residue: MoM did swing -0.42 -> +0.07, so the monthly direction turned positive, just far too weak to move YoY; and the late-July crude spike lands in the AUGUST print on pass-through lag. An August YoY back above 3.5 would be a NEW observation, not a rescue of this entry - the window was the July print, scored as written. | Owner, seed | 2026-08-12 |
| resolved | June shelter +0.1% MoM is the first sub-0.2% print since Jan-2021 — tracking a trend-break vs a one-month blip. NO directional bet. NEUTRAL watch-item, scored on the next two prints. A directional 'blip' call was declined as framework-bias risk (the debasement lens wants shelter sticky) — logged neutral to keep the scorecard honest (Ruling 5, SETTLED). *** OBSERVABLE NOW SERVED, 2026-09-07. *** Until today this entry named an observable the platform did not collect: shelter appeared nowhere in config/, derive/ or api/ except in this note, so it would have reached its own 2026-09-11 window unscoreable. CUSR0000SAH1 is now collected (release 10, resolved live) — see the collect commit of 2026-09-07. PREMISE CORRECTION, recorded BESIDE the claim and deliberately NOT merged into it (the §28 pattern: a dated entry is not rewritten after the fact, the correction sits next to it). 'First sub-0.2% print since Jan-2021' is FALSE on the current vintage — 2021-02 (+0.179) and 2021-08 (+0.172) are both sub-0.2%. What IS true, and is plainly what the seed meant: June-2026 (+0.118) is the LOWEST shelter MoM since Jan-2021 (+0.097). VINTAGE CAVEAT, and it matters: CPI seasonal factors are re-estimated annually, so those 2021 prints may well have read >= 0.2 on the vintage visible on 2026-07-14. This corrects the wording against today's vintage; it is NOT evidence the seed was wrong when written. SCORING STATE at 2026-09-07 — print 1 of the 2 is ALREADY IN and nobody could see it: July = +0.139, sub-0.2%, corroborating trend-break. June +0.118 and July +0.139 sit against a Jan–May 2026 run-rate of +0.220 / +0.232 / +0.266 / +0.607 / +0.318. The August print is the second and final scoring print and releases before this window closes. DATA GAP: 2025-10 is absent at source, the only calendar gap since 1990. It does not touch this entry's June/July/August path, but any longer shelter window rests on a missing base month. RESOLUTION RULE — *** RULED 2026-09-07 by owner ratification ('I'll go with your recommendation'). SETTLED, and settled BEFORE the deciding August print, which was the entire point of asking early. *** Written when it was because the August print was not out yet and this argument is unwinnable afterwards: TREND-BREAK if the August print is also sub-0.2% (three consecutive); BLIP if August prints >= 0.2% and returns toward the Jan–May run-rate. Because the entry carries NO directional bet (column none, axis), EITHER outcome resolves as a recorded OBSERVATION and scores for and against nobody — it must not be counted as evidence for the debasement thesis in either direction. Scored on the served CUSR0000SAH1 vintage current on 2026-09-11. *** SCORED 2026-09-12 on the pre-written rule: BLIP. *** The August print landed in the 2026-09-11 CPI release (source_available_at 2026-09-11) and reached the payload in the 2026-09-12T00:30Z collector run; read on build e041db12a129. Served index 430.227 (2026-08) against 429.095 (2026-07) = +0.264% MoM — AT OR ABOVE the 0.2% line and back inside the Jan-May run-rate (+0.220 / +0.232 / +0.266 / +0.607 / +0.318), so the three-consecutive condition fails on its third print after June +0.118 and July +0.139. MoM is arithmetic on two served index values; the same arithmetic reproduces the June and July figures recorded above to the third decimal, which is why it is trusted here. Corroborating ladder, served: 3m SAAR 2.10, 6m 3.47, YoY 3.03 (39th pct of 38.9y), impulse decelerating positive. NO DIRECTIONAL CREDIT OR DEBIT, per the ruling above: this scores for and against nobody and must never be cited as evidence for or against the debasement thesis. STATUS VOCABULARY, named because it is a limitation and not a verdict: `partial` is the only TERMINAL status that fits (derive/tracker.py treats verified/falsified/partial as resolved). It is used here for the SPLIT ON THE WORDING already recorded above — 'lowest since Jan-2021' true, 'first sub-0.2% since Jan-2021' false on the current vintage — and NOT to imply the watch half-succeeded. A neutral terminal status for no-bet observations is proposed for a ruling in RULINGS_PENDING; if it is granted, this entry moves to it unchanged. | Owner, seed | 2026-09-11 |
| resolved | 10y (DGS10) stays < 5.0% through 2026 FALSIFIED 2026-09-15, marked 2026-09-24 on owner ratification ('go with your recommendations'): served DGS10 closed 5.00 on 2026-09-15 (not < 5.0), then 5.01 (09-16, 09-18) and 5.11 (09-23, highest since 2007-07-13), build 95444956b014; flagged by the 2026-09-24 Yardeni check. This counter failing says the benign bound broke; it does not by itself confirm acute fiscal dominance. Falsifies acute fiscal-dominance timing. Observable: DGS10 (scoreable now). REWRITTEN 2026-07-15 (review I-8): the original bundled '...with no 10y/30y auction tails in the extreme band', whose observable (auction internals) is NOT built — a falsifier pointing at a nonexistent observable looks honest while it isn't. The auction-tail rider is split out below as UNSCOREABLE until the TreasuryDirect collector lands, so it stops counting as a live falsifier. | Benign counter (BIAS_REVIEW) | 2026-12-31 |
| resolved | The next Fed policy-rate change is a HIKE (stated confidence '100%'; 'must be coming'; >=6 more months of intensified tightening). id: howell_2026_next_move_hike. VERIFIED - resolved 2026-09-16 (the next move was a hike), marked 2026-09-24 on owner ratification ('go with your recommendations'). The FOMC raised the range +25bp to 3.75-4.00%, 12-0; served footprint DFF 3.63 (09-16) -> 3.88 (09-17), build 95444956b014. Reported in the same breath: his words nearest the meeting leaned away from an imminent move - 08-21 'tactical postponement', 09-07 (New Era Finance, recorded nine days before) only 'it may well be that they decide', ~09-09 'pantomime' (check 2026-09-24-howell-session05-new-era-finance.md). Checkpoints: Jackson Hole late-Aug 2026 (his stated tell); FOMC 2026-09-15/16; FOMC late-Oct (one week pre-election); December. FALSIFIED BY: a cut as the next move, OR year-end with no hike (the 'must be coming' framing was near-term). Context at seed: rate_path DGS2-DFF +59bp 'hikes-priced' (supports); Fed funds futures ~2/3 odds of no September move per MacroVoices desk 2026-08-13 (timing pressure against). Error type accepted: scoring a near-term-framed call on its stated year-end window may under-credit a hike that lands in early 2027 — the window is his framing, not ours. RESTATED, NOT REVERSED — recorded 2026-09-12 on owner ratification, under the RULINGS §28 discipline (a dated entry is not rewritten by what the analyst says inside its own window). 08-21 (Risk Takers, Session 03): 'Warsh was not put in there to cut rates. What he was probably put there to do was to work alongside... Bessent and keep rates where they were or where they are', and 'my guess would be that this is a tactical postponement... ultimately interest rates have to go up'. ~09-09 (Triangle, Session 04): central banks are 'buying themselves time' and the funds decision is 'a big pantomime'. TIMING SOFTENED, DIRECTION UNCHANGED. He never restated the opposite — no cut, no 'there will be no hike' — so this is NOT a reversal and this entry resolves exactly as written. A September hold ADVANCES THE CLOCK and earns NO credit for 'postponement': that line is not a second call, it is deliberately not seeded, and crediting it would make the seat unfalsifiable in aggregate. A September hike CONFIRMS, and every report of that confirmation carries the softening in the same breath, because his words nearest the meeting leaned away from an imminent move. Served context at recording: DGS2 4.19 (08-20) -> 4.43 (09-09) and T10Y2Y 0.50 (08-21) -> 0.39 (09-10), a bear flattener moving toward the 08-13 version of the call rather than the 08-21 one. See data/panel/checks/2026-09-11-howell-sessions02-04.md §1. | Howell (CrossBorder Capital) — MacroVoices #545, produced 2026-08-13 (primary) | 2026-12-31 |
| resolved | The 2026 rise in the 10y is supply/policy-priced REAL rate, not growth or inflation repricing: through the Sept FOMC (2026-09-16) the 20-session decomposition stays real-led or mixed with breakevens contributing under a third of the nominal move, while the inflation-fundamentals impulse never turns accelerating. id: read_20260818_real_led_supply. FALSIFIED 2026-09-11 on its own first clause, marked by owner ratification the same day (RULINGS §30): decomposition.windows.20s served tag inflation-led with d_breakeven +10bp against d_real +3bp (+13bp nominal) on the 2026-09-09 vintage (build 1ad0dc3b816e, collector 2026-09-11T13:24Z) — inside the window. The oil rally ran through breakevens over the 20 sessions even though YTD (+52 real of +64) and since-trough (+78 real of +86) stayed real-led: the read got the year right and the window wrong, and the window is what was written. Scored as written, not rescued. SEEDED AT (served, build 8a9dcd25d166): 10y 4.68 / real 2.41 / BE 2.27; YTD +49bp nominal = +47 real +2 BE; since the 252s trough (2025-10-22) +71 = +73 real -2 BE; ACM TP 0.83 (54.2nd pct of 36.6y, rising-not-high) carrying ~half the 20s move (tp_share 0.52, real_driver mixed); inflation_fundamentals mixed/decelerating with ALL FIVE components decelerating (impulse norm -1.0); growth_fundamentals expansion/ decelerating (impulse -0.444, 6 of 9 slowing). The read: with inflation compensation flat and growth momentum cooling, the residual drivers of +73bp real are SUPPLY (fiscal engine: interest/receipts 25.55% at the 100th pct of 10.8y; marginal coupon 4.39 vs avg 3.44) and POLICY-PATH repricing (front-end spread +52bp; Howell hike call) - not growth optimism. FALSIFIED IF: decomposition.windows.20s serves tag inflation-led with |d_breakeven_bp| > |d_real_bp| on any vintage inside the window, OR inflation_fundamentals.impulse.direction serves 'accelerating' on any vintage inside the window. Either outcome means the disinflation/real-led frame mis-read the move - which is the debasement-guardrail failure mode, logged here precisely so it can score against me. ERROR ACCEPTED: type-1. A dovish Jackson Hole + clean auction week can fade the term premium and leave this read moot-but-unfalsified; an unfalsified outcome is NOT a win and scores qualitatively against the decomposition path, never as a binary pass. DISCRIMINATORS THIS WEEK: 20Y auction 2026-08-19 + 30Y TIPS reopen 2026-08-20 (a TIPS reopening, NOT a nominal 30Y - corrected 2026-08-21; absorption z's, tail_manual_pending - owner paste supersedes proxy tails); Jackson Hole late Aug (Howell's stated tell). Cross-references: howell_2026_next_move_hike, dale_1998_correction (both checkpoint FOMC 2026-09-15/16) - do not double-count. | read 2026-08-18 | 2026-09-16 |
| resolved | Treasury will act to arrest disorderly long-end repricing and will escalate the tool as the strain level rises: the 2026-08-19 doubling of the 10-20y/20-30y buyback cap is the second point in a reaction function (after the 2023 bill-shift), and further strain produces further response. id: read_20260820_long_end_repression. Owner-framework relevant. This entry is the OBSERVABLE that separates the two readings of 2026-08-19 — dossier §1d.1 / MacroVoices desk ('duration management, not YCC') vs Dale, El-Erian ('already YCC' / 'broader deployment of YCC'). CONFIRMS, ascending: (1) realised sizes at/near the $4bn cap in the 10-20 and 20-30 buckets from the first enlarged op 2026-09-10 (then 09-24, 10-01, 10-08, 10-15, 10-27, 11-04 — tentative schedule 2026-08-05); (2) a further cap raise or off-schedule op before 11-04; (3) program extended or enlarged at the 11-04 QRA; (4) SLR exemption for Treasuries (the named escalation). FALSIFIED IF: DGS30 clears above 5.50% on a sustained basis (>= 5 sessions) with no new support op, no cap raise and no issuance-mix response. ALSO FALSIFIED IF the program lapses at 11-04 while the long end is still under strain (expiry into a calm market proves nothing). NOT falsifiers: gold, equities or the 30Y falling. SECONDARY TEST (does it hold): DGS30 back above 5.30% within four weeks of 09-09 with ops running at the new cap => intervention cosmetic, claim weakens. CONTAMINATION: the 08-19 20Y and 08-20 30Y auctions are VOID as organic-demand reads (the announcement landed ~4h before the 20Y). SEEDED AT (served, build 8b79e3cf21ca): DGS30 5.19 (08-19) off 5.31 close 08-17 / 5.34 intraday; 20Y auction 08-19 dealer_pct_z +0.55, bucket capacity_exhausted; ACM TP 0.84 (54.8th pct of 36.6y). ERROR ACCEPTED: type-1 — a cap-sized op into a calm market will look like confirmation without strain; score realised size AGAINST strain, never alone. Observable: DGS30, auction_internals, calendar buyback ops. Realised op sizes ARE COLLECTED from 2026-08-21 (collect/sources/treasury_buybacks.py, FiscalData; BUYBACK_LS_10Y20Y_* / BUYBACK_LS_20Y30Y_*: OFFERED_BN, ACCEPTED_BN, CAP_BN, FILL_PCT). CALIBRATION 2026-08-21 from the history: long-end ops filled AT CAP in 23 of 25 since 2025-11 with offers 4-18x the cap, so confirm (1) 'realised at the $4bn cap' is the DEFAULT outcome, not a signal - downgraded to a precondition. The informative reads are OFFERED_BN (trend DOWN: ~36bn/op in Mar-Apr 2026 -> ~20bn in Jul-Aug) and FILL_PCT < 100 (Treasury rejecting on price, as on 2026-03-19 at 10% and 2025-11-20 at 40%). Put-like signature = 100% fill with RISING offered volume after 09-10; price-discipline signature = fill below cap. BESSENT 2026-08-20 (CNBC via Reuters): 'it could be more than the $4 billion per issue' and 'part of it is signaling' — escalation narrated one day after the announcement, before any enlarged op; counts as CONFIRMS-(2) talk, not action, until a cap raise or off-schedule op is published. Cross-references: read_20260818_real_led_supply, dale_1998_correction, howell_2026_next_move_hike — do not double-count. | read 2026-08-20 (QRA_BASELINE.md §2b; Treasury sb0607 primary) | 2026-11-04 |
| resolved | The four-quarter trailing real GDP trend prints a one-handle (<= 1.9%) at Q3-2026 — the 'growth turndown' to roughly 1.5% he called on 2026-09-06. id: rosenberg_q3_gdp_onehandle. disclosed-interest (§28: first promotional session). FALSIFIED - resolved 2026-10-29 (BEA Q3-2026 advance estimate), marked 2026-11-02 on owner ratification ('go with your recommendations'). GDPC1 four-quarter rate +2.07% on the advance vintage (bar: >= 2.0 falsifies); the served mechanical check agreed (build 39c28edddb17). Q3 grew +3.37% annualised against +3.10% needed to falsify; a repeat of trend would have printed +1.77% (confirm). The base effect held (2.19 -> 2.07); Q3 accelerated. Not a boom signal - one advance estimate. Check: data/panel/checks/2026-11-02-rosenberg-q3-gdp-onehandle.md. LEAN: caution / cyclical counter. OBSERVABLE: GDPC1, four-quarter trailing growth, at the BEA advance estimate for Q3-2026. CONFIRMS: <= 1.9%. FALSIFIED IF: >= 2.0%. *** THE BASE-EFFECT FINDING IS PART OF THIS ENTRY, NOT A FOOTNOTE. *** Seeded on the owner's ruling that it ships WITH the measurement or not at all. Measured 2026-09-07 on build 9f9176311288: the four-quarter trailing rate is +2.10% (GDPC1 as_of 2026-04-01), and for it to print 1.5% at Q3 the quarter need only come in at +1.95% annualised. If Q3 merely REPEATS the current four-quarter trend (+2.06%), the trailing rate prints +1.53% ANYWAY, because Q3-2025's +4.38% annualised quarter rolls out of the window. Roughly three quarters of the fall is arithmetic. Q3 would have to ACCELERATE to about +4.0% annualised for the trailing rate to keep a 2-handle. SCORING INSTRUCTION: score the call, do not inherit the alarm. A CONFIRM here is evidence he read the arithmetic correctly and is NOT evidence of a growth scare; any brief quoting this entry as support for S5 must carry the base-effect sentence with it. ERROR ACCEPTED: type-1 — an arithmetically near-certain call scoring as foresight. Accepted deliberately, and the note above is the mitigation. Cross-reference rosenberg_2027_disinflation (same seat, different measure) — do not double-count them as two independent confirmations of one thesis.
| Rosenberg (Rosenberg Research) — Thoughtful Money, published 2026-09-06 (primary; checks in data/panel/checks/2026-09-07-rosenberg-thoughtful-money-session03.md) | 2026-11-30 |
| resolved | Warsh does not raise rates at the 2026-09-16 FOMC — 'he's not, he can't' — because true interest expense is already ~105% of receipts and growing at twice receipts. id: gromen_no_sep_hike. FALSIFIED - resolved 2026-09-16, marked 2026-09-24 on owner ratification ('go with your recommendations'). The FOMC raised the range +25bp to 3.75-4.00%, 12-0; served footprint DFF 3.63 (09-16) -> 3.88 (09-17), build 95444956b014. No escape hatch. His fiscal arithmetic (106.83% on his basket) stands separately; the prediction built on it did not. He did not score the miss in his next primary session (~09-21 GoldRepublic; check 2026-09-24-gromen-goldrepublic-macroscopic.md section 1): the hike is folded into 'both Fed cuts and Fed rate hikes are inflationary'. LEAN: fiscal dominance / no-policy-room. COLUMN none:state deliberately — a Fed that cannot hike is evidence about the REPRESSION REGIME, not progress along any of the six paths. The mirror entry rosenberg_sep_hike is s6 because a HIKE would be evidence for nominal boom; this one verifying is not evidence for s5, and must not be counted as such. *** THE DELIBERATE MIRROR. *** Same observable, same meeting, same resolution rule as rosenberg_sep_hike, inverted. A HAWKISH HOLD CONFIRMS GROMEN AND FALSIFIES ROSENBERG. That symmetry is the reason to seed it: one meeting now scores two seats against each other on one instrument, with both rules written before the event. OBSERVABLE: the FOMC decision, DFF daily_close as its served footprint. CONFIRMS: the 2026-09-16 meeting holds or cuts. FALSIFIED IF: the target range rises. NO ESCAPE HATCH — he offered none, and none is supplied. Contrast rosenberg_sep_hike, whose VOID clause exists only because Rosenberg stated the condition himself. ERROR ACCEPTED: type-1 — a hold for reasons unrelated to his mechanism (true interest expense) scores as a win. His mechanism is not separately observable at the meeting, and inventing a proxy for it would manufacture a test he did not accept. MECHANISM VERIFIED SEPARATELY, and it is the strongest fiscal arithmetic on file: his 105%-of-receipts claim measured 106.83% on his own basket, growing +7.53% against receipts +3.57% (2.11x) — check file §2.
| Gromen (FFTT) — BTC Sessions joint session, published 2026-09-04 (primary; checks in data/panel/checks/2026-09-07-gromen-alden-btc-sessions-joint.md) | 2026-09-16 |
| resolved | Warsh hikes at the 2026-09-16 FOMC, absent a very weak payroll print (his stated bar: roughly +0.2 on U3 and 0.1 on core). id: rosenberg_sep_hike. VERIFIED - resolved 2026-09-16, marked 2026-09-24 on owner ratification ('go with your recommendations'). The FOMC raised the range +25bp to 3.75-4.00%, 12-0; served footprint DFF 3.63 (09-16) -> 3.88 (09-17), build 95444956b014. Escape hatch did not fire (UNRATE Aug 4.1 = Jul 4.1; core CPI +0.3 vs 0.2). RULINGS SECTION 30 SENTENCE: he reversed this within days of making it - on the record by 2026-09-05, eleven days before it resolved; his 09-06 reason (vote arithmetic, 'doesn't have seven') was wrong: 12-0. Trajectory: no-hike (08-05, 08-12) -> hike (09-01, this entry) -> no-hike (~09-01..03, 09-06) -> on decision morning a hike assumed rather than disputed (09-16 Excess Returns; check 2026-09-24-rosenberg-excess-returns-sep16.md). A hit on a call he had withdrawn; report it that way. LEAN: caution / policy-tightening — same side as howell_2026_next_move_hike, arrived at from the opposite seat. Rosenberg's own framework says the Fed SHOULD NOT hike (rosenberg_2027_disinflation); this entry is his PREDICTION that it WILL. Normative and predictive are not a contradiction and must never be scored as one. MECHANISM: a new chair establishing an anti-inflation reputation acts on the tariff-driven goods impulse rather than the decelerating services side. OBSERVABLE: the FOMC decision, with DFF daily_close as its served footprint (effective fed funds steps to the new target range within days); UNRATE and CPILFESL for the escape hatch. RESOLUTION RULE — written 2026-09-02, BEFORE the event, at the owner's instruction, because this is precisely the argument that is unwinnable afterwards. CONFIRMS: the 2026-09-16 FOMC raises the target range. Nothing else confirms. FALSIFIED IF: the meeting holds or cuts and the escape hatch below did not trigger. A HAWKISH HOLD IS FALSIFIED, NOT PARTIAL — this is the case the rule exists to settle. A hold with hard guidance, a hawkish dot plot, or a dissent-for-hike scores the claim WRONG. The claim is about an ACTION at a DATED meeting, not about a stance. Scoring a signal-hard hold as a win would make the entry unfalsifiable, and a non-firing mechanism must never count as confirmation. Record the hawkishness as context; award it no credit. VOID — not activated, never scored as wrong: the last Employment Situation report published before the 2026-09-16 decision prints U3 at least 0.2pp above its prior vintage, or the corresponding core print comes in 0.1pp soft against consensus. That is HIS OWN stated escape and it routes to not-activated. AMBIGUITY, DECLARED: the surviving source says only "0.1 core" and does not name the measure. Where the escape hatch is the DECIDING factor and the reading is genuinely ambiguous, this entry resolves VOID rather than being scored on our reading of his words — same discipline as the declines above. The escape hatch can only VOID; it can never convert a miss into a win. *** REVERSED BY ITS OWN AUTHOR, 2026-09-06 — RULINGS §28, RULED 2026-09-07. *** Five days after making this call he said the opposite, on Thoughtful Money: "I don't think they're going to raise rates," on vote arithmetic (the chair needs seven and, even with the three July dissenters, Rosenberg does not think he has them). THIS ENTRY IS NOT AMENDED, VOIDED OR SUPERSEDED BY THAT. It resolves on 2026-09-16 exactly as the rule above was written on 2026-09-02, before the event. THE GENERAL RULE §28 ESTABLISHES, binding on every entry in this file: a dated entry is not amended, voided or superseded by the analyst restating it inside its own window. The restatement is recorded beside it as a REVERSAL. Reason: the resolution rule was pre-written precisely so the record could not be renegotiated once the answer was known, and a reversal ten days out is a renegotiation with better manners. REPORTING RULE: the pair is one reversal, never two independent calls. The 09-06 view is deliberately NOT seeded — seeding both sides of one meeting would make him unfalsifiable in aggregate, which is the single outcome the pre-written rule exists to prevent. Whatever this entry resolves to, any report of it must carry the sentence "he reversed this within days of making it — on the record by 2026-09-05, eleven days before it resolved." (AMENDED 2026-09-11 by owner ratification, RULINGS §30: the sentence originally read "on 2026-09-06, ten days before it resolved"; the Jimmy Connor session, published 2026-09-05 and recorded before the 2026-09-04 payroll, already carries the no-hike view — see data/panel/checks/2026-09-11-rosenberg-jimmy-connor.md §1. The ruling's substance is unchanged: this entry stands and resolves as written.) See data/panel/checks/2026-09-07-rosenberg-thoughtful-money-session03.md §7. ERROR ACCEPTED: type-2 — the escape hatch is generous and will let a genuine miss resolve VOID. Deliberate: he stated the condition himself, and a falsifier that ignores a caller's own stated escape is not scoring him, it is scoring a strawman. CONTEXT AT SEED (served, build 78f16b8c334e, 2026-09-02T21:04Z): DFF 3.63 (2026-08-31, 52.3rd pct of 40.1y); DGS2 4.34, DGS10 4.75, 2s10s +41bp; T10YIE 2.35 (2026-09-01). CROSS-REFERENCES: howell_2026_next_move_hike (same direction, WIDER window to 2026-12-31 and unconditional on the meeting — a Sept hold falsifies this entry and merely advances that one's clock); dale_1998_correction (conditional, checkpoints the same meeting); rosenberg_2027_disinflation (the same analyst's normative opposite). DO NOT DOUBLE-COUNT — one meeting, four exposures, scored independently, per the FOMC-September pair note above.
| Rosenberg (Rosenberg Research) — David Lin session, 2026-09-01 (owner-supplied draft, processed 2026-09-02; checks in data/panel/checks/2026-09-02-rosenberg-lin-session.md) | 2026-09-16 |
| resolved | The Fed raises the target range at the 2026-09-16 FOMC ('honestly, I would say 90%', against a market 50/50). id: wang_sep_hike_90. VERIFIED - resolved 2026-09-16, marked 2026-09-24 on owner ratification ('go with your recommendations'). The FOMC raised the range +25bp to 3.75-4.00%, 12-0; served footprint DFF 3.63 (09-16) -> 3.88 (09-17), build 95444956b014. Sub-claim (a skip means December) moot. He scored it himself on 09-16: 'I did call this perfectly' (check 2026-09-24-wang-markets-weekly-0912-0919-fomc-live.md). LEAN: policy-tightening, from the plumbing seat. OBSERVABLE: the FOMC decision; DFF daily_close as its served footprint (effective fed funds steps to the new range within days). CONFIRMS: the 2026-09-16 FOMC raises the target range. FALSIFIED IF: it holds or cuts. A hawkish hold is FALSIFIED, not partial — the same rule as rosenberg_sep_hike. His stated 90% is context, not a softener: a hold scores the call wrong. SUB-CLAIM, recorded not scored: a skipped September means December, not October (the midterm calendar, his Framework 4). NO ESCAPE HATCH — he offered none, and none is supplied. CROSS-REFERENCES: rosenberg_sep_hike, howell_2026_next_move_hike, bianco_sep_hike (same side); gromen_no_sep_hike (mirror). One meeting, scored independently — do not merge. ERROR ACCEPTED: type-1 — a hike confirms a 90% call no more strongly than a 60% one. CONTEXT AT SEED (served, build 1ad0dc3b816e, collector 2026-09-11T13:24Z): DFF 3.63 (2026-09-09); DGS2 4.43; front-end spread 80bp, 78.7th pct of 40.1y.
| Wang (Fed Guy / Monetary Macro) — Markets Weekly 2026-08-29, restated as strengthened 2026-09-05 (primary captions; checks in data/panel/checks/2026-09-11-wang-markets-weekly-0829-0905.md) | 2026-09-16 |
| resolved | The Fed raises the target range at the 2026-09-16 FOMC unless it is 'a real shocker week'; it should. id: bianco_sep_hike. VERIFIED - resolved 2026-09-16, marked 2026-09-24 on owner ratification ('go with your recommendations'). The FOMC raised the range +25bp to 3.75-4.00%, 12-0; served footprint DFF 3.63 (09-16) -> 3.88 (09-17), build 95444956b014. His vote count (5 hike / 6 hold / 1 unknown, check 2026-09-11 section 6) was wrong: 12-0; he reframed it as chair-decides rather than scoring it (check 2026-09-24-bianco-fomc-week.md). LEAN: policy-tightening, from the bond-vigilante seat. OBSERVABLE: the FOMC decision; DFF daily_close as its served footprint. CONFIRMS: the target range rises on 2026-09-16. FALSIFIED IF: it holds or cuts. A hawkish hold is FALSIFIED, not partial. ESCAPE HATCH (his, loosely stated): 'a real shocker week'. Assessed BEFORE the event: the only data between his statement and the meeting were PPI (2026-09-10) and CPI (2026-09-11, core +0.3 m/m against a 0.2 consensus per BLS) — neither is a dovish shock, so the hatch did not trigger. It could only ever VOID, never convert a miss into a win. CROSS-REFERENCES: rosenberg_sep_hike, howell_2026_next_move_hike, wang_sep_hike_90 (same side); gromen_no_sep_hike (mirror); bianco_hike_caps_long_end (his conditional on what the decision does to the long end — a separate claim, scored separately). ERROR ACCEPTED: type-1 — a hike for reasons unrelated to his vigilante mechanism scores as a win; the mechanism is tested by bianco_hike_caps_long_end, not here.
| Bianco (Bianco Research) — The Money Path, CrossCheck Media, 2026-09-09 (primary captions; re-posted by Bianco Research 2026-09-10, same interview; checks in data/panel/checks/2026-09-11-bianco-money-path.md) | 2026-09-16 |
| resolved | Bond traders can stop panicking when the Fed starts panicking: if the Fed hikes on 2026-09-16 the 10-year peaks; if it does not act, the 10-year goes 'shooting through 5%'. id: bianco_hike_caps_long_end. FALSIFIED ON DAY ONE (hike branch) - resolved 2026-09-16, marked 2026-09-24 on owner ratification ('go with your recommendations'). Rule: FALSIFIED IF DGS10 closes at or above 5.00 inside the span despite the hike. Served DGS10 09-16 = 5.01 (unrevised through 09-24), then 5.01 (09-18) and 5.11 (09-23), build 95444956b014. The entry's own accepted error covers exactly this. His later conditionals (09-17 'marginally higher, not dramatically higher'; 09-21 'yields would go down') are recorded in his profile beside this entry, not in it (section 28). THE WINDOW IS OURS, NOT HIS. He stated no horizon. RULINGS §30 (owner-ratified 2026-09-11) authorised supplying one — a bounded departure from the §23 practice of declining windowless calls, allowed because the claim is keyed to a dated event, both branches are fixed before it, and the window ends at the next structural date the long end prices (the 2026-11-04 refunding). BRANCH BY OUTCOME, fixed before the meeting: HIKE branch — CONFIRMS: DGS10 makes no daily close at or above 5.00 from 2026-09-16 through 2026-11-04. FALSIFIED IF: DGS10 closes at or above 5.00 inside that span despite the hike. HOLD (or cut) branch — CONFIRMS: DGS10 closes at or above 5.00 inside the span. FALSIFIED IF: it never does. OBSERVABLE: DGS10 daily close — the cash series, never the stitched future (§21). CROSS-ANALYST MIRROR: rosenberg_nov_refunding expects a hold plus the 11-04 refunding to rally the 10-year; a hold with DGS10 at or above 5.00 confirms this entry's hold branch and weighs against his. gromen_bond_rupture_q4 (DGS10 at or above 5.52 by 2026-12-04) is the fiscal-dominance version of the hold branch — do not double-count a single breach. ERROR ACCEPTED: type-2 — a peak that forms after 2026-11-04 is outside a window he never set; and a hike followed by a brief 5.00 close scores FALSIFIED even if yields then peak. CONTEXT AT SEED (served, build 1ad0dc3b816e): DGS10 4.83 (2026-09-09), 60.1st pct of 40.1y; the highest close in served history since 2023 is 4.98 (2023-10-19).
| Bianco (Bianco Research) — The Money Path, CrossCheck Media, 2026-09-09 (checks in data/panel/checks/2026-09-11-bianco-money-path.md) | 2026-11-04 |
| resolved | The 2026-09-30 release — a fresh set of totally revised PCE deflator data — will make inflation look a lot lower than it does now. id: rosenberg_pce_revision_lower. disclosed-interest (section 28 conduct trigger — the session pitches ROSY and names its prospectus). LEAN: benign / disinflation — his seat. VERIFIED (CONFIRMS) - resolved 2026-09-30, marked the same day on the owner's instruction ('yes score rosenberg'). ALFRED realtime 2026-09-29 vs 2026-09-30: PCEPILFE 2026-07 YoY 3.34 -> 2.98 (-0.36pp, bar 0.20pp); PCEPI 3.70 -> 3.36 (-0.34pp); history revised back through 2025, so not VOID. Pre-release figures reproduce CONTEXT AT SEED exactly; post-release levels match the served store (build 70e85dce6ce2, run-2026-09-30T13:00:04Z). A lower history is not a disinflation print — the new 2026-08 observation (+0.25% core MoM) is not scored here. Check: data/panel/checks/2026-09-30-rosenberg-pce-revision.md. MECHANISM: the revision re-estimates the deflator history; he expects the revised path to show lower inflation than the pre-revision vintage. OBSERVABLE: PCEPILFE (core) YoY for 2026-07 — the latest month common to both vintages — on the last pre-release vintage vs the first post-release vintage (vintage rows are preserved in the store; ALFRED realtime 2026-09-29 vs 2026-09-30 is the cross-check). PCEPI headline is the secondary leg. CONFIRMS: 2026-07 core PCE YoY revised DOWN by 0.20pp or more. PARTIAL: down 0.10-0.19pp, or core down less than 0.20pp while headline is down 0.20pp or more. FALSIFIED IF: 2026-07 core PCE YoY is revised down less than 0.10pp, unchanged, or UP. VOID IF: the 09-30 release revises no month before 2026-08 — the revision he named did not occur. THE 0.20pp BAR IS OURS (he said 'a lot lower'); set by owner ratification 2026-09-24, before the release, and never moved after it. NOT SCORED HERE: the new 2026-08 observation and any later disinflation — rosenberg_2027_disinflation's job. ERROR ACCEPTED: type-2 — a revision that lowers earlier years but leaves 2026-07 roughly unchanged scores FALSIFIED. CONTEXT AT SEED (served, build 95444956b014, collector 2026-09-24T13:22Z): PCEPILFE 2026-07 YoY 3.34, m3 SAAR 3.05; PCEPI 2026-07 YoY 3.70; next_release_at 2026-09-30 for both. CROSS-REFERENCES: rosenberg_2027_disinflation (its core-PCE falsifier reads the revised vintage from 09-30 on — a downward revision is NOT a disinflation print; do not double-count). CHECKPOINT: score on the first collector run after the 2026-09-30 release. Check: data/panel/checks/2026-09-24-rosenberg-excess-returns-sep16.md section 7.
| Rosenberg (Rosenberg Research) — Excess Returns, recorded 2026-09-16 before the FOMC decision, published 2026-09-17 (primary) | 2026-10-07 |
| resolved | SPX 8000 before the election is my call — Trump needs a strong stock market for the midterm, and his tools (social media, a ceasefire or peace declaration, oil, jawboning Treasuries) are understated id: owner_spx_8000_before_midterm. FALSIFIED 2026-11-04, marked on the owner's "commit" after "No 8000 SPX" (in chat): no SP500 close >= 8,000.00 in 2026-10-12..11-02. Highest close in the window 7,807.24 (10-12), -2.4% short; last session 11-02 closed 7,541.40 (served FRED SP500, build 90f0284363e3, run 2026-11-04T00:30Z). House counter HV-016 (40%) resolves on the same close: Brier 0.16. The mechanism named one lever that did fire — Brent -7.6% on 11-02 (BZ_F_F27 113.43 -> 104.77) — and the index did not follow. Mechanism unscored, as the rule says. LEAN: benign / risk-on — the political-put reading of the midterm. MECHANISM (the owner's): the White House needs equities strong on 2026-11-03 and its levers are larger than the market credits; "200 points in 2 days is a nothing burger" in a tape with an "insatiable appetite for equities at any price". The principals layer carries the want as INFERRED (white_house: equities, gasoline and jobs in good order on 2026-11-03). OBSERVABLE: SP500 daily_close (never live_overlay; the SPY-derived mark is not a close). CONFIRMS: any SP500 close >= 8,000.00 from 2026-10-12 through 2026-11-02. FALSIFIED IF: no such close. "Before the election" is read as the last session before election day (11-02); a close of 8,000 on 11-03 itself does not count. ERROR ACCEPTED: type-1 — the mechanism is not scored; a close at 8,000 for any reason confirms. CONTEXT AT SEED (served, build 7632d72e1a59): SP500 7,811.54 (2026-10-09 close), +2.41% to the bar, 16 sessions in the window; live SPY-derived mark 7,807.33 (10-12 13:30Z, not a close); closing high 7,818.93 (10-06); VIXCLS live 15.2; positioning leverage sub-read crowded, 94.6th pct of 28.6y; RSP/SPY at its lowest since 2006 (10-07). COUNTER (same event, opposite side): house view HV-016 — the same close at 40%. Measured 2026-10-12 on FRED SP500 2016-10-10 -> 2026-10-09, +2.41% inside 16 sessions: 49% of all starts; 35% of starts within 1% of the one-year high; 30% (15 sessions) for starts near the high with Trump in office; 1 of 4 pre-election windows (2018, 2020, 2022, 2024) reached it. CROSS-REFERENCES: HV-015 (8,000 by 12-31, 80%) — confirms on the same close if it lands by 11-02; score the number once and the date separately. yardeni_2027_spx_8400_mid (different bar). CHECKPOINTS: 10-14 CPI; mid-October buyback blackout; the 10-28 FOMC (P hike 0.18 at 10-09); analyst's stated update rule — near 7,900 by ~10-23 moves the house above even, under 7,700 after CPI moves it under a third.
| Owner, seed — analyst session 2026-10-12 (in chat) | 2026-11-02 |
| resolved | Iran is invested in disrupting the administration before the midterms — the inverse of the frictionless 8,000 call; Brent futures close at $120 or higher before the election. id: owner_iran_escalates_brent_120. FALSIFIED 2026-11-04, marked on the owner's "commit" (in chat): no BZ_F close >= 120.00 in 2026-10-16..11-02. Highest close 117.64 (10-30, the December contract before the roll), $2.36 short; 11-02 closed 104.77 (BZ_F = BZ_F_F27, -7.6% on the day from 113.43), build 90f0284363e3, run 2026-11-03T13:00Z. The roll caveat named at seed worked against a hit as stated; on the explicit January contract the window high was 113.43 (10-30). Cause of the 11-02 fall not recorded in the payload. LEAN: caution / supply shock — against owner_spx_8000_before_midterm in spirit (both can resolve the same way; they are not the same event, so neither scores the other). MECHANISM (the owner's): Iran wants to hurt the incumbent party at the pump before 11-03 and oil is its lever. Analyst's counter at seed: escalation before the vote hands the White House a rally-round-the-flag moment and a reason to resume the strikes it paused "until after the election"; a slow grind may serve Tehran better. Tier 4 (narrative) both ways. The mechanism is not scored; a $120 close for any reason confirms. OBSERVABLE: BZ_F daily close (ICE Brent continuous front; the instrument ruled 2026-08-14 for Brent-benchmark claims). DCOILBRENTEU (Dated Brent, physical) is NOT the bar — it ran ~$20 over the future at seed (126.55 on 10-14 vs BZ_F 106.06) and would confirm on a different market. ROLL CAVEAT: the December contract expires ~10-30; in steep backwardation the continuous front prints LOWER on the roll — that works against a hit and is accepted, not adjusted. CONFIRMS: any BZ_F close >= 120.00 from 2026-10-16 through 2026-11-02. FALSIFIED IF: no such close. BASE RATE (measured 2026-10-16 on Yahoo BZ=F daily 2007-07 -> 2026-10-16, 12-session forward max, overlapping starts): +10% (~120) in 12.4% of all starts; 21.8% when a >=5% day occurred in the prior 20 sessions (not the case at seed: max |day| 4.1%). Symmetric risk: -10% in 13.4% / 21.7%. +15% (~126): 4.6% / 10.3%. No house probability stated (owner's call). ERROR ACCEPTED: type-1 — a hurricane, an OPEC+ act or a roll quirk that prints $120 confirms. CONTEXT AT SEED: BZ_F 105.92 (10-15 close, served brent_wti_spread, build 6cebcd399d9d); live 109.41 owner's BZ1! = Yahoo 109.42 (10-16, not a close); BNO live 67.125 vs 64.86 close; Brent-WTI 7.95 (89.6th pct of 39.4y); WTI DCOILWTICO 101.91 (10-14). Press, not served: tanker attacks at a wartime high (11 in the week to 10-04); Hormuz exports ~8.5 mb/d, ~40% under pre-war (Kpler); US strikes paused until after the midterm; MOAB reported under consideration. CROSS-REFERENCES: owner_spx_8000_before_midterm (same window end); the Johnson Brent entries; the 2026-08-14 BZ_F >= 85 entry (fired); owner_10y_yield_topped_531 (oil is the inflation channel against it). CHECKPOINTS: weekly Kpler Hormuz flows; the 10-28 FOMC; the Dec/Jan roll ~10-30; 11-02 close.
| Owner, seed — analyst session 2026-10-16 (in chat): 'This could be the inverse call to my frictionless SPX 8000 call. Iran is probably very invested to disrupt the current administration for the mid terms.' Level ($120) proposed by the analyst, accepted on 'seed it'. | 2026-11-02 |
| 2026-11-09 | 3-Year note auction, $58.0bn — results ingest at the next 00:30Z runannounced — TreasuryDirect first-party upcoming feed |
| 2026-11-10 | 10-Year note auction, $42.0bn — results ingest at the next 00:30Z runannounced — TreasuryDirect first-party upcoming feed |
| 2026-11-12 | 30-Year bond auction, $25.0bn — results ingest at the next 00:30Z runannounced — TreasuryDirect first-party upcoming feed |
Whether the PLATFORM is working — not a market gauge. Status degraded.
in-app indicator: it cannot report that the app itself is unreachable — if the process is down, nothing answers here at all
Every dated, probability-stated call the analyst has made (data/panel/HOUSE_VIEW.md §7), settled by arithmetic against the served closes on its own window.
The analyst influences every read on this board and was accountable for none of them — the asymmetry the panel exists to remove, reproduced at the centre. A scored mediocre view beats an unscored authoritative one. Opened 2026-08-22; empty for 21 days; auto-scored since 2026-09-12 so it cannot go quiet unnoticed.
data/panel/house_view.yaml mirrors each prose entry as series · operator · level · window · touch/never · p. A `touch` leg verifies on the first qualifying close and falsifies once the window is fully observed without one; `never` is the mirror. Brier = mean (p − outcome)² over resolved legs. Base rates are measured with scripts/base_rate.py and live in the prose entry.
AWAITING DATA = the window closed but the series has not published through it — not a miss. `brier` 0.25 is a coin flip; lower is better; read n beside it. SILENT = no new call for 30 days, which HOUSE_VIEW.md §4 treats as a finding.
Pre-registered fields are never edited; a change of mind is a new HV id citing the old one, which keeps its verdict.
Every dated prediction we or the panel have made, with its window and status.
Forecasts only teach if they can be WRONG on a date. Scoring the panel (and ourselves) honestly is what separates calibration from narrative — falsified entries are announcements, not embarrassments.
Hand-curated in config/tracker.yaml; urgency computed from the window.
OVERDUE = past its window, score it now. Standing = structural claims with long horizons.