From dc4e3adc9986fcfbddf7853003136f9b5857cdd0 Mon Sep 17 00:00:00 2001
From: tegwick <bernd.worsch@gmail.com>
Date: Wed, 14 Oct 2026 02:31:02 +0200
Subject: [PATCH] Close loose ends: WP-0018 finished (T04 capture link tested,
 T05 reference-model decision), WP-0016 finished (J4 from China data), WP-0014
 blocked (owner declaration + grant v2 confirmed; waits on rein-aharness
 publisher and 2027), WP-0015 blocked (EU Pm measured, noise calibrated; AR
 data and Q4 2026 data pending); Calibration v0.6

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 14913@bnt-lap001
Assistant-Session: add7ef69-b901-4756-9a67-34cf1125d96f
---
 .../assembly-checkpoint.declaration.yaml      |  85 ++++++++
 checkpoints/2026/2026-10-13-review.md         |  11 ++
 models/Calibration.md                         |   4 +-
 models/CrossBorder.md                         |  15 ++
 models/GADM-0.5.md                            |  44 +++++
 research/ModelMatching.md                     |  10 +
 simulations/gadm02/calibrate_data.py          |   4 +-
 simulations/gadm02/calibration_targets.json   |  56 ++++++
 simulations/gadm02/cross_border.py            |   2 +-
 simulations/gadm02/gadm02.py                  |   5 +-
 simulations/gadm02/gadm03.py                  |   1 +
 simulations/gadm02/matching.py                |  53 +++++
 .../gadm02/results/calibrated_readings.md     |  16 +-
 .../gadm02/results/calibrated_starts.json     | 181 ++++++++++++------
 simulations/gadm02/results/calibration.md     |  37 +++-
 simulations/gadm02/results/cross_border.md    |  78 ++++----
 simulations/gadm02/results/exports.md         |  32 ++--
 simulations/gadm02/results/matching.md        |  56 ++++--
 simulations/gadm02/results/world_scenarios.md |  58 +++---
 simulations/gadm02/world_scenarios.py         |   5 +-
 ...MBLY-WP-0014-checkpoint-operations-2027.md |  10 +-
 ...SSEMBLY-WP-0015-first-checkpoint-review.md |  10 +-
 ...SEMBLY-WP-0016-open-economy-refinements.md |  15 +-
 .../ASSEMBLY-WP-0018-gadm05-grounding.md      |  17 +-
 24 files changed, 606 insertions(+), 199 deletions(-)
 create mode 100644 activity-definitions/assembly-checkpoint.declaration.yaml

diff --git a/activity-definitions/assembly-checkpoint.declaration.yaml b/activity-definitions/assembly-checkpoint.declaration.yaml
new file mode 100644
index 0000000..728599c
--- /dev/null
+++ b/activity-definitions/assembly-checkpoint.declaration.yaml
@@ -0,0 +1,85 @@
+# Owner declaration — assembly-checkpoint-* (ASSEMBLY-WP-0014-T01)
+#
+# Answers glas-harness GLAS-WP-0017-T02 (grant-shape confirmation) and rein-aharness
+# REINAH-WP-0003-T04 / legacy-runtime-inventory step 2. Shape copied from the FI template
+# (freedom-intelligence/activity-definitions/fi-daily-research-brief.declaration.yaml).
+# This file is assembly-economics' authority over WHAT runs and under WHICH grant. It does not
+# schedule (activity-core), pick a sandbox (glas-harness) or execute and publish (rein-aharness).
+# Human-readable version and founder approval: checkpoints/EXECUTION_DECLARATION.md (approved
+# 2026-10-13), addendum in docs/decisions/2026-10-12-checkpoint-origin-publication.md.
+declaration_version: "1"
+definition_ids:
+  - assembly-checkpoint-setup-daily      # until 2026-12-31
+  - assembly-checkpoint-monthly-2026     # until 2026-12-31
+  - assembly-checkpoint-quarterly
+  - assembly-checkpoint-annual
+owner: assembly-economics
+approved_by: Bernd Worsch
+approved_at: "2026-10-13"
+status: proposed            # -> active at cutover to the governed route (target: December monthly run)
+
+execution:
+  shape: blueprint          # operator decision 2026-10-13, hub decision 0636c577: tenant-owned blueprint
+  harness_profile_ref: none # no Glas profile unless the assessment becomes an agent loop
+  steps:
+    - deterministic: python3 checkpoints/run_checkpoint.py --kind {kind} --date {date}
+    - bounded_model_call: one llm-connect request inside run_checkpoint.py (assessment); numbers
+        not present in the fetched data are rejected; --no-llm yields a valid record
+  selection: explicit definition id only. No label or approach_hint routing after cutover.
+
+inputs:
+  kind: {type: enum, values: [light, full], source: definition}
+  date: {type: date, tz: Europe/Berlin, source: activity-core due item (context.item.date)}
+  repositories:
+    - coulomb/assembly-economics at a verified Forgejo commit (main)
+    - coulomb/web-fetcher at a verified Forgejo commit (main)
+  code: checkpoints/run_checkpoint.py and web_fetcher (standard library only; no package install)
+  network:
+    egress_hosts:   # from checkpoints/sources.json, plus the llm-connect endpoint
+      - www.census.gov
+      - ec.europa.eu
+      - fred.stlouisfed.org
+      - ghoapi.azureedge.net
+      - api.worldbank.org
+      - api.openalex.org
+  secrets: none for data sources (all public). llm-connect per its own route. The push credential
+    belongs to the publisher, never to the run.
+
+outputs:
+  artifacts:
+    - checkpoints/{yyyy}/{date}-{kind}.md
+    - checkpoints/data/{date}-{kind}.jsonl
+  out_of_window: successful no-op (JSON summary status "out-of-window"), no commit
+
+repository_grant:
+  version: "2"
+  allowed_paths:
+    - checkpoints/**
+  commit_count: {min: 0, max: 1}     # 0 for out-of-window days
+  publication:
+    remote: origin
+    ref: main
+    mode: fast-forward-only
+    rebase_retries: 1                # only if every unpublished commit stays within allowed_paths
+    force: never
+    decision: docs/decisions/2026-10-12-checkpoint-origin-publication.md
+  commit_message: "Checkpoint {date} ({kind}) — activity-core"
+
+completion:
+  event: assembly_checkpoint
+  required_detail:
+    repo: assembly-economics
+    date: inputs.date
+    kind: inputs.kind
+    path: outputs.artifacts[0]
+    pushed: true (or status out-of-window)
+  on_failure: executor_run ok=false, no assembly_checkpoint event; the next scheduled run is
+    unaffected. Source failures are not run failures (recorded as not_found / error rows).
+
+rollback:
+  pins:
+    - this declaration and the four definitions, by git revision in this repo
+  procedure: revert the checkpoint commit (records are append-only by convention, so a revert is
+    visible), or set enabled false in the definition and sync activity-core definitions.
+    Until the legacy expiry (2026-12-31) the compatibility route (rein-aharness approach
+    assembly-checkpoint, 06bd08c) remains available.
diff --git a/checkpoints/2026/2026-10-13-review.md b/checkpoints/2026/2026-10-13-review.md
index d0bcc82..ad60b8a 100644
--- a/checkpoints/2026/2026-10-13-review.md
+++ b/checkpoints/2026/2026-10-13-review.md
@@ -63,6 +63,17 @@ the mean of DE, FR, IT and ES.
 Secondary-source figures are marked. They should be checked against the
 originals at the next review.
 
+### Added 2026-10-14
+
+| Indicator | Jurisdiction | Value | Source |
+|---|---|---|---|
+| Top-10 constituents' share of the STOXX Europe 600 | EU | about 21 % (Nestlé 3.0, ASML 2.7, Novo Nordisk 2.4, LVMH 2.3, AstraZeneca 2.1, Roche 2.0, Novartis 2.0, Shell 1.9, HSBC 1.5, TotalEnergies 1.3 %) | [STOXX components list](https://www.stoxx.com/document/Bookmarks/CurrentComponents/SXXP.pdf) (undated) |
+| MERVAL concentration | AR | more than half of the index weight in energy firms; no top-firm share found | [El Cronista](https://www.cronista.com/finanzas-mercados/las-tres-mejores-y-las-tres-peores-acciones-del-sp-merval-en-2026-donde-hay-valor-todavia/), [iProfesional](https://www.iprofesional.com/finanzas/461395-crece-a-pasos-agigantados-la-brecha-entre-las-acciones-que-mas-suben-y-las-que-mas-bajan) |
+
+The MERVAL is a narrow index (about 20 stocks), so a top-firm share would not
+be comparable with the US and EU operators. AR market concentration stays a
+prior.
+
 ## Operators applied
 
 - Discovery channel Pd = leading search share.
diff --git a/models/Calibration.md b/models/Calibration.md
index 4b42314..ffc8247 100644
--- a/models/Calibration.md
+++ b/models/Calibration.md
@@ -1,6 +1,8 @@
 # Calibration of GADM-0.4 to measured data
 
-**Status:** v0.5 (2026-10-13). ASSEMBLY-WP-0013, then the 2026-10-13 review
+**Status:** v0.6 (2026-10-14): EU market concentration measured (STOXX
+Europe 600 top-10 about 21 %), China (J4) added; 50 of 50 targets in range,
+42 measured. Readings R1–R4 unchanged. Earlier: v0.5 (2026-10-13). ASSEMBLY-WP-0013, then the 2026-10-13 review
 (§ 3.1): 36 of 40 targets measured, all 40 inside their ranges after fitting.
 v0.5 adds Argentina's ecological headroom (global planetary-boundary measure),
 which fits AR's `ereg` (0.0058) instead of the default. Calibrates starting states
diff --git a/models/CrossBorder.md b/models/CrossBorder.md
index b9912c6..40634a6 100644
--- a/models/CrossBorder.md
+++ b/models/CrossBorder.md
@@ -372,3 +372,18 @@ Readings:
 
 Calibration note: runs without `d0` keep the original hook, so all earlier
 results are unchanged.
+
+## 9. J4 bloc from China data (WP-0016-T03)
+
+The J4 reference bloc now starts from its own calibrated state
+(`calibrate_data`, jurisdiction "CN", J4 levers):
+
+- leading search share 0.53 (Baidu 46.9 % in StatCounter, which undercounts
+  China's domestic market, so the range is 0.40–0.65);
+- mobile-OS HHI 0.56;
+- sentiment 0.78 (Ipsos: 83 % see more benefits than drawbacks);
+- adoption level and growth as priors.
+
+With it, J4 adopts slightly faster (K_Ω 0.43 at t 10) and keeps high sentiment
+(0.77 at t 100, against 0.28 with the US start). The H1 and H2 verdicts are
+unchanged (`results/world_scenarios.md`, `results/exports.md`).
diff --git a/models/GADM-0.5.md b/models/GADM-0.5.md
index 27f5eb7..06b52d6 100644
--- a/models/GADM-0.5.md
+++ b/models/GADM-0.5.md
@@ -356,3 +356,47 @@ future AI use shifts toward automating physical production.
 
    Decoupling holds at the task-data φ_r and becomes partial or disappears at
    the dematerialization fits.
+
+## 8. Selection weights grounded and the capture link tested (ASSEMBLY-WP-0018-T04)
+
+- **Speed.** Leader displacement has declined over decades since 2000
+  (Bessen et al.), a time scale of roughly 20 years. The adjustment speed of
+  0.05 per year matches that. Lobbying moves enactment probabilities only
+  marginally (Kang 2016), which fits the moderate capture strength. Both stay
+  as set.
+- **Capture link.** `o_pd` (base parameter, default 0) makes organism
+  selection quality fall with discovery concentration,
+  `O = Q (1 − disc)(1 − o_pd · Pd)`. In GADM-0.3+ the channel stocks are
+  hidden from the base auxiliaries, so Pd is passed through explicitly
+  (`Pd_ch`). With `o_pd` = 1, endogenous capture changes VFAC by at most
+  0.0003 (attention oligopoly, baseline, generative regime at B2–B3).
+- **Reading.** Capture of selection is not a material channel for VFAC in
+  GADM, even when captured channels deliver worse quality. Selection quality
+  σ reaches the stocks only through value flows, with a small elasticity.
+  The governance concern about captured selection is therefore not supported
+  quantitatively by this model.
+
+## 9. Decision: GADM-0.5 is not yet the reference model (ASSEMBLY-WP-0018-T05)
+
+Decided 2026-10-14. **GADM-0.4 stays the reference model** for the band
+probe, the calibration (`models/Calibration.md`), model matching and the
+cross-border world. **GADM-0.5 is the module for ecological and demand
+questions**: classes, demand split, non-rival output.
+
+Reasons:
+
+1. BR-011 is open. The central GADM-0.5 reading, genuine decoupling, holds at
+   the task-data φ_r but not at the dematerialization fits. Making GADM-0.5 the
+   reference now would build every downstream reading on an unresolved
+   parameter.
+2. The structural gains are already usable where they matter (`gadm05`,
+   `simulate_eco`, `simulate_sel`) without changing the reference results.
+
+Known limitation of the reference model, recorded here: GADM-0.4 keeps
+`e2` = 0.03. That overstates compute's ecological draw 10–20 times (§ 7.1).
+This affects the ecological headroom readings of GADM-0.4 runs at high
+K_Ω.
+
+Revisit when post-2024 values of E5 (material intensity) and L4 (task shares)
+are in. Those decide φ_r and thus BR-011, which is the trigger in the break
+register.
diff --git a/research/ModelMatching.md b/research/ModelMatching.md
index 35dce0c..90d26d4 100644
--- a/research/ModelMatching.md
+++ b/research/ModelMatching.md
@@ -228,3 +228,13 @@ Readings:
 `results/matching.md`, which is re-run at the review of each checkpoint. The
 unattended runner does not compute regimes, so the approved execution
 declaration stays unchanged.
+
+### 4.4 Observation noise from real residuals (WP-0015-T07, step 1)
+
+Residual standard deviations around the best static regime (US, B3/loop
+off): adoption 0.026, investment (L5b) 0.078, discovery 0.009. With these
+(never below the assumed noise), the filter result is unchanged: B3 for two
+quarters, B4 at 1.00 in Q3 2026, no confirmed flip (`results/matching.md`).
+The filter's sharpness therefore comes from model-structure uncertainty
+(all candidates share the GADM structure), not from underestimated
+measurement noise. The B3 → B4 check waits for Q4 2026 data.
diff --git a/simulations/gadm02/calibrate_data.py b/simulations/gadm02/calibrate_data.py
index d38658b..a4bf4f6 100644
--- a/simulations/gadm02/calibrate_data.py
+++ b/simulations/gadm02/calibrate_data.py
@@ -41,7 +41,7 @@ from jurisdictions import ARCHETYPES  # noqa: E402
 HERE = Path(__file__).resolve().parent
 OUT = HERE / "results"
 TARGETS = json.loads((HERE / "calibration_targets.json").read_text())["targets"]
-NEAREST = {"US": "J3", "EU": "J1", "DE": "J1", "AR": "J2"}
+NEAREST = {"US": "J3", "EU": "J1", "DE": "J1", "AR": "J2", "CN": "J4"}
 FALLBACK = {"DE": "EU"}           # DE uses EU targets where it has none
 CHANNEL_KEYS = ("Pm", "Pi", "Pd", "Pmem", "Plin")
 ADOPT = ("aO", "ell", "oseed")
@@ -181,7 +181,7 @@ def main() -> None:
                "| Jurisdiction | " + " | ".join(gr.BAND) + " |", "|---|" + "---|" * len(gr.BAND)]
     resid_lines = []
     resid_all = {}
-    for jur in ("US", "EU", "DE", "AR"):
+    for jur in ("US", "EU", "DE", "AR", "CN"):
         t = targets(jur)
         p, s = start_state(jur, BASE_ANCHOR)
         if "Ko_ratio_1y" not in t:
diff --git a/simulations/gadm02/calibration_targets.json b/simulations/gadm02/calibration_targets.json
index 310c8db..6ca7b49 100644
--- a/simulations/gadm02/calibration_targets.json
+++ b/simulations/gadm02/calibration_targets.json
@@ -136,6 +136,12 @@
     "hi": 0.62,
     "basis": "measured",
     "op": "Special Eurobarometer 554/557 (fieldwork 2024): 55 % expect AI to affect daily life positively, 35 % negatively; Sigma = 0.5 + (pos - neg)/200 = 0.60; different question from Pew, so the range is wide and skewed down (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "Pm": {
+    "lo": 0.17,
+    "hi": 0.3,
+    "basis": "measured",
+    "op": "top-10 constituents' share of the STOXX Europe 600, about 21 % (STOXX components list, undated; secondary search 2026-10-14); same operator family as the US (top-7 share of the S&P 500, 34 %) (checkpoints/2026/2026-10-13-review.md)"
    }
   },
   "DE": {
@@ -225,6 +231,56 @@
     "basis": "measured",
     "op": "Planetary Health Check 2025 (Stockholm Resilience Centre / PIK): 7 of 9 boundaries breached; E = share still in the safe zone = 2/9 = 0.22 (global) (checkpoints/2026/2026-10-13-review.md)"
    }
+  },
+  "CN": {
+   "Ko": {
+    "lo": 0.06,
+    "hi": 0.2,
+    "basis": "prior",
+    "op": "no all-firm share; 74.6 % of China's Fortune 500 deploy or use generative AI (Omdia 2025H1), 42 % of the population are generative-AI users (CNIC, end of 2025); range like the US"
+   },
+   "Ko_ratio_1y": {
+    "lo": 1.3,
+    "hi": 1.8,
+    "basis": "prior",
+    "op": "no firm series; user numbers roughly doubled within a year (CNIC); range above the US/EU observed growth"
+   },
+   "Pd": {
+    "lo": 0.4,
+    "hi": 0.65,
+    "basis": "measured",
+    "op": "leading search share Baidu 46.9 % (StatCounter Sept 2026); StatCounter undercounts China's domestic market (Bing 21.6 %, Yandex 15 % suggest foreign-browser bias), so the range is wide (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "Pmem": {
+    "lo": 0.46,
+    "hi": 0.66,
+    "basis": "measured",
+    "op": "mobile OS HHI (Android 67.5 %, iOS 31.9 %) = 0.56 (StatCounter Sept 2026) (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "Pi": {
+    "lo": 0.55,
+    "hi": 0.8,
+    "basis": "prior",
+    "op": "domestic cloud market led by Alibaba, Huawei, Tencent; top-3 share prior"
+   },
+   "Plin": {
+    "lo": 0.5,
+    "hi": 0.85,
+    "basis": "prior",
+    "op": "domestic lineages (DeepSeek, Qwen, others); concentration prior"
+   },
+   "Sig": {
+    "lo": 0.7,
+    "hi": 0.85,
+    "basis": "measured",
+    "op": "Ipsos (via Stanford AI Index 2025): 83 % of Chinese respondents see more benefits than drawbacks (US 39 %); Sigma = 0.5 + (benefit - drawback)/200 = 0.83, range widened downward for the different question (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "E": {
+    "lo": 0.15,
+    "hi": 0.32,
+    "basis": "measured",
+    "op": "Planetary Health Check 2025 (Stockholm Resilience Centre / PIK): 7 of 9 boundaries breached; E = share still in the safe zone = 2/9 = 0.22 (global) (checkpoints/2026/2026-10-13-review.md)"
+   }
   }
  }
 }
diff --git a/simulations/gadm02/cross_border.py b/simulations/gadm02/cross_border.py
index 7fd17ea..641191c 100644
--- a/simulations/gadm02/cross_border.py
+++ b/simulations/gadm02/cross_border.py
@@ -48,7 +48,7 @@ def base() -> dict:
     st = json.loads((OUT / "calibrated_starts.json").read_text())
     def par(j):
         return {k: v for k, v in st[j]["params"].items() if k in g.PARAMS}
-    return {j: (par(j), dict(st[j]["init"])) for j in ("US", "EU", "AR")}
+    return {j: (par(j), dict(st[j]["init"])) for j in ("US", "EU", "AR", "CN") if j in st}
 
 
 def world(cal: dict, coupled: bool = True, us_lev: dict | None = None, eu_access: dict | None = None,
diff --git a/simulations/gadm02/gadm02.py b/simulations/gadm02/gadm02.py
index 675db91..ea0eb12 100644
--- a/simulations/gadm02/gadm02.py
+++ b/simulations/gadm02/gadm02.py
@@ -46,6 +46,9 @@ PARAMS = {
     # (physical, limited by ecological headroom); 1 = GADM up to 0.4 (all output scaled by sqrt(E)).
     # xn = rate at which non-rival output, as far as it is openly inheritable, flows into commons X.
     "phi_r": 1.0, "xn": 0.0,
+    # ASSEMBLY-WP-0018-T04: organism selection quality falls with discovery concentration,
+    # O = Q (1 - disc) (1 - o_pd * Pd). 0 = GADM up to 0.5 (default).
+    "o_pd": 0.0,
     # commons (§ 7)
     "aX": 0.30, "rho0": 0.30, "rhoO": 0.40, "IX": 0.02, "dX": 0.04,
     "lX": 0.06, "lXo": 0.06,
@@ -105,7 +108,7 @@ def auxiliaries(s: dict, p: dict) -> dict:
     V = min(1.0, v_h + p["vo"] * Ko * Q)
 
     M = math.sqrt(A * dem)
-    O = Q * (1 - p["disc"])
+    O = Q * (1 - p["disc"]) * (1 - p.get("o_pd", 0.0) * s.get("Pd_ch", 0.0))
     w = [p["wm"], p["ws"], p["wc"], p["wp"], p["wo"] * Ko]
     q = [M, Q, N * (1 - P), S, O]
     sigma = sum(a * b for a, b in zip(w, q)) / sum(w)
diff --git a/simulations/gadm02/gadm03.py b/simulations/gadm02/gadm03.py
index 34945c8..ea1145a 100644
--- a/simulations/gadm02/gadm03.py
+++ b/simulations/gadm02/gadm03.py
@@ -95,6 +95,7 @@ COEF = {"Pm": "c_m", "Pi": "c_i", "Pd": "c_d", "Pmem": "c_mem", "Plin": "c_lin",
 def derivatives(s: dict, p: dict) -> tuple[dict, dict]:
     s2 = {k: v for k, v in s.items() if k not in CHANNELS}
     s2["P"] = total_p(s)
+    s2["Pd_ch"] = s["Pd"]          # discovery channel, visible to the base auxiliaries (o_pd link)
     d, a = g2.derivatives(s2, p)
     d.pop("P")
     for k, (drv, r, ctr) in _channel_terms(s, p, a).items():
diff --git a/simulations/gadm02/matching.py b/simulations/gadm02/matching.py
index b50c623..b9b8b69 100644
--- a/simulations/gadm02/matching.py
+++ b/simulations/gadm02/matching.py
@@ -193,6 +193,7 @@ def main_observed() -> None:
                   + ", ".join(f"{a} {w:.2f}" for a, w in anchors.items())
                   + f". Loop weights: off {loops['off']:.2f}, on {loops['on']:.2f}.", ""]
     lines += observed_imm_lines()
+    lines += calibrated_imm_lines()
     (OUT / "matching.md").write_text("\n".join(lines))
     print("\n".join(lines))
 
@@ -556,6 +557,58 @@ def observed_imm_lines() -> list[str]:
     return lines
 
 
+# --- WP-0015-T07: observation noise calibrated from real residuals ------------------------------
+
+def residual_noise(jur: str = "US") -> dict:
+    """Residual standard deviation per indicator around the best static regime (model error included)."""
+    adop = SERIES["US_adoption"]
+    static = score(jur, path, adop, level=cd.targets(jur)["Ko"])
+    b = best(static)
+    rb = path(jur, *b)
+    t_from = CAL["US_adoption"]
+    def mt(t):
+        return static[b]["t0"] + (t - t_from)
+    out = {}
+    for name, kind in KIND.items():
+        pts = [(CAL[name] + tau, y) for tau, y in SERIES[name]["points"]]
+        pts = [(t, y) for t, y in pts if mt(t) >= STEP]
+        if kind == "pd":
+            r = [y - observe(rb, mt(t), kind) for t, y in pts]
+            r = [x - sum(r) / len(r) for x in r]     # level offset is a measurement scale, not noise
+        else:
+            r = [math.log(y) - math.log(observe(rb, mt(t), kind)) for t, y in pts]
+            r = [x - sum(r) / len(r) for x in r]
+        n = len(r)
+        out[kind] = (sum(x * x for x in r) / max(1, n - 1)) ** 0.5
+    return {"best": b, "sd": out}
+
+
+def calibrated_imm_lines() -> list[str]:
+    global NOISE
+    info = residual_noise("US")
+    keep = dict(NOISE)
+    for kind, sd in info["sd"].items():
+        mode, _old = NOISE[kind]
+        NOISE[kind] = (mode, max(sd, _old))     # never below the sampling noise assumed so far
+    ev, _i = observed_imm("US")
+    NOISE.clear(); NOISE.update(keep)
+    lines = ["", "## US: IMM filter with observation noise calibrated from residuals (WP-0015-T07)", "",
+             f"Residual standard deviations around the best static regime ({fmt(info['best'])}), model error included:",
+             "", "| Indicator | assumed noise | residual sd | used |", "|---|---|---|---|"]
+    for kind, sd in info["sd"].items():
+        lines.append(f"| {kind} | {keep[kind][1]} | {sd:.4f} | {max(sd, keep[kind][1]):.4f} |")
+    lines += ["", "| Calendar t | B1 | B2 | B3 | B4 |", "|---|---|---|---|---|"]
+    for t, res in ev:
+        an = marginal(res, 0)
+        lines.append(f"| {t:.2f} | " + " | ".join(f"{an[(a,)]['w']:.2f}" for a in ("B1", "B2", "B3", "B4")) + " |")
+    an = marginal(ev[-1][1], 0)
+    top = max(an, key=lambda k: an[k]["w"])
+    fl = flips(ev, by=0)
+    lines += ["", f"Last quarter: {top[0]} {an[top]['w']:.2f}. Confirmed flips: "
+              + ("; ".join(f"{t:.2f}: {a[0]} → {b_[0]}" for t, a, b_ in fl) if fl else "none") + ".", ""]
+    return lines
+
+
 if __name__ == "__main__":
     if "--imm" in sys.argv:
         main_imm()
diff --git a/simulations/gadm02/results/calibrated_readings.md b/simulations/gadm02/results/calibrated_readings.md
index a34f33b..ad7e55b 100644
--- a/simulations/gadm02/results/calibrated_readings.md
+++ b/simulations/gadm02/results/calibrated_readings.md
@@ -22,14 +22,14 @@ already contains today's permissiveness).
 
 | Regime | Loop | K_Ω t5 / t25 / peak | Ȳ 0–25 | VFAC t100 | G | A | P t0 → t100 | floors broken |
 |---|---|---|---|---|---|---|---|---|
-| J1 | off | 0.25 / 0.30 / 0.31 | 0.261 | 0.289 | 0.41 | 0.41 | 0.72 → 0.82 | P |
-| J1 | on | 0.24 / 0.18 / 0.26 | 0.242 | 0.306 | 0.41 | 0.47 | 0.72 → 0.80 | P |
-| J2 | off | 0.55 / 0.52 / 0.59 | 0.318 | 0.193 | 0.33 | 0.18 | 0.72 → 0.91 | P,A |
-| J2 | on | 0.43 / 0.14 / 0.43 | 0.233 | 0.246 | 0.38 | 0.31 | 0.72 → 0.88 | P |
-| J3 | off | 0.49 / 0.49 / 0.54 | 0.308 | 0.186 | 0.32 | 0.17 | 0.72 → 0.92 | P,A |
-| J3 | on | 0.38 / 0.11 / 0.38 | 0.223 | 0.237 | 0.38 | 0.30 | 0.72 → 0.90 | P |
-| J4 | off | 0.50 / 0.51 / 0.55 | 0.313 | 0.207 | 0.33 | 0.24 | 0.72 → 0.90 | P |
-| J4 | on | 0.47 / 0.35 / 0.50 | 0.287 | 0.228 | 0.36 | 0.29 | 0.72 → 0.89 | P |
+| J1 | off | 0.25 / 0.31 / 0.32 | 0.265 | 0.288 | 0.41 | 0.41 | 0.70 → 0.82 | P |
+| J1 | on | 0.24 / 0.19 / 0.27 | 0.245 | 0.306 | 0.41 | 0.48 | 0.70 → 0.80 | P |
+| J2 | off | 0.55 / 0.53 / 0.59 | 0.320 | 0.192 | 0.33 | 0.18 | 0.70 → 0.91 | P,A |
+| J2 | on | 0.44 / 0.14 / 0.44 | 0.234 | 0.246 | 0.38 | 0.31 | 0.70 → 0.88 | P |
+| J3 | off | 0.49 / 0.50 / 0.55 | 0.310 | 0.186 | 0.32 | 0.17 | 0.70 → 0.92 | P,A |
+| J3 | on | 0.38 / 0.11 / 0.38 | 0.224 | 0.237 | 0.38 | 0.30 | 0.70 → 0.90 | P |
+| J4 | off | 0.50 / 0.51 / 0.56 | 0.315 | 0.206 | 0.33 | 0.24 | 0.70 → 0.90 | P |
+| J4 | on | 0.48 / 0.35 / 0.50 | 0.289 | 0.228 | 0.36 | 0.29 | 0.70 → 0.89 | P |
 
 ## AR (current regime ≈ J2)
 
diff --git a/simulations/gadm02/results/calibrated_starts.json b/simulations/gadm02/results/calibrated_starts.json
index 31e5961..efaac1d 100644
--- a/simulations/gadm02/results/calibrated_starts.json
+++ b/simulations/gadm02/results/calibrated_starts.json
@@ -61,42 +61,103 @@
   },
   "t_star": 1.1
  },
+ "CN": {
+  "init": {
+   "As": 0.49300303196473755,
+   "Aw": 0.49791492955158234,
+   "Cch": 0.4302546754224776,
+   "Cd": 0.3622557560708628,
+   "Cs": 0.5266049749545682,
+   "Cw": 0.5568429588015028,
+   "Dch": 0.4362407628962177,
+   "Dd": 0.4019755119474779,
+   "Ds": 0.539234768802821,
+   "Dw": 0.5438975284092785,
+   "E": 0.23483345509982823,
+   "F": 0.40252055679926135,
+   "G": 0.47241433298399665,
+   "Gs": 0.45499599771736576,
+   "Gw": 0.47822150069277375,
+   "Ko": 0.12400013263270267,
+   "Kp": 0.4906170696505522,
+   "Lg": 0.7749999999999999,
+   "N": 0.4836863755546663,
+   "Nch": 0.22729946215662053,
+   "Nd": 0.09090387668390394,
+   "Ns": 0.17049784862648207,
+   "Nw": 0.5114005378433795,
+   "Pd": 0.5253629120311295,
+   "Pi": 0.6757309041042,
+   "Plin": 0.677657906631317,
+   "Pm": 0.5005796261214728,
+   "Pmem": 0.5607016992076187,
+   "Ppol": 0.500441943438186,
+   "Q": 0.4966583186115709,
+   "S": 0.49909615249141687,
+   "Sig": 0.7749999999999999,
+   "W": 0.13319776829790653,
+   "X": 0.3884559648900623
+  },
+  "k": 2.0,
+  "params": {
+   "L": 0.2,
+   "aext": 0.58,
+   "c_d": 0.4026,
+   "c_i": 0.3543,
+   "c_lin": 0.3494,
+   "c_m": 0.0693,
+   "c_mem": 0.3759,
+   "c_pol": 0.3874,
+   "care_load": 0.03,
+   "ereg": 0.006502027921568628,
+   "leg_lag": 10.0,
+   "lin_open": 0.2,
+   "m": 0.2,
+   "r_d": 0.6,
+   "r_i": 0.45,
+   "r_lin": 0.45,
+   "r_mem": 0.45,
+   "sig_gain": 0.5,
+   "sig_ref": 0.7749999999999999
+  },
+  "t_star": 1.0
+ },
  "DE": {
   "init": {
-   "As": 0.5599288424043934,
-   "Aw": 0.5595642615483107,
-   "Cch": 0.44606561637805564,
-   "Cd": 0.34624573024760946,
-   "Cs": 0.5194055619662479,
-   "Cw": 0.563580622192313,
-   "Dch": 0.4522600176030964,
-   "Dd": 0.4023715575613355,
-   "Ds": 0.5340519790596704,
-   "Dw": 0.5412211368980316,
-   "E": 0.23486643582643155,
-   "F": 0.40376104620616343,
-   "G": 0.44808171081056014,
-   "Gs": 0.4299840929679978,
-   "Gw": 0.45411530817377127,
-   "Ko": 0.11335989592626683,
-   "Kp": 0.48192186244043217,
+   "As": 0.5598041090399595,
+   "Aw": 0.5594406283690844,
+   "Cch": 0.4461157073944171,
+   "Cd": 0.346246380664374,
+   "Cs": 0.5194182228592499,
+   "Cw": 0.5636122720939143,
+   "Dch": 0.4523514525061948,
+   "Dd": 0.40237155808739855,
+   "Ds": 0.5340519980836524,
+   "Dw": 0.5412220665920002,
+   "E": 0.23484990520131996,
+   "F": 0.40377394583544624,
+   "G": 0.45311924160733974,
+   "Gs": 0.4353894545615529,
+   "Gw": 0.4590302071996272,
+   "Ko": 0.11485564105840863,
+   "Kp": 0.48273487422759537,
    "Lg": 0.535,
-   "N": 0.46214558382237997,
+   "N": 0.46458431790460164,
    "Nch": 0.227299207505914,
    "Nd": 0.0909055888190386,
    "Ns": 0.17049683002365584,
    "Nw": 0.5114007924940861,
-   "Pd": 0.8804855464849508,
-   "Pi": 0.6619591288561137,
-   "Plin": 0.8767697856588446,
-   "Pm": 0.5012800826351705,
-   "Pmem": 0.6002862986481262,
-   "Ppol": 0.502247695174575,
-   "Q": 0.49172382234185763,
-   "S": 0.48339721687145243,
+   "Pd": 0.8804417432960704,
+   "Pi": 0.661876305600853,
+   "Plin": 0.8768161282536577,
+   "Pm": 0.23554956571681512,
+   "Pmem": 0.6002937620843378,
+   "Ppol": 0.5019583275343558,
+   "Q": 0.49259417188036014,
+   "S": 0.4863702853040337,
    "Sig": 0.535,
-   "W": 0.12902620854069197,
-   "X": 0.3780888608429845
+   "W": 0.12904547830757415,
+   "X": 0.37914891949767066
   },
   "k": 1.4142135623730951,
   "params": {
@@ -106,9 +167,9 @@
    "c_d": 0.1627,
    "c_i": 0.3161,
    "c_lin": 0.0741,
-   "c_m": 0.0679,
+   "c_m": 0.0641,
    "c_mem": 0.1358,
-   "c_pol": 0.4022,
+   "c_pol": 0.3996,
    "care_load": 0.03,
    "ell": 0.5868986283848344,
    "ereg": 0.0065505098039215694,
@@ -129,40 +190,40 @@
  },
  "EU": {
   "init": {
-   "As": 0.547253539111859,
-   "Aw": 0.548224625767804,
-   "Cch": 0.4272032249357156,
-   "Cd": 0.3657481101191295,
-   "Cs": 0.5300003504494355,
-   "Cw": 0.5574458274764564,
-   "Dch": 0.43082056226817644,
-   "Dd": 0.40185192143119564,
-   "Ds": 0.5402872251399204,
-   "Dw": 0.544443167177392,
-   "E": 0.23494814917684614,
-   "F": 0.40163610217993473,
-   "G": 0.43280316908266564,
-   "Gs": 0.41316669929349137,
-   "Gw": 0.43934985809595245,
-   "Ko": 0.09150868360327788,
-   "Kp": 0.48813726708482263,
+   "As": 0.547208952633877,
+   "Aw": 0.5481801731864364,
+   "Cch": 0.4272214655833899,
+   "Cd": 0.36574823554100594,
+   "Cs": 0.5300047637492451,
+   "Cw": 0.5574576804433082,
+   "Dch": 0.43085482224420174,
+   "Dd": 0.4018519214548069,
+   "Ds": 0.5402872270108603,
+   "Dw": 0.5444433510945222,
+   "E": 0.2349426062734158,
+   "F": 0.40164020411821666,
+   "G": 0.4378733888117984,
+   "Gs": 0.41860910680062213,
+   "Gw": 0.44429599250510354,
+   "Ko": 0.09242910832258855,
+   "Kp": 0.48858757497583744,
    "Lg": 0.535,
-   "N": 0.4765639299171411,
+   "N": 0.47805006052205135,
    "Nch": 0.22729951420364686,
    "Nd": 0.09090351693347513,
    "Ns": 0.17049805681458743,
    "Nw": 0.5114004857963531,
-   "Pd": 0.8901863182693627,
-   "Pi": 0.6607094653142659,
-   "Plin": 0.875695868800961,
-   "Pm": 0.5003873745051807,
-   "Pmem": 0.5701365716974683,
-   "Ppol": 0.5009311028707848,
-   "Q": 0.49481882396446897,
-   "S": 0.4869511365194517,
+   "Pd": 0.8901708551489159,
+   "Pi": 0.6606801143118258,
+   "Plin": 0.8757129720606179,
+   "Pm": 0.23510926188591255,
+   "Pmem": 0.5701394197070061,
+   "Ppol": 0.5008252732712836,
+   "Q": 0.4953556298212024,
+   "S": 0.488790198860286,
    "Sig": 0.535,
-   "W": 0.1531825186893347,
-   "X": 0.386286326958335
+   "W": 0.15318948181824038,
+   "X": 0.3869341435739321
   },
   "k": 1.5422108254079407,
   "params": {
@@ -172,9 +233,9 @@
    "c_d": 0.1507,
    "c_i": 0.3161,
    "c_lin": 0.0741,
-   "c_m": 0.073,
+   "c_m": 0.0718,
    "c_mem": 0.1387,
-   "c_pol": 0.4247,
+   "c_pol": 0.4215,
    "care_load": 0.03,
    "ell": 0.6400174925442954,
    "ereg": 0.006262980392156864,
diff --git a/simulations/gadm02/results/calibration.md b/simulations/gadm02/results/calibration.md
index b2f2fd1..3666e2c 100644
--- a/simulations/gadm02/results/calibration.md
+++ b/simulations/gadm02/results/calibration.md
@@ -10,19 +10,21 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.
 |---|---|---|---|---|---|---|---|---|
 | US | J3 | 2.00 | 1.2 | 0.125 | 0.478 | 0.0062 | 0.387 | 0.26 |
 | EU | J1 | 1.54 | 0.9 | 0.092 | 0.523 | 0.0063 | 0.719 | 0.60 |
-| DE | J1 | 1.41 | 1.5 | 0.113 | 0.616 | 0.0066 | 0.702 | 0.60 |
+| DE | J1 | 1.41 | 1.5 | 0.115 | 0.616 | 0.0066 | 0.702 | 0.60 |
 | AR | J2 | 2.00 | 1.1 | 0.110 | 0.387 | 0.0058 | 0.580 | 0.60 |
+| CN | J4 | 2.00 | 1.0 | 0.124 | 0.600 | 0.0065 | 0.580 | 0.60 |
 
-**Fit summary:** 40 of 40 targets inside their range; measured targets 36 of 36.
+**Fit summary:** 50 of 50 targets inside their range; measured targets 42 of 42.
 
 ## Fixed anchors aligned by growth: K_Ω level miss
 
 | Jurisdiction | B1 | B2 | B3 | B4 |
 |---|---|---|---|---|
 | US | cannot reach growth | K_Ω 0.055, z -0.9 | K_Ω 0.125, z +0.0 | K_Ω 0.220, z +1.3 |
-| EU | cannot reach growth | K_Ω 0.055, z -0.7 | K_Ω 0.125, z +0.6 | K_Ω 0.203, z +2.0 |
-| DE | cannot reach growth | K_Ω 0.075, z -0.6 | K_Ω 0.173, z +0.9 | K_Ω 0.353, z +3.7 |
+| EU | cannot reach growth | K_Ω 0.056, z -0.7 | K_Ω 0.127, z +0.6 | K_Ω 0.216, z +2.2 |
+| DE | cannot reach growth | K_Ω 0.078, z -0.6 | K_Ω 0.175, z +0.9 | K_Ω 0.362, z +3.8 |
 | AR | cannot reach growth | cannot reach growth | K_Ω 0.110, z +0.0 | K_Ω 0.162, z +0.9 |
+| CN | cannot reach growth | K_Ω 0.057, z -1.0 | K_Ω 0.124, z -0.1 | K_Ω 0.213, z +1.2 |
 
 ## Residuals at the calibrated start
 
@@ -47,31 +49,33 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.
 
 | Target | Range | Model at t* | z | basis |
 |---|---|---|---|---|
-| Ko | 0.04–0.15 | 0.092 | -0.06 | prior |
-| Ko_ratio_1y | 1.35–1.6 | 1.452 | -0.18 | measured |
+| Ko | 0.04–0.15 | 0.092 | -0.05 | prior |
+| Ko_ratio_1y | 1.35–1.6 | 1.458 | -0.14 | measured |
 | Pd | 0.84–0.94 | 0.890 | +0.00 | measured |
 | Pi | 0.56–0.76 | 0.661 | +0.01 | measured |
 | Pmem | 0.47–0.67 | 0.570 | +0.00 | measured |
 | Plin | 0.8–0.95 | 0.876 | +0.01 | measured |
-| B | 0.6–0.69 | 0.636 | -0.21 | measured |
+| B | 0.6–0.69 | 0.635 | -0.21 | measured |
 | A | 0.45–0.65 | 0.548 | -0.02 | measured |
 | E | 0.15–0.32 | 0.235 | -0.00 | measured |
 | Sig | 0.45–0.62 | 0.535 | +0.00 | measured |
+| Pm | 0.17–0.3 | 0.235 | +0.00 | measured |
 
 ### DE
 
 | Target | Range | Model at t* | z | basis |
 |---|---|---|---|---|
-| Ko | 0.05–0.18 | 0.113 | -0.03 | prior |
-| Ko_ratio_1y | 1.2–1.45 | 1.318 | -0.06 | measured |
+| Ko | 0.05–0.18 | 0.115 | -0.00 | prior |
+| Ko_ratio_1y | 1.2–1.45 | 1.322 | -0.03 | measured |
 | Pd | 0.83–0.93 | 0.880 | +0.01 | measured |
 | Pi | 0.56–0.76 | 0.662 | +0.02 | measured |
 | Pmem | 0.5–0.7 | 0.600 | +0.00 | measured |
 | Plin | 0.8–0.95 | 0.877 | +0.02 | measured |
-| B | 0.66–0.76 | 0.693 | -0.34 | measured |
+| B | 0.66–0.76 | 0.693 | -0.35 | measured |
 | A | 0.48–0.65 | 0.560 | -0.06 | measured |
 | E | 0.15–0.32 | 0.235 | -0.00 | measured |
 | Sig | 0.45–0.62 | 0.535 | +0.00 | measured |
+| Pm | 0.17–0.3 | 0.236 | +0.01 | measured |
 
 ### AR
 
@@ -85,3 +89,16 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.
 | Pmem | 0.68–0.88 | 0.780 | +0.00 | measured |
 | Ko_ratio_1y | 1.25–1.65 | 1.443 | -0.04 | prior |
 | E | 0.15–0.32 | 0.235 | -0.00 | measured |
+
+### CN
+
+| Target | Range | Model at t* | z | basis |
+|---|---|---|---|---|
+| Ko | 0.06–0.2 | 0.124 | -0.09 | prior |
+| Ko_ratio_1y | 1.3–1.8 | 1.546 | -0.02 | prior |
+| Pd | 0.4–0.65 | 0.525 | +0.00 | measured |
+| Pmem | 0.46–0.66 | 0.561 | +0.01 | measured |
+| Pi | 0.55–0.8 | 0.676 | +0.01 | prior |
+| Plin | 0.5–0.85 | 0.678 | +0.02 | prior |
+| Sig | 0.7–0.85 | 0.775 | +0.00 | measured |
+| E | 0.15–0.32 | 0.235 | -0.00 | measured |
diff --git a/simulations/gadm02/results/cross_border.md b/simulations/gadm02/results/cross_border.md
index 8afab1d..b226729 100644
--- a/simulations/gadm02/results/cross_border.md
+++ b/simulations/gadm02/results/cross_border.md
@@ -15,86 +15,86 @@ at most a ≈ 0.006. Counterfactual strong rule: a 0.15.
 | Case | Jurisdiction | P t0 → t25 → t100 | Pd / Pmem / Plin t100 | K_Ω t25 | VFAC t100 | G | A | floor |
 |---|---|---|---|---|---|---|---|---|
 | closed model, current levers | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| closed model, current levers | EU | 0.72 → 0.80 → 0.82 | 0.91 / 0.60 / 0.90 | 0.30 | 0.289 | 0.41 | 0.41 | broken |
+| closed model, current levers | EU | 0.70 → 0.77 → 0.82 | 0.91 / 0.60 / 0.90 | 0.31 | 0.288 | 0.41 | 0.41 | broken |
 | closed model, current levers | AR | 0.75 → 0.85 → 0.86 | 0.95 / 0.79 / 0.90 | 0.31 | 0.229 | 0.34 | 0.27 | broken |
 | coupled, current levers | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, current levers | EU | 0.72 → 0.80 → 0.82 | 0.88 / 0.63 / 0.91 | 0.30 | 0.281 | 0.41 | 0.42 | broken |
+| coupled, current levers | EU | 0.70 → 0.77 → 0.82 | 0.88 / 0.63 / 0.91 | 0.31 | 0.281 | 0.41 | 0.42 | broken |
 | coupled, current levers | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, DMA as observed (a fitted) | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, DMA as observed (a fitted) | EU | 0.72 → 0.80 → 0.82 | 0.88 / 0.63 / 0.91 | 0.30 | 0.281 | 0.41 | 0.42 | broken |
+| coupled, DMA as observed (a fitted) | EU | 0.70 → 0.77 → 0.82 | 0.88 / 0.63 / 0.91 | 0.31 | 0.281 | 0.41 | 0.42 | broken |
 | coupled, DMA as observed (a fitted) | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, strong access rule a 0.15 (σ 0) | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, strong access rule a 0.15 (σ 0) | EU | 0.72 → 0.66 → 0.68 | 0.14 / 0.07 / 0.91 | 0.34 | 0.333 | 0.44 | 0.39 | ok |
+| coupled, strong access rule a 0.15 (σ 0) | EU | 0.70 → 0.61 → 0.68 | 0.14 / 0.07 / 0.91 | 0.35 | 0.332 | 0.44 | 0.39 | ok |
 | coupled, strong access rule a 0.15 (σ 0) | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, strong access rule, spillover σ 0.5 | US | 0.68 → 0.70 → 0.75 | 0.68 / 0.12 / 0.90 | 0.38 | 0.279 | 0.37 | 0.31 | broken |
-| coupled, strong access rule, spillover σ 0.5 | EU | 0.72 → 0.66 → 0.68 | 0.14 / 0.07 / 0.91 | 0.34 | 0.332 | 0.44 | 0.39 | ok |
+| coupled, strong access rule, spillover σ 0.5 | EU | 0.70 → 0.61 → 0.69 | 0.14 / 0.07 / 0.91 | 0.35 | 0.331 | 0.44 | 0.39 | ok |
 | coupled, strong access rule, spillover σ 0.5 | AR | 0.75 → 0.73 → 0.76 | 0.68 / 0.12 / 0.90 | 0.33 | 0.265 | 0.36 | 0.26 | broken |
 | coupled, US moves to J1 levers | US | 0.68 → 0.58 → 0.60 | 0.86 / 0.02 / 0.28 | 0.49 | 0.379 | 0.46 | 0.47 | ok |
-| coupled, US moves to J1 levers | EU | 0.72 → 0.62 → 0.63 | 0.86 / 0.02 / 0.35 | 0.37 | 0.359 | 0.45 | 0.38 | ok |
+| coupled, US moves to J1 levers | EU | 0.70 → 0.57 → 0.64 | 0.86 / 0.02 / 0.35 | 0.38 | 0.358 | 0.45 | 0.38 | ok |
 | coupled, US moves to J1 levers | AR | 0.75 → 0.63 → 0.66 | 0.86 / 0.02 / 0.28 | 0.37 | 0.303 | 0.37 | 0.25 | ok |
 | coupled, US J1 + DMA as observed | US | 0.68 → 0.58 → 0.60 | 0.86 / 0.02 / 0.28 | 0.49 | 0.379 | 0.46 | 0.47 | ok |
-| coupled, US J1 + DMA as observed | EU | 0.72 → 0.62 → 0.63 | 0.86 / 0.02 / 0.35 | 0.37 | 0.359 | 0.45 | 0.38 | ok |
+| coupled, US J1 + DMA as observed | EU | 0.70 → 0.57 → 0.64 | 0.86 / 0.02 / 0.35 | 0.38 | 0.358 | 0.45 | 0.38 | ok |
 | coupled, US J1 + DMA as observed | AR | 0.75 → 0.63 → 0.66 | 0.86 / 0.02 / 0.28 | 0.37 | 0.303 | 0.37 | 0.25 | ok |
 | coupled, US J1 + strong access rule | US | 0.68 → 0.58 → 0.60 | 0.86 / 0.02 / 0.28 | 0.49 | 0.379 | 0.46 | 0.47 | ok |
-| coupled, US J1 + strong access rule | EU | 0.72 → 0.47 → 0.42 | 0.15 / 0.01 / 0.35 | 0.41 | 0.426 | 0.51 | 0.35 | ok |
+| coupled, US J1 + strong access rule | EU | 0.70 → 0.41 → 0.42 | 0.15 / 0.01 / 0.35 | 0.43 | 0.427 | 0.51 | 0.35 | ok |
 | coupled, US J1 + strong access rule | AR | 0.75 → 0.63 → 0.66 | 0.86 / 0.02 / 0.28 | 0.37 | 0.303 | 0.37 | 0.25 | ok |
 | coupled, EU lineage access rule a 0.15 on Plin | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU lineage access rule a 0.15 on Plin | EU | 0.72 → 0.71 → 0.73 | 0.88 / 0.63 / 0.19 | 0.33 | 0.317 | 0.43 | 0.40 | broken |
+| coupled, EU lineage access rule a 0.15 on Plin | EU | 0.70 → 0.67 → 0.73 | 0.88 / 0.63 / 0.19 | 0.34 | 0.316 | 0.43 | 0.40 | broken |
 | coupled, EU lineage access rule a 0.15 on Plin | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU lineage portability 0.5 | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU lineage portability 0.5 | EU | 0.72 → 0.80 → 0.82 | 0.88 / 0.63 / 0.92 | 0.30 | 0.284 | 0.41 | 0.42 | broken |
+| coupled, EU lineage portability 0.5 | EU | 0.70 → 0.77 → 0.83 | 0.88 / 0.63 / 0.92 | 0.31 | 0.283 | 0.41 | 0.42 | broken |
 | coupled, EU lineage portability 0.5 | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU lineage access 0.15 + portability 0.5 | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU lineage access 0.15 + portability 0.5 | EU | 0.72 → 0.74 → 0.76 | 0.88 / 0.63 / 0.56 | 0.32 | 0.311 | 0.42 | 0.41 | broken |
+| coupled, EU lineage access 0.15 + portability 0.5 | EU | 0.70 → 0.70 → 0.76 | 0.88 / 0.63 / 0.56 | 0.33 | 0.310 | 0.42 | 0.40 | broken |
 | coupled, EU lineage access 0.15 + portability 0.5 | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU cloud switching (Data Act) a 0.15 on Pi | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU cloud switching (Data Act) a 0.15 on Pi | EU | 0.72 → 0.74 → 0.77 | 0.88 / 0.63 / 0.91 | 0.31 | 0.303 | 0.42 | 0.41 | broken |
+| coupled, EU cloud switching (Data Act) a 0.15 on Pi | EU | 0.70 → 0.71 → 0.77 | 0.88 / 0.63 / 0.91 | 0.32 | 0.302 | 0.42 | 0.41 | broken |
 | coupled, EU cloud switching (Data Act) a 0.15 on Pi | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | EU | 0.72 → 0.43 → 0.35 | 0.14 / 0.07 / 0.57 | 0.41 | 0.438 | 0.53 | 0.34 | ok |
+| coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | EU | 0.70 → 0.36 → 0.35 | 0.14 / 0.07 / 0.57 | 0.43 | 0.438 | 0.53 | 0.34 | ok |
 | coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.237 | 0.34 | 0.26 | broken |
 | coupled, endogenous foreign shares, current levers | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, endogenous foreign shares, current levers | EU | 0.72 → 0.80 → 0.82 | 0.88 / 0.63 / 0.90 | 0.30 | 0.282 | 0.41 | 0.42 | broken |
+| coupled, endogenous foreign shares, current levers | EU | 0.70 → 0.77 → 0.82 | 0.88 / 0.63 / 0.90 | 0.31 | 0.281 | 0.41 | 0.42 | broken |
 | coupled, endogenous foreign shares, current levers | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU commons push (IX × 3), fixed shares | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU commons push (IX × 3), fixed shares | EU | 0.72 → 0.79 → 0.81 | 0.88 / 0.63 / 0.91 | 0.33 | 0.328 | 0.42 | 0.40 | broken |
+| coupled, EU commons push (IX × 3), fixed shares | EU | 0.70 → 0.76 → 0.81 | 0.88 / 0.63 / 0.91 | 0.34 | 0.327 | 0.42 | 0.40 | broken |
 | coupled, EU commons push (IX × 3), fixed shares | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, EU commons push (IX × 3), endogenous shares | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, EU commons push (IX × 3), endogenous shares | EU | 0.72 → 0.79 → 0.81 | 0.88 / 0.63 / 0.90 | 0.33 | 0.327 | 0.42 | 0.40 | broken |
+| coupled, EU commons push (IX × 3), endogenous shares | EU | 0.70 → 0.76 → 0.81 | 0.88 / 0.63 / 0.90 | 0.34 | 0.326 | 0.42 | 0.40 | broken |
 | coupled, EU commons push (IX × 3), endogenous shares | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, split hook (domestic segment measured), current levers | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, split hook (domestic segment measured), current levers | EU | 0.72 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.30 | 0.283 | 0.41 | 0.42 | broken |
+| coupled, split hook (domestic segment measured), current levers | EU | 0.70 → 0.76 → 0.82 | 0.88 / 0.63 / 0.90 | 0.31 | 0.282 | 0.41 | 0.42 | broken |
 | coupled, split hook (domestic segment measured), current levers | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, split hook, endogenous shares | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, split hook, endogenous shares | EU | 0.72 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.30 | 0.282 | 0.41 | 0.42 | broken |
+| coupled, split hook, endogenous shares | EU | 0.70 → 0.76 → 0.82 | 0.88 / 0.63 / 0.90 | 0.31 | 0.281 | 0.41 | 0.42 | broken |
 | coupled, split hook, endogenous shares | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, split hook, commons push, endogenous shares | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, split hook, commons push, endogenous shares | EU | 0.72 → 0.78 → 0.79 | 0.88 / 0.63 / 0.90 | 0.33 | 0.332 | 0.42 | 0.40 | broken |
+| coupled, split hook, commons push, endogenous shares | EU | 0.70 → 0.75 → 0.79 | 0.88 / 0.63 / 0.90 | 0.34 | 0.331 | 0.42 | 0.40 | broken |
 | coupled, split hook, commons push, endogenous shares | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 | coupled, split hook, lineage portability 0.5 | US | 0.68 → 0.79 → 0.82 | 0.88 / 0.63 / 0.90 | 0.36 | 0.253 | 0.36 | 0.32 | broken |
-| coupled, split hook, lineage portability 0.5 | EU | 0.68 → 0.77 → 0.80 | 0.88 / 0.63 / 0.86 | 0.31 | 0.293 | 0.41 | 0.41 | broken |
+| coupled, split hook, lineage portability 0.5 | EU | 0.66 → 0.73 → 0.80 | 0.88 / 0.63 / 0.86 | 0.32 | 0.292 | 0.41 | 0.41 | broken |
 | coupled, split hook, lineage portability 0.5 | AR | 0.75 → 0.81 → 0.83 | 0.88 / 0.63 / 0.90 | 0.32 | 0.241 | 0.35 | 0.27 | broken |
 
 ## EU domination at t 100: effect of each lever (coupled model)
 
 | Lever | EU P t100 | Δ vs. coupled current |
 |---|---|---|
-| coupled, current levers | 0.821 | +0.000 |
-| coupled, DMA as observed (a fitted) | 0.821 | -0.000 |
-| coupled, strong access rule a 0.15 (σ 0) | 0.683 | -0.138 |
-| coupled, strong access rule, spillover σ 0.5 | 0.684 | -0.137 |
-| coupled, US moves to J1 levers | 0.635 | -0.186 |
-| coupled, US J1 + DMA as observed | 0.635 | -0.186 |
-| coupled, US J1 + strong access rule | 0.423 | -0.398 |
-| coupled, EU lineage access rule a 0.15 on Plin | 0.727 | -0.094 |
-| coupled, EU lineage portability 0.5 | 0.824 | +0.004 |
-| coupled, EU lineage access 0.15 + portability 0.5 | 0.757 | -0.064 |
-| coupled, EU cloud switching (Data Act) a 0.15 on Pi | 0.766 | -0.055 |
-| coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | 0.350 | -0.471 |
-| coupled, endogenous foreign shares, current levers | 0.818 | -0.003 |
-| coupled, EU commons push (IX × 3), fixed shares | 0.807 | -0.014 |
-| coupled, EU commons push (IX × 3), endogenous shares | 0.807 | -0.014 |
-| coupled, split hook (domestic segment measured), current levers | 0.817 | -0.004 |
-| coupled, split hook, endogenous shares | 0.818 | -0.003 |
-| coupled, split hook, commons push, endogenous shares | 0.793 | -0.027 |
-| coupled, split hook, lineage portability 0.5 | 0.802 | -0.019 |
+| coupled, current levers | 0.822 | +0.000 |
+| coupled, DMA as observed (a fitted) | 0.822 | -0.000 |
+| coupled, strong access rule a 0.15 (σ 0) | 0.685 | -0.137 |
+| coupled, strong access rule, spillover σ 0.5 | 0.685 | -0.137 |
+| coupled, US moves to J1 levers | 0.635 | -0.187 |
+| coupled, US J1 + DMA as observed | 0.635 | -0.187 |
+| coupled, US J1 + strong access rule | 0.417 | -0.405 |
+| coupled, EU lineage access rule a 0.15 on Plin | 0.729 | -0.093 |
+| coupled, EU lineage portability 0.5 | 0.825 | +0.003 |
+| coupled, EU lineage access 0.15 + portability 0.5 | 0.758 | -0.064 |
+| coupled, EU cloud switching (Data Act) a 0.15 on Pi | 0.767 | -0.055 |
+| coupled, EU all host rules (Pd, Pmem, Plin, Pi 0.15; portability 0.5) | 0.347 | -0.475 |
+| coupled, endogenous foreign shares, current levers | 0.819 | -0.002 |
+| coupled, EU commons push (IX × 3), fixed shares | 0.808 | -0.013 |
+| coupled, EU commons push (IX × 3), endogenous shares | 0.808 | -0.014 |
+| coupled, split hook (domestic segment measured), current levers | 0.818 | -0.004 |
+| coupled, split hook, endogenous shares | 0.819 | -0.003 |
+| coupled, split hook, commons push, endogenous shares | 0.795 | -0.027 |
+| coupled, split hook, lineage portability 0.5 | 0.803 | -0.019 |
diff --git a/simulations/gadm02/results/exports.md b/simulations/gadm02/results/exports.md
index d5b96d1..19268f9 100644
--- a/simulations/gadm02/results/exports.md
+++ b/simulations/gadm02/results/exports.md
@@ -7,19 +7,19 @@ own start. H2 holds if AR beats the EU on relative output (years 0–100) and on
 
 | Context | Loop | Export share | AR Ȳ 0–25 / 0–100 | EU Ȳ 0–25 / 0–100 | AR − EU VFAC t100 | AR P t100 | H2 |
 |---|---|---|---|---|---|---|---|
-| X1 US-led concentrated frontier | off | closed (no exports) | 0.99 / 0.81 | 1.01 / 0.85 | -0.041 | 0.83 | no |
-| X1 US-led concentrated frontier | off | today, upper bound (1.5 %) | 1.00 / 0.82 | 1.01 / 0.85 | -0.042 | 0.83 | no |
-| X1 US-led concentrated frontier | off | sector goal (4.5 %) | 1.03 / 0.85 | 1.01 / 0.85 | -0.045 | 0.83 | no |
-| X1 US-led concentrated frontier | off | export hub (10 %) | 1.07 / 0.90 | 1.02 / 0.85 | -0.052 | 0.83 | no |
-| X1 US-led concentrated frontier | on | closed (no exports) | 0.90 / 0.68 | 0.95 / 0.76 | -0.036 | 0.78 | no |
-| X1 US-led concentrated frontier | on | today, upper bound (1.5 %) | 0.91 / 0.69 | 0.95 / 0.76 | -0.037 | 0.78 | no |
-| X1 US-led concentrated frontier | on | sector goal (4.5 %) | 0.92 / 0.70 | 0.95 / 0.76 | -0.038 | 0.78 | no |
-| X1 US-led concentrated frontier | on | export hub (10 %) | 0.93 / 0.71 | 0.95 / 0.76 | -0.040 | 0.78 | no |
-| X2 open-lineage world | off | closed (no exports) | 1.06 / 0.93 | 1.09 / 1.01 | -0.055 | 0.66 | no |
-| X2 open-lineage world | off | today, upper bound (1.5 %) | 1.07 / 0.94 | 1.09 / 1.01 | -0.057 | 0.66 | no |
-| X2 open-lineage world | off | sector goal (4.5 %) | 1.10 / 0.96 | 1.09 / 1.01 | -0.061 | 0.66 | no |
-| X2 open-lineage world | off | export hub (10 %) | 1.14 / 1.00 | 1.09 / 1.01 | -0.071 | 0.66 | no |
-| X2 open-lineage world | on | closed (no exports) | 0.99 / 0.84 | 1.05 / 0.96 | -0.048 | 0.62 | no |
-| X2 open-lineage world | on | today, upper bound (1.5 %) | 1.00 / 0.85 | 1.05 / 0.96 | -0.049 | 0.62 | no |
-| X2 open-lineage world | on | sector goal (4.5 %) | 1.01 / 0.86 | 1.05 / 0.96 | -0.051 | 0.62 | no |
-| X2 open-lineage world | on | export hub (10 %) | 1.03 / 0.87 | 1.05 / 0.96 | -0.055 | 0.62 | no |
+| X1 US-led concentrated frontier | off | closed (no exports) | 0.99 / 0.81 | 1.02 / 0.86 | -0.040 | 0.83 | no |
+| X1 US-led concentrated frontier | off | today, upper bound (1.5 %) | 1.00 / 0.82 | 1.02 / 0.86 | -0.041 | 0.83 | no |
+| X1 US-led concentrated frontier | off | sector goal (4.5 %) | 1.03 / 0.85 | 1.02 / 0.86 | -0.044 | 0.83 | no |
+| X1 US-led concentrated frontier | off | export hub (10 %) | 1.07 / 0.90 | 1.03 / 0.86 | -0.051 | 0.83 | no |
+| X1 US-led concentrated frontier | on | closed (no exports) | 0.90 / 0.68 | 0.96 / 0.76 | -0.036 | 0.78 | no |
+| X1 US-led concentrated frontier | on | today, upper bound (1.5 %) | 0.91 / 0.68 | 0.96 / 0.76 | -0.036 | 0.78 | no |
+| X1 US-led concentrated frontier | on | sector goal (4.5 %) | 0.92 / 0.69 | 0.96 / 0.76 | -0.037 | 0.78 | no |
+| X1 US-led concentrated frontier | on | export hub (10 %) | 0.93 / 0.70 | 0.96 / 0.76 | -0.039 | 0.78 | no |
+| X2 open-lineage world | off | closed (no exports) | 1.06 / 0.93 | 1.10 / 1.03 | -0.055 | 0.66 | no |
+| X2 open-lineage world | off | today, upper bound (1.5 %) | 1.07 / 0.94 | 1.10 / 1.03 | -0.057 | 0.66 | no |
+| X2 open-lineage world | off | sector goal (4.5 %) | 1.10 / 0.96 | 1.10 / 1.03 | -0.061 | 0.66 | no |
+| X2 open-lineage world | off | export hub (10 %) | 1.14 / 1.00 | 1.10 / 1.03 | -0.071 | 0.66 | no |
+| X2 open-lineage world | on | closed (no exports) | 0.99 / 0.84 | 1.07 / 0.98 | -0.049 | 0.62 | no |
+| X2 open-lineage world | on | today, upper bound (1.5 %) | 1.00 / 0.85 | 1.07 / 0.98 | -0.050 | 0.62 | no |
+| X2 open-lineage world | on | sector goal (4.5 %) | 1.01 / 0.86 | 1.07 / 0.98 | -0.052 | 0.62 | no |
+| X2 open-lineage world | on | export hub (10 %) | 1.03 / 0.87 | 1.07 / 0.98 | -0.055 | 0.62 | no |
diff --git a/simulations/gadm02/results/matching.md b/simulations/gadm02/results/matching.md
index 4c23e5e..44df227 100644
--- a/simulations/gadm02/results/matching.md
+++ b/simulations/gadm02/results/matching.md
@@ -29,16 +29,16 @@ Series: Eurostat isoc_eb_ai, EU27_2020, reference years 2023-2025 (2021: 7.65, b
 
 | Candidate | log-likelihood | weight | t0 | K_Ω at window end |
 |---|---|---|---|---|
-| B3/loop on | -2.1 | 0.501 | 0.45 | 0.193 |
-| B3/loop off | -2.2 | 0.458 | 0.50 | 0.198 |
-| B2/loop off | -5.6 | 0.016 | 0.00 | 0.098 |
-| B2/loop on | -5.7 | 0.015 | 0.00 | 0.098 |
-| B4/loop on | -6.4 | 0.007 | 0.90 | 0.290 |
-| B4/loop off | -7.5 | 0.002 | 1.00 | 0.308 |
-| B1/loop off | -58.0 | 0.000 | 0.00 | 0.058 |
-| B1/loop on | -58.0 | 0.000 | 0.00 | 0.058 |
+| B3/loop on | -2.3 | 0.484 | 0.45 | 0.196 |
+| B3/loop off | -2.3 | 0.452 | 0.50 | 0.201 |
+| B2/loop off | -5.0 | 0.031 | 0.00 | 0.099 |
+| B2/loop on | -5.1 | 0.028 | 0.00 | 0.099 |
+| B4/loop on | -7.1 | 0.004 | 0.95 | 0.302 |
+| B4/loop off | -8.1 | 0.001 | 1.00 | 0.316 |
+| B1/loop off | -56.7 | 0.000 | 0.00 | 0.058 |
+| B1/loop on | -56.7 | 0.000 | 0.00 | 0.058 |
 
-**Best:** B3/loop on (weight 0.50). Anchor weights: B1 0.00, B2 0.03, B3 0.96, B4 0.01. Loop weights: off 0.48, on 0.52.
+**Best:** B3/loop on (weight 0.48). Anchor weights: B1 0.00, B2 0.06, B3 0.94, B4 0.01. Loop weights: off 0.48, on 0.52.
 
 ## DE
 
@@ -46,16 +46,16 @@ Series: Eurostat isoc_eb_ai, DE, reference years 2023-2025 (2021: 10.56, exclude
 
 | Candidate | log-likelihood | weight | t0 | K_Ω at window end |
 |---|---|---|---|---|
-| B3/loop on | -1.2 | 0.295 | 0.65 | 0.211 |
-| B3/loop off | -1.3 | 0.263 | 0.65 | 0.214 |
-| B2/loop off | -1.4 | 0.225 | 0.00 | 0.101 |
-| B2/loop on | -1.5 | 0.216 | 0.00 | 0.101 |
-| B4/loop on | -6.7 | 0.001 | 1.25 | 0.344 |
-| B4/loop off | -8.3 | 0.000 | 1.35 | 0.370 |
-| B1/loop off | -43.4 | 0.000 | 0.00 | 0.059 |
-| B1/loop on | -43.5 | 0.000 | 0.00 | 0.059 |
+| B2/loop off | -1.1 | 0.277 | 0.00 | 0.102 |
+| B2/loop on | -1.2 | 0.267 | 0.00 | 0.102 |
+| B3/loop on | -1.2 | 0.249 | 0.65 | 0.215 |
+| B3/loop off | -1.4 | 0.205 | 0.70 | 0.221 |
+| B4/loop on | -7.1 | 0.001 | 1.25 | 0.352 |
+| B4/loop off | -9.3 | 0.000 | 1.40 | 0.385 |
+| B1/loop off | -42.3 | 0.000 | 0.00 | 0.060 |
+| B1/loop on | -42.3 | 0.000 | 0.00 | 0.060 |
 
-**Best:** B3/loop on (weight 0.29). Anchor weights: B1 0.00, B2 0.44, B3 0.56, B4 0.00. Loop weights: off 0.49, on 0.51.
+**Best:** B2/loop off (weight 0.28). Anchor weights: B1 0.00, B2 0.54, B3 0.45, B4 0.00. Loop weights: off 0.48, on 0.52.
 
 ## US: IMM regime filter on the observed series (WP-0017-T05)
 
@@ -80,3 +80,23 @@ Robustness of the last evaluation (anchor weights):
 | without investment | 0.00 | 0.00 | 0.31 | 0.69 |
 
 **Last evaluation (US):** anchor B4 with weight 1.00; previous quarter B3 1.00. A flip needs the new anchor at ≥ 0.75 in two consecutive quarters; a change seen in one quarter is a **flip candidate** to check at the next review. Confirmed flips: none.
+
+
+## US: IMM filter with observation noise calibrated from residuals (WP-0015-T07)
+
+Residual standard deviations around the best static regime (B3/loop off), model error included:
+
+| Indicator | assumed noise | residual sd | used |
+|---|---|---|---|
+| adopt | 0.03 | 0.0258 | 0.0300 |
+| invest | 0.05 | 0.0784 | 0.0784 |
+| pd | 0.01 | 0.0086 | 0.0100 |
+
+| Calendar t | B1 | B2 | B3 | B4 |
+|---|---|---|---|---|
+| 0.32 | 0.00 | 0.38 | 0.40 | 0.21 |
+| 0.57 | 0.00 | 0.00 | 1.00 | 0.00 |
+| 0.82 | 0.00 | 0.00 | 1.00 | 0.00 |
+| 1.07 | 0.00 | 0.00 | 0.00 | 1.00 |
+
+Last quarter: B4 1.00. Confirmed flips: none.
diff --git a/simulations/gadm02/results/world_scenarios.md b/simulations/gadm02/results/world_scenarios.md
index 661e274..c88843f 100644
--- a/simulations/gadm02/results/world_scenarios.md
+++ b/simulations/gadm02/results/world_scenarios.md
@@ -10,65 +10,65 @@ output levels are normalized model units and not comparable across jurisdictions
 | Loop | Jurisdiction | K_Ω t10 | Ȳ 0–25 / 0–100 | VFAC t100 | P t100 | G | A | Σ end |
 |---|---|---|---|---|---|---|---|---|
 | off | US | 0.39 | 1.04 / 0.87 | 0.253 | 0.82 | 0.36 | 0.32 | 0.28 |
-| off | EU | 0.31 | 1.01 / 0.85 | 0.281 | 0.82 | 0.41 | 0.41 | 0.54 |
+| off | EU | 0.32 | 1.02 / 0.86 | 0.280 | 0.82 | 0.41 | 0.41 | 0.54 |
 | off | AR | 0.34 | 0.99 / 0.81 | 0.241 | 0.83 | 0.35 | 0.27 | 0.43 |
-| off | J4 | 0.40 | 1.07 / 0.92 | 0.285 | 0.77 | 0.38 | 0.41 | 0.28 |
+| off | J4 | 0.43 | 1.10 / 0.93 | 0.280 | 0.77 | 0.39 | 0.33 | 0.77 |
 | on | US | 0.27 | 0.90 / 0.69 | 0.293 | 0.77 | 0.39 | 0.39 | 0.16 |
-| on | EU | 0.27 | 0.95 / 0.76 | 0.310 | 0.77 | 0.42 | 0.46 | 0.44 |
-| on | AR | 0.27 | 0.90 / 0.68 | 0.274 | 0.78 | 0.37 | 0.31 | 0.31 |
-| on | J4 | 0.38 | 1.03 / 0.85 | 0.296 | 0.76 | 0.39 | 0.43 | 0.24 |
+| on | EU | 0.28 | 0.96 / 0.76 | 0.311 | 0.77 | 0.42 | 0.46 | 0.43 |
+| on | AR | 0.27 | 0.90 / 0.68 | 0.275 | 0.78 | 0.37 | 0.32 | 0.31 |
+| on | J4 | 0.39 | 1.05 / 0.85 | 0.305 | 0.73 | 0.41 | 0.36 | 0.70 |
 
 ## X2 open-lineage world
 
 | Loop | Jurisdiction | K_Ω t10 | Ȳ 0–25 / 0–100 | VFAC t100 | P t100 | G | A | Σ end |
 |---|---|---|---|---|---|---|---|---|
 | off | US | 0.45 | 1.16 / 1.10 | 0.379 | 0.60 | 0.46 | 0.47 | 0.28 |
-| off | EU | 0.34 | 1.09 / 1.01 | 0.359 | 0.63 | 0.45 | 0.38 | 0.54 |
+| off | EU | 0.35 | 1.10 / 1.03 | 0.358 | 0.64 | 0.45 | 0.38 | 0.54 |
 | off | AR | 0.37 | 1.06 / 0.93 | 0.303 | 0.66 | 0.37 | 0.25 | 0.43 |
-| off | J4 | 0.40 | 1.07 / 0.92 | 0.285 | 0.77 | 0.38 | 0.41 | 0.28 |
+| off | J4 | 0.43 | 1.10 / 0.93 | 0.280 | 0.77 | 0.39 | 0.33 | 0.77 |
 | on | US | 0.44 | 1.14 / 1.08 | 0.394 | 0.57 | 0.47 | 0.48 | 0.26 |
-| on | EU | 0.32 | 1.05 / 0.96 | 0.378 | 0.60 | 0.46 | 0.40 | 0.46 |
+| on | EU | 0.33 | 1.07 / 0.98 | 0.378 | 0.60 | 0.46 | 0.41 | 0.46 |
 | on | AR | 0.32 | 0.99 / 0.84 | 0.330 | 0.62 | 0.40 | 0.28 | 0.31 |
-| on | J4 | 0.38 | 1.03 / 0.85 | 0.296 | 0.76 | 0.39 | 0.43 | 0.24 |
+| on | J4 | 0.39 | 1.05 / 0.85 | 0.305 | 0.73 | 0.41 | 0.36 | 0.70 |
 
 ## X3 accelerating frontier
 
 | Loop | Jurisdiction | K_Ω t10 | Ȳ 0–25 / 0–100 | VFAC t100 | P t100 | G | A | Σ end |
 |---|---|---|---|---|---|---|---|---|
 | off | US | 0.73 | 1.32 / 1.20 | 0.218 | 0.84 | 0.33 | 0.23 | 0.28 |
-| off | EU | 0.39 | 1.08 / 0.95 | 0.267 | 0.84 | 0.40 | 0.38 | 0.54 |
-| off | AR | 0.41 | 1.06 / 0.89 | 0.226 | 0.85 | 0.34 | 0.24 | 0.43 |
-| off | J4 | 0.74 | 1.33 / 1.23 | 0.246 | 0.80 | 0.35 | 0.31 | 0.28 |
+| off | EU | 0.40 | 1.09 / 0.95 | 0.267 | 0.84 | 0.40 | 0.38 | 0.54 |
+| off | AR | 0.42 | 1.06 / 0.89 | 0.226 | 0.85 | 0.34 | 0.24 | 0.43 |
+| off | J4 | 0.75 | 1.34 / 1.23 | 0.238 | 0.80 | 0.34 | 0.23 | 0.77 |
 | on | US | 0.44 | 0.99 / 0.74 | 0.288 | 0.78 | 0.39 | 0.37 | 0.13 |
-| on | EU | 0.34 | 1.01 / 0.82 | 0.305 | 0.78 | 0.42 | 0.43 | 0.41 |
-| on | AR | 0.33 | 0.95 / 0.74 | 0.268 | 0.78 | 0.37 | 0.30 | 0.27 |
-| on | J4 | 0.67 | 1.24 / 1.05 | 0.274 | 0.78 | 0.37 | 0.37 | 0.19 |
+| on | EU | 0.35 | 1.01 / 0.81 | 0.306 | 0.78 | 0.42 | 0.44 | 0.41 |
+| on | AR | 0.33 | 0.95 / 0.72 | 0.269 | 0.78 | 0.37 | 0.30 | 0.28 |
+| on | J4 | 0.66 | 1.22 / 1.01 | 0.282 | 0.75 | 0.39 | 0.30 | 0.66 |
 
 ## X4 fragmented blocs
 
 | Loop | Jurisdiction | K_Ω t10 | Ȳ 0–25 / 0–100 | VFAC t100 | P t100 | G | A | Σ end |
 |---|---|---|---|---|---|---|---|---|
 | off | US | 0.39 | 1.04 / 0.87 | 0.253 | 0.82 | 0.36 | 0.32 | 0.28 |
-| off | EU | 0.30 | 1.00 / 0.84 | 0.282 | 0.83 | 0.41 | 0.42 | 0.54 |
+| off | EU | 0.31 | 1.01 / 0.85 | 0.282 | 0.83 | 0.41 | 0.42 | 0.54 |
 | off | AR | 0.34 | 0.99 / 0.80 | 0.235 | 0.85 | 0.34 | 0.27 | 0.43 |
-| off | J4 | 0.40 | 1.07 / 0.92 | 0.285 | 0.77 | 0.38 | 0.41 | 0.28 |
+| off | J4 | 0.43 | 1.10 / 0.93 | 0.280 | 0.77 | 0.39 | 0.33 | 0.77 |
 | on | US | 0.27 | 0.90 / 0.69 | 0.293 | 0.77 | 0.39 | 0.39 | 0.16 |
-| on | EU | 0.26 | 0.93 / 0.71 | 0.307 | 0.79 | 0.42 | 0.47 | 0.44 |
+| on | EU | 0.26 | 0.94 / 0.72 | 0.307 | 0.79 | 0.42 | 0.48 | 0.43 |
 | on | AR | 0.25 | 0.87 / 0.64 | 0.266 | 0.81 | 0.37 | 0.33 | 0.32 |
-| on | J4 | 0.38 | 1.03 / 0.85 | 0.296 | 0.76 | 0.39 | 0.43 | 0.24 |
+| on | J4 | 0.39 | 1.05 / 0.85 | 0.305 | 0.73 | 0.41 | 0.36 | 0.70 |
 
 ## Hypothesis re-test
 
 | Context | Loop | H1 EU adoption lags US and AR (K_Ω t10) | EU − US: ΔȲ 0–25 / ΔVFAC | H2 AR − EU: ΔȲ 0–25 / ΔȲ 0–100 / ΔVFAC | H2 holds? |
 |---|---|---|---|---|---|
-| X1 US-led concentrated frontier | off | yes (0.31 vs 0.39, 0.34) | -0.031 / +0.028 | -0.016 / -0.040 / -0.041 | no |
-| X1 US-led concentrated frontier | on | no (0.27 vs 0.27, 0.27) | +0.050 / +0.017 | -0.048 / -0.073 / -0.036 | no |
-| X2 open-lineage world | off | yes (0.34 vs 0.45, 0.37) | -0.074 / -0.020 | -0.028 / -0.083 / -0.055 | no |
-| X2 open-lineage world | on | no (0.32 vs 0.44, 0.32) | -0.093 / -0.017 | -0.060 / -0.117 / -0.048 | no |
-| X3 accelerating frontier | off | yes (0.39 vs 0.73, 0.41) | -0.235 / +0.049 | -0.024 / -0.052 / -0.041 | no |
-| X3 accelerating frontier | on | no (0.34 vs 0.44, 0.33) | +0.022 / +0.017 | -0.053 / -0.082 / -0.037 | no |
-| X4 fragmented blocs | off | yes (0.30 vs 0.39, 0.34) | -0.038 / +0.029 | -0.016 / -0.044 / -0.047 | no |
-| X4 fragmented blocs | on | no (0.26 vs 0.27, 0.25) | +0.027 / +0.013 | -0.055 / -0.073 / -0.040 | no |
+| X1 US-led concentrated frontier | off | yes (0.32 vs 0.39, 0.34) | -0.015 / +0.027 | -0.031 / -0.050 / -0.040 | no |
+| X1 US-led concentrated frontier | on | no (0.28 vs 0.27, 0.27) | +0.061 / +0.017 | -0.059 / -0.082 / -0.036 | no |
+| X2 open-lineage world | off | yes (0.35 vs 0.45, 0.37) | -0.058 / -0.021 | -0.043 / -0.099 / -0.055 | no |
+| X2 open-lineage world | on | no (0.33 vs 0.44, 0.32) | -0.076 / -0.016 | -0.077 / -0.138 / -0.049 | no |
+| X3 accelerating frontier | off | yes (0.40 vs 0.73, 0.42) | -0.225 / +0.049 | -0.032 / -0.057 / -0.041 | no |
+| X3 accelerating frontier | on | no (0.35 vs 0.44, 0.33) | +0.026 / +0.018 | -0.063 / -0.088 / -0.037 | no |
+| X4 fragmented blocs | off | yes (0.31 vs 0.39, 0.34) | -0.027 / +0.028 | -0.027 / -0.054 / -0.047 | no |
+| X4 fragmented blocs | on | no (0.26 vs 0.27, 0.25) | +0.038 / +0.013 | -0.066 / -0.083 / -0.040 | no |
 
 ## Local-in-context vs. global (EU, context X1, loop off)
 
@@ -76,5 +76,5 @@ EU simulated alone against a US path from a US-only run, compared with the full
 
 | Setting | EU P t100 local | global | EU K_Ω t25 local | global | max |ΔP| over time |
 |---|---|---|---|---|---|
-| no spillover | 0.821 | 0.821 | 0.303 | 0.303 | 0.000 |
-| strong EU access rule, spillover σ 0.5 | 0.683 | 0.684 | 0.340 | 0.340 | 0.000 |
+| no spillover | 0.822 | 0.822 | 0.312 | 0.312 | 0.000 |
+| strong EU access rule, spillover σ 0.5 | 0.685 | 0.685 | 0.354 | 0.354 | 0.000 |
diff --git a/simulations/gadm02/world_scenarios.py b/simulations/gadm02/world_scenarios.py
index 926de64..1ec269e 100644
--- a/simulations/gadm02/world_scenarios.py
+++ b/simulations/gadm02/world_scenarios.py
@@ -57,7 +57,10 @@ def build(cal: dict, ctx: dict, sent: float) -> dict:
         "US": w.make({**us_p, **ctx.get("us", {}), "sent": sent}, us_s),
         "EU": w.make({**eu_p, "sent": sent}, eu_s, home="US", f=f_eu, kappa=kap),
         "AR": w.make({**ar_p, "sent": sent}, ar_s, home="US", f=f_ar, kappa=kap),
-        "J4": w.make({**us_p, **J4, **ctx.get("j4", {}), "sent": sent, "sig_gain": 0.5, "leg_lag": 10.0}, us_s),
+        # WP-0016-T03: the J4 bloc starts from its own calibrated state (China data, calibrate_data "CN")
+        "J4": w.make({**cal["CN"][0], **ctx.get("j4", {}), "sent": sent, "sig_gain": 0.5, "leg_lag": 10.0},
+                     cal["CN"][1]) if "CN" in cal else
+              w.make({**us_p, **J4, **ctx.get("j4", {}), "sent": sent, "sig_gain": 0.5, "leg_lag": 10.0}, us_s),
     }
 
 
diff --git a/workplans/ASSEMBLY-WP-0014-checkpoint-operations-2027.md b/workplans/ASSEMBLY-WP-0014-checkpoint-operations-2027.md
index 370d080..09fd62b 100644
--- a/workplans/ASSEMBLY-WP-0014-checkpoint-operations-2027.md
+++ b/workplans/ASSEMBLY-WP-0014-checkpoint-operations-2027.md
@@ -4,13 +4,13 @@ type: workplan
 title: "Checkpoint operations: governed execution and 2027 transition"
 domain: communication
 repo: assembly-economics
-status: active
+status: blocked
 flavor: residual
 depends_on: [ASSEMBLY-WP-0010]
 owner: claude-code
 topic_slug: coulomb-social
 created: "2026-10-13"
-updated: "2026-10-13"
+updated: "2026-10-14"
 state_hub_workstream_id: "e07ebc2e-2a59-5d7a-aebe-54d8622617a4"
 ---
 
@@ -28,11 +28,13 @@ security close-down checklist row 20
 
 ```task
 id: ASSEMBLY-WP-0014-T01
-status: progress
+status: wait
 priority: high
 state_hub_task_id: "2d0f60c8-30c6-589f-ae7c-f26a01f03687"
 ```
 
+**2026-10-14, blocked on rein-aharness:** operator decisions 0636c577 (tenant-owned blueprint, no Glas profile) and c6f23433 (publisher owned by rein-aharness as a separately granted capability). Our side is done: owner declaration `activity-definitions/assembly-checkpoint.declaration.yaml` with the `repository_grant` v2 block, confirmed to glas-harness (GLAS-WP-0017-T02) and rein-aharness. Waiting for the publisher capability (REINAH-WP-0003-T04), then the canary run before 2026-12-15.
+
 **2026-10-13:** step 2 of rein-aharness's migration sequence prepared:
 `checkpoints/EXECUTION_DECLARATION.md`, **approved by the founder 2026-10-13**
 (addendum to the 2026-10-12 decision record).
@@ -94,6 +96,8 @@ priority: medium
 state_hub_task_id: "cc430881-dd0f-57c6-8092-0f884b0f8f6e"
 ```
 
+**2026-10-14:** local dry run `--kind full --no-llm` (date 2027-01-20, output deleted): 37 ok, 4 not_found (the manual light specs), 0 errors, including E5. Waits for the annual run: the X2 OpenAlex year must move to the last complete year then.
+
 The first `kind: full` run uses the 22 full specs added in WP-0010-T08.
 Before it runs, move the publication year in the X2 OpenAlex URLs to the
 last complete year (see the spec note). Run `--kind full --no-llm` locally
diff --git a/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md b/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md
index b0cc22f..a6f357b 100644
--- a/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md
+++ b/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md
@@ -4,13 +4,13 @@ type: workplan
 title: "First checkpoint review: measured targets, manual indicators, re-calibration"
 domain: communication
 repo: assembly-economics
-status: active
+status: blocked
 flavor: residual
 depends_on: [ASSEMBLY-WP-0013, ASSEMBLY-WP-0009]
 owner: claude-code
 topic_slug: coulomb-social
 created: "2026-10-13"
-updated: "2026-10-13"
+updated: "2026-10-14"
 state_hub_workstream_id: "74d02cb9-e4f3-5943-8c33-6db81c8b5acf"
 ---
 
@@ -143,6 +143,8 @@ priority: medium
 state_hub_task_id: "8cce2578-9046-502f-b4a6-bcab272accab"
 ```
 
+**2026-10-14:** EU market concentration measured (STOXX Europe 600 top-10 about 21 %), re-calibrated (Calibration v0.6: 50/50 targets, 42 measured); readings R1–R4 unchanged; cross-border and world results re-run. **Blocked:** AR market concentration (the MERVAL is too narrow for a comparable operator) and AR adoption growth (needs a second Google/IDB or INDEC data point).
+
 Waits for data that does not exist yet as of 2026-10-13.
 
 - **AR adoption growth:** the second data point of the Google/IDB study, or
@@ -161,11 +163,13 @@ R1–R4 change. Also verify the secondary-source figures marked in
 
 ```task
 id: ASSEMBLY-WP-0015-T07
-status: todo
+status: wait
 priority: high
 state_hub_task_id: "73b358e7-d09e-56b8-a1c3-1659e70554f1"
 ```
 
+**2026-10-14:** step 1 done. Observation noise from real residuals: adoption 0.026, investment 0.078, discovery 0.009. The filter result is unchanged (B4 1.00 in Q3 2026, no confirmed flip), so its sharpness comes from model-structure uncertainty, not from measurement noise. **Blocked:** step 2, the B3 → B4 check, needs Q4 2026 data (next review).
+
 From ASSEMBLY-WP-0017. The IMM filter is overconfident on observed data:
 weights jump to 1.00 within a quarter (`research/ModelMatching.md` § 4.3).
 
diff --git a/workplans/ASSEMBLY-WP-0016-open-economy-refinements.md b/workplans/ASSEMBLY-WP-0016-open-economy-refinements.md
index cb6559b..454b833 100644
--- a/workplans/ASSEMBLY-WP-0016-open-economy-refinements.md
+++ b/workplans/ASSEMBLY-WP-0016-open-economy-refinements.md
@@ -4,13 +4,13 @@ type: workplan
 title: "Open-economy refinements: lineage access rules, endogenous foreign shares, calibrated coupling"
 domain: communication
 repo: assembly-economics
-status: active
+status: finished
 flavor: residual
 depends_on: [ASSEMBLY-WP-0011]
 owner: claude-code
 topic_slug: coulomb-social
 created: "2026-10-13"
-updated: "2026-10-13"
+updated: "2026-10-14"
 state_hub_workstream_id: "94ce26e9-139a-5da5-947d-8a582495b38f"
 ---
 
@@ -62,12 +62,12 @@ point. Use indicator CB1 (operator shares over time) to calibrate the speed.
 
 ```task
 id: ASSEMBLY-WP-0016-T03
-status: progress
+status: done
 priority: medium
 state_hub_task_id: "8c6e4c04-5749-52a6-a296-e6fd79eab379"
 ```
 
-**Partly done 2026-10-13:** κ is not identified by today's data (EU one-year ratio 1.453 at κ 0, 1.470 at 0.8, observed 1.48); 0.8 does not double-count. Open: a calibrated J4 start (no tracked data for a J4 jurisdiction).
+**Partly done 2026-10-13:** κ is not identified by today's data (EU one-year ratio 1.453 at κ 0, 1.470 at 0.8, observed 1.48); 0.8 does not double-count. **Done 2026-10-14:** the J4 bloc now starts from China data (`calibrate_data` "CN": search and mobile-OS shares measured with a StatCounter caveat, sentiment Ipsos 83 % benefits > drawbacks, adoption level and growth priors); `world_scenarios.py` uses it. H1 and H2 verdicts unchanged.
 
 κ = 0.8 is a prior. Calibrate it from the lag between frontier releases and
 host adoption (indicator CB5, BTOS vs. Eurostat timing). Give the J4 bloc its
@@ -96,3 +96,10 @@ Add C6 (trade in organism output), C7 (talent and founders) and C8
 whether H2 ("Argentina may thrive") changes with an export channel. That is
 the strongest argument for J2-type thriving that the model does not yet
 represent (`models/AdoptionBand.md` § 4.5).
+
+## Close-out 2026-10-14
+
+T01–T04 are done (`models/CrossBorder.md` §§ 6–8, `world.py`, `exports.py`,
+J4 start from China data). The remaining priors (d0, λ, γ, δ, κ) are tracked
+by the coupling indicators CB1–CB6 in `checkpoints/IndicatorCatalog.md`. No
+new workplan.
diff --git a/workplans/ASSEMBLY-WP-0018-gadm05-grounding.md b/workplans/ASSEMBLY-WP-0018-gadm05-grounding.md
index 571a755..d11e3c5 100644
--- a/workplans/ASSEMBLY-WP-0018-gadm05-grounding.md
+++ b/workplans/ASSEMBLY-WP-0018-gadm05-grounding.md
@@ -4,7 +4,7 @@ type: workplan
 title: "GADM-0.5 grounding: dematerialization, ecological classes, non-rival output"
 domain: communication
 repo: assembly-economics
-status: active
+status: finished
 flavor: residual
 depends_on: [ASSEMBLY-WP-0007]
 owner: claude-code
@@ -96,11 +96,13 @@ AI-generated output, for example code on public repositories).
 
 ```task
 id: ASSEMBLY-WP-0018-T04
-status: todo
+status: done
 priority: low
 state_hub_task_id: "60940505-b8f9-55f6-b55f-aa4075fe10ff"
 ```
 
+**Done 2026-10-14:** the speed (0.05/yr) and strengths are consistent with Bessen et al. and Kang 2016; the capture link `o_pd` was tested (with Pd passed to the base auxiliaries): VFAC effect ≤ 0.0003. Capture of selection is not material (`models/GADM-0.5.md` § 8).
+
 Ground the weight adjustment speed (0.05 per year) and the capture
 strengths. Test the candidate link `O = Q (1 − Pd)` (organism selection
 quality falls with discovery concentration). Only that would make capture of
@@ -110,12 +112,21 @@ selection material for VFAC.
 
 ```task
 id: ASSEMBLY-WP-0018-T05
-status: todo
+status: done
 priority: medium
 state_hub_task_id: "3aace583-cc0b-58f7-9eb0-e241dcfddc1e"
 ```
 
+**Done 2026-10-14 (decision):** GADM-0.4 stays the reference model; GADM-0.5 is the module for ecology and demand questions. Revisit when post-2024 E5/L4 data decide BR-011 (`models/GADM-0.5.md` § 9).
+
 After T01–T03, decide whether GADM-0.5 (classes, demand split, non-rival
 output) becomes the reference model for the band probe, the calibration
 (WP-0013/0015) and the cross-border world (WP-0011/0016). Re-run the key
 readings and record what changes in the break register.
+
+## Close-out 2026-10-14
+
+T01–T05 are done. The open question, BR-011 (decoupling vs. φ_r), is tracked
+in the break register, with the trigger "post-2024 E5 and L4 data". E5 is
+wired into the full checkpoint, and L4 is collected at reviews
+(WP-0015-T01). No new workplan.