Monitor v7 — live research track

Data as of 2026-11-13 · generated 2026-11-14T02:02:17+00:00 · previous state: PHASE_2A

RESEARCH TRACK — v7 framework, parameters frozen at first live deploy 2026-10-01 (44 days since frozen). v7 is the rotation framework: portfolio transitions (Stage 2A/2B/TAIL) key off these states. The v7 arming layer supersedes the v5 Tier-1 gates: Defensive Level 1 (≥2 of 4 raw signals) and Level 2 (B OAS ≥ 450 bps) arm the machine; the G3 yield-exhaustion gate is removed (22.7% 6-month drawdown rate — worse than the 13.8% base rate). The B+65/CCC+200 credit legs moved to the confirmation layer (2A/2B escalation only). 2A/2B exits, the Fed-rollover flag and TAIL_OVERRIDE are unchanged from v5. Evidence: backtest v7 · methodology + track archive: handover.

Tactical track (Track B). Strategic context: /expected-returns/ (Track A — strategic posture, quarterly cadence).

Credit basis: the v7 credit legs read FRED ICE BofA series (B: BAMLH0A2HYB, CCC: BAMLH0A3HYC, daily) with the Bloomberg-calibrated thresholds applied as-is (owner decision 2026-10-01). The ICE BofA basis runs ~8 bps below Bloomberg on B and ~139 bps on CCC (2026-09-30, backtest v6), so the level/velocity triggers are marginally harder to hit live than in the evidence base — the conservative direction.

Fed-rollover flag (automated): FALSE · DGS2 4.86% vs target midpoint 3.875% (DFEDTARU 4.00 / DFEDTARL 3.75) − 0.25 · computed from FRED daily; config.json is not used · PHASE 2B requires this flag on top of 2A

DATA-SUSPECT: the Brent spot input (FRED DCOILBRENTEU) deviates > 15% from the BZ=F front-month close; the BZ=F close stands in on 56 suspect session(s) for the tail-override and oil-proxy inputs. The series is stale at source (last print 2026-11-11; confirmed via FRED's public CSV, not a fetch/parse bug) and the BZ=F close backstops 49 missing session(s).

v7 state

PHASE_2A

23 sessions in state

Level 1 legs firing (raw)

1 / 4 (inactive, need 2)

raw daily fires — deliberately no latch (v7)

Level 2 (B OAS ≥ 450 bps)

COLD

B 339 bps [as-of 2026-11-12] · supersedes Level 1 when hot

Confirmation quorum (latched)

0 / 2 (not met, need 1)

2A/2B escalation only · 15-session latches

TAIL_OVERRIDE inputs

inactive (0d)

10Y weekly close 5.10% > 5.40% (week ending 2026-11-13, partial) · Brent spot $96.08 > $120 (source: BZ=F close (sanity/backstop - DCOILBRENTEU stale/suspect at source), raw $92.14) [as-of 2026-11-11] · joint: False · joint condition never activated

Duration Risk Flag (DRF) — duration overlay, not a v7 state

cold score 0/4 · tranche-2 gate: PASS (DRF cold) · effective IDTL: 2A 25% / 2B 32% (prescribed 25% / 32%)

Leg	Latest vs trigger	State	Note
OIL — supply shock	Brent 20d -19.5% vs +12%	COLD	mandatory leg
SB — hedge failure	10Y 20d +0.01pp vs +0.25pp · SPX 20d -1.2% vs ≤0	COLD
BE — inflation-led	T10YIE 20d -0.11pp vs +0.10pp	COLD
DEB — debasement	GLD 20d -7.0% vs +4% · USD 20d +0.4% vs ≤0	COLD

DRF cold: full prescribed duration permitted; tranche 2 gate passes. Fire = OIL AND ≥1 of {SB, BE, DEB}, ×3 sessions; clears after 15 quiet sessions. Evidence: 2011+ backtest, 26 episodes, 3m fwd duration −2.7% vs +0.4% base; caught Jan-2022 before −15%. Pure debasement (no oil) does NOT trigger — see handover addendum 2026-10-28 §2.3.

Gold gate — 2026-11-01 amendment · effective SGLD: base 10%

Channel A: DEB leg active today? no (GLD 20d -7.0% vs +4% · DTWEXBGS 20d +0.4% vs ≤0); 0/3 consecutive sessions toward set.
Channel B: GLD last W-FRI weekly close $355.45 vs 200-session SMA $408.42 — below.
Overrides: DRF active? no · TAIL? no — either forces full gold.

Gold gate (2026-11-01 amendment): base SGLD 10%; full weight via Channel A (DEB ×3, DRF-hot, or TAIL) or Channel B (GLD W-FRI close > 200d SMA); clear after A = DEB quiet 15 sessions AND below 200d.

State-entry provenance

PHASE_2A since 2026-10-14 (23 sessions)

Entered PHASE_2A on 2026-10-14: Level 1 hot (B OAS velocity + SPHB/SPLV) with CCC +200bps/20d confirmation latched.

Level 1 as of entry

Leg	State	Value
B OAS velocity	FIRED	B 329 bps · 20d +51 bps
SPHB/SPLV	FIRED	1.971 vs SMA50 2.027
10Y-2Y curve	COLD	+0.37 vs trigger +0.45
VIX	COLD	16.9 vs trigger 23.1
Level 2 (B OAS ≥ 450 bps)	COLD	329 bps

Confirmation as of entry

Leg	State	Value
B +65bps/20d	COLD	20d +51 bps (level 329 bps)
CCC +200bps/20d	FIRED + LATCHED	20d +206 bps (level 1,282 bps)

Other inputs as of entry

Fed-rollover flag: FALSE

TAIL joint condition: FALSE

10Y weekly close 5.19% · Brent $106.06

Current episode: NORMAL (2026-10-13) → PHASE_2A (2026-10-14)

Defensive Level 1 legs (raw daily — no latch; need ≥2 of 4)

L1 · B OAS velocity

COLD

B 339 bps · 20d +4 bps [as-of 2026-11-12]
B OAS +50 bps over 20 sessions (ICE BofA B, BAMLH0A2HYB)
last fire 2026-11-06 (5 sessions ago)

L1 · SPHB/SPLV ratio

FIRED

1.702 vs SMA50 1.959 [as-of 2026-11-13]
< 50-day SMA
last fire 2026-11-13 (0 sessions ago)

How to read the beta leg: raw fires are informative when SPY is above its 200d SMA and near highs — stealth rotation, ~24% hit rate vs 19.8% base. Standalone fires in a down tape carry no signal (~17%, below base). This never changes the L1 quorum: down-tape fires must coincide with a second stress leg, and that subset is the higher-precision one (~28%). The leg stays raw by design (tested 2026-10; conditioning = no-op at quorum level: 113 vs 117 fires, identical trigger dates in all 6 episodes).

L1 · 10Y-2Y curve

COLD

+0.27 vs trigger +0.45 [as-of 2026-11-13]
>= trailing 12-month min + 0.25
last fire 2026-10-08 (26 sessions ago)

L1 · VIX

COLD

18.7 vs trigger 29.0 [as-of 2026-11-12]
> 1.5 x 20-day SMA (VIXCLS)
last fire 2026-11-05 (6 sessions ago)

Confirmation layer — 2A/2B escalation only (15-session latch)

#	Confirmation	Latest vs trigger	State (raw / 15d latch)	Recency
1	CF-1 · B OAS velocity +65 bps over 20 sessions (ICE BofA B, BAMLH0A2HYB)	20d +4 bps (level 339 bps) [as-of 2026-11-12]	COLD	last fire 2026-10-20 (18 sessions ago)
2	CF-2 · CCC OAS velocity +200 bps over 20 sessions (ICE BofA CCC, BAMLH0A3HYC)	20d +17 bps (level 1,314 bps) [as-of 2026-11-12]	COLD	last fire 2026-10-23 (15 sessions ago)

Legacy v5 Tier-2 context (raw; not in the v7 state machine)

Legacy leg	Latest vs trigger	Raw state
C1 · Initial claims, 4-week MA >= trailing 1-year min + 25k	195,250 vs trigger 220,250	COLD
C2 · GLD/CPER ratio > 1.15 x 50-day SMA	9.730 vs trigger 11.115	COLD
C4 · Brent term structure (roll-yield proxy) BNO 60-day minus Brent spot 60-day <= -2%	+11.94%	COLD

Context only — not in the v7 state machine

KOSPI watch

KOSPI drawdown from peak is at a depth historically seen only with SPX already −18–30% — unprecedented configuration with SPX near highs; treated as readiness, not a trigger. Tested twice as a v7 leg/gate and rejected (2026-10; see handover addendum §3). Card invalidates on the first of: KOSPI weekly close ≥ 6,800 · 10Y < 4.85% · mid-Jan-2027.

Context only — not in the v7 state machine

AI Credit Watch

Master trigger = Anthropic IPO pricing vs ~$2T mark. Tripwires: CRWV CDS > 800bp · Jupiter marks < 85¢ · ORCL CDS widening vs BB · OpenAI round ≤ $852B. Display only — no v7 linkage.
Update 2026-10-28: CRWV CDS tripwire BREACHED — printing ~950–1,000bp mid-October vs the 800bp line; ORCL downgraded BBB-. Anthropic IPO master trigger still pending.

Rules summary (v7, frozen 2026-10-01)

Defensive Level 1 (ELEVATED): ≥2 of 4 legs fire on the day (raw, no latch): B OAS +50 bps over 20 sessions · SPHB/SPLV < 50-day SMA · T10Y2Y ≥ trailing 12-month min + 0.25 · VIX > 1.5 × 20-day SMA. Defensive Level 2 (MAXIMUM): B OAS level ≥ 450 bps (supersedes Level 1). Expected frequencies (evidence base, 2000–2026): Level 1 ~3 triggers/yr with a ~35–37% 6-month drawdown rate; Level 2 ~1.6/yr, ~35%.
PHASE 2A: Level 1 or Level 2 hot AND ≥1 of 2 confirmation legs latched hot (B OAS +65 bps / 20 sessions; CCC OAS +200 bps / 20 sessions; 15-session latches). PHASE 2B: 2A plus the automated Fed-rollover flag above.
Exits from 2A/2B (first wins): HY < 50d SMA (≈309 bps) ×3 consecutive sessions · SPHB/SPLV > 50d SMA (1.959) ×3 consecutive sessions · time stop 23/60 sessions. Cooldown after any exit: 0/10 sessions remaining.
TAIL_OVERRIDE: 10Y weekly close > 5.40% AND Brent spot > $120 — supersedes all states and suppresses 2A/2B rotation instructions while active; exits when either condition lapses for 5 sessions.
Removed vs v5: the G3 yield-exhaustion gate (DGS10 ≥ 5.10% or z ≥ +3.0; 22.7% 6-month drawdown rate, worse than random) and the pre-peak-window metric (replaced by forward drawdown from the signal day). Full evidence: backtest v7.
Fail-safe data handling: a series that fails to load or is stale > 3 sessions never counts as triggered and is listed under data quality / errors. Every data card carries an as-of stamp (its underlying series' last observation): amber when > 2 sessions old, red when > 5.