docs(iit,#13410): densite ICT-13 (1029->1243) + ICT-15f (994->1347) -- 8 lectures chiffrees

[jsboige]

Grain: DEEP/notebook-python -- lane myia-po-2026:CoursIA -- prev: DEEP/notebook-python #16357

Livrable

Tranche densite #13410 : ICT-13-AxelrodStrategicMorphodynamics (1029 → 1243) + ICT-15f-Bridge1bis-DecoupledFamily (994 → 1347), plancher 1200, marge +43/+147. Markdown-only : +56 lignes de prose (8 cellules md inserees par ancres), 0 cellule code touchee, 0 output edite (C.2/C.3). Worktree dedie depuis origin/main 5a1989a92, commit b1f40fee5.

Les 8 lectures (chaque nombre lu dans les sorties committes)

ICT-13 (5 lectures locales au point de mesure) :

1. Duel nu (§1) : AllD=5,00 vs AllC=0,00 sur 50 coups sans bruit — la fausse evidence que le tournoi demonte ; les 4 bits de comportement (tft 1er coup = 0, grim apres C = 0, pavlov stay/shift 0/1) lus comme genetique du tournoi.
2. Gate 2 : seuil analytique (T−R)/(T−P) = 0,500, croisement numerique 0,550 — l'ecart de 0,05 lu comme « accord a 1 pas de grille » (verbatime de la sortie), pas comme une divergence de modele.
3. Gate 3, chiffre par chiffre : les 6 chutes sous bruit b=0→0,4 dans l'ordre exact tft +0,399 > gtft +0,392 > pavlov +0,376 > allc +0,369 > grim +0,287, alld −0,001 — hierarchie inversee du tournoi, direction explicite (valeur positive = perte de score).
4. Gate 5a (suite) : la trajectoire turbulente lue contre le decompte (2 visitees/1 changement en instantane et uniforme, contre 6 visitees/104 changements en turbulent).
5. Gate 5c, ligne par ligne : R1 0,0000 ; R2 etale 0,0259→0,4106 moyenne 0,2304 avec le run 0,2444 replique du 6b ; R3 0,0000 ×5 ; R4 0,0000 ×5 ; delta R3−R4 = +0,0000 — la lecture ferme la question avant le verdict ARTEFACT_DE_GENERATEUR.

ICT-15f (3 lectures) :
6. Decouplage constructif : rho(sigma,width) = +0,000 (critere |.|<0,2 verifie) MAIS fuite rho(sigma,barrier) = +0,463 ; recovery min=0,400 max=1,000 std=0,181 sur n=30 — la fuite motive la purge a 2 covariables de la section 4.
7. Deterministe vs stochastique : 0,800 vs 0,333 sur le meme substrat (sigma=6,0, w=1,000, barrier=0,750), facteur 2,4 — ce que le choix de mesure revele.
8. Controle nul chiffre par chiffre : rho(sw)=+1,000, rho(wr)=+0,714, partial(sigma|w)=+0,000, fronce reproduite True — croise avec le regime decouple (partials 0,128→0,173 sous null p95 0,324→0,385 sur 4 graines) : deux regimes de couplage opposes, meme absence de pouvoir propre.

Validation

• pedagogy_density.py <2 notebooks> → « All judged notebooks meet the density floor » (1243 / 1347).
• detect_markdown_rendering.py --check → OK sur chacun (0 nouvelle violation ERROR).
• Pre-commit : gitleaks, H.3, Aucun garde ne parse la source des cellules : une cellule non compilable portant une sortie traverse les 60+ checks #13326, rendering bloc — tous Passed. Le hook a normalise au basename les 2 chemins papermill absolus herites d'ICT-13 (D:\Dev\CoursIA-13009-repair2...) — seule normalisation toleree, defect preexistant signale ici.
• Densite re-mesuree post-commit depuis la CLI officielle du depot.

Preflight

Par chemins exacts (search GitHub + liste PR open) : 0 PR open sur ICT-13-AxelrodStrategicMorphodynamics et ICT-15f-Bridge1bis-DecoupledFamily. Les autres sous-plancher du scan live sont couverts par les PRs en vol (#16432, #16433, #16434, #16435, #16437, #16424, #16427).

See #13410

🤖 Generated with Claude Code

[github-actions]

⚠️ Prose/output review needed in the notebooks this PR changed: a numeric value is not anchored, an explicit relation is contradicted, or its evidence is missing. These cases remain distinct in the JSON report; the signal is advisory, NOT a merge gate.

Scope = notebooks CHANGED in this PR, not the whole corpus. Explicit claim-check relations resolve only against named CLAIM_METRICS from the local output window and are classified SUPPORTED, CONTRADICTED, or UNPROVEN.
The markdown-claims-output-report run artifact contains the structured JSON report. See python scripts/check_markdown_claims_output.py --help for re-running locally.
Detector rationale: c.290 / c.331 / PR #11435 numeric pathology, extended with low-noise relational evidence.

[github-actions]

Notebook outputs-required (H.4 schema): PASS (every code cell carries an outputs: list)

[github-actions]

Golden-Set Execution (H.7 P3)

✅ 8/8 notebooks passed (certified reproducible)

Notebook	Status	Time
2.1-Workflow-ML.ipynb	✅ SUCCESS	3.7s
2.2-Descente-de-gradient.ipynb	✅ SUCCESS	4.3s
2.3-Regression-lineaire-logistique.ipynb	✅ SUCCESS	5.3s
2.4-Arbres-Forets-Ensembles.ipynb	✅ SUCCESS	5.0s
Search-01-StateSpace.ipynb	✅ SUCCESS	3.8s
SL-1-LogicalLearning.ipynb	✅ SUCCESS	2.8s
rl_4_multi_armed_bandits.ipynb	✅ SUCCESS	15.9s
GameTheory-04c-NashExistence-Python.ipynb	✅ SUCCESS	2.5s

Pinned lockfile: scripts/notebook_tools/golden_set.lock.txt (H.7 P3, axe A #4208)

[github-actions]

Notebook PR Validation: PASS

• Notebooks checked: 2
• Code cells validated: 29
• Result: All passed

Checks: H.1 (no errors), H.3 (execution_count), C.1 (no banned patterns)
Non-Python kernels (.NET/Lean): C.1 + errors only (execution_count advisory)
QuantConnect notebooks: C.1 + errors only (require QC Cloud for execution)

[myia-ai-01]

Approving exact head b1f40fee53121b81c825dfcaeb490a1a08a716b9 after reading the complete PR surface and independently verifying the strict markdown-only delta, density calculations, numeric anchors, current-main integration, collisions, checks, and closing semantics.

The two notebooks preserve every pre-existing cell byte-for-byte, including code, outputs, execution counts and cell metadata. The only inherited non-cell change is the disclosed normalization of two absolute Papermill paths in ICT-13 to basenames. The official density metric reproduces ICT-13 1029→1243 and ICT-15f 994→1347. All eight added readings match committed outputs, including tournament payoffs, threshold crossing, strategy drops, trajectory counts, R1–R4 table values, width/recovery correlations, deterministic-vs-stochastic ratio, null control and four-seed partial correlations.

Canonical main 5a1989a92e2185678763a06a76032b7384cd70e5 is the parent, current content checks are green, no open PR overlaps either notebook, and closingIssuesReferences is empty so #13410 remains open. Two prose interpretations are loose but non-blocking: rho=0.714 is not two-thirds variance explained, and grim ranks near the top in the clean tournament despite being called mediocre.

Content is approved. The sole current failure is the 120-minute DWELL timer, maturing at 2026-09-16T21:45:42Z. Do not push or rerun before maturity; afterward require a successful PR-gate reaggregation at this unchanged head and the separate quota signal.

[jsboige]

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2025:CoursIA-2
pr: 16444
head: b1f40fe
complete: true
body: read
comments-reviewed: 4
reviews-reviewed: 1
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: 28ba0aa10135215cfd9a99bb6920c8ae56986e0f93faee3dd78ecc69cb7a2c97
diff-files: 2
diff-additions: 58
diff-deletions: 2
checks: latest-wins-green
b0: clear
scope: pass
domain: pass
verdict: READY
[/ADJOINT PREFLIGHT]