fix(notebook,#18124): research_commodity_term_structure -- real 21-futures universe replaces np.random roll simulation

[jsboige]

Grain: MED/notebook-python -- lane myia-po-2026:CoursIA -- prev: #18165

Summary

SOTA repair of Research-Executor/research_commodity_term_structure.ipynb
(re-audit of #18124 against current state found it still degraded: the
central cell was np.random.normal(0, 0.15, 21) dressed up as "Simulated
Roll Returns", the ending printed the author's numbers as results).

• Real universe: 19/21 front-month commodity futures via yfinance (2
  unavailable on Yahoo, reported), 1-year realized returns measured:
  Heating Oil +92.4%, Gasoline +60.7%, Crude +46.7% ... Coffee -25.3%,
  Orange Juice -36.3%.
• The quintile long-short mechanism computed on real returns: long leg
  (top quintile) mean +66.6%, short leg -31.1%, spread +97.7%.
• Honest scope statement instead of fabricated roll numbers: roll
  returns need the distant-expiry contract, which Yahoo does not expose
  reliably -- the true term-structure screen (near/distant ratio) runs on
  QC Cloud with multi-contract data. The substitution (realized front-month
  returns) is stated in the output, not disguised.
• The author's numbers (-15.71% 5Y CAGR, 80.8% MaxDD, then +38.85% quarter
  with 3M Sharpe 10.5) are quoted as a design reference and read
  critically: that pattern is the signature of a fragile regime-dependent
  strategy, not an edge.

Validation

• Papermill: 7/7 cells, 0 errors, all execution_count set (C.2)
• C.1: no raise NotImplementedError / assert False / 1/0 (verified)
• Catalogue byte-identical to main (single .ipynb changed)
• SOTA verdict: SOTA-OK for the measurable part (real futures data);
  the roll-return screen itself is explicitly INTRINSIC to multi-contract
  data and routed to QC Cloud -- no proxy faked as roll returns

See #18124 (8 of 10 repaired by this lane; remaining: runner.ipynb single
unexecuted harness cell).

🤖 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]

Golden-Set Execution (H.7 P3)

✅ 8/8 notebooks passed (certified reproducible)

Notebook	Status	Time
2.1-Workflow-ML.ipynb	✅ SUCCESS	11.5s
2.2-Descente-de-gradient.ipynb	✅ SUCCESS	8.3s
2.3-Regression-lineaire-logistique.ipynb	✅ SUCCESS	4.3s
2.4-Arbres-Forets-Ensembles.ipynb	✅ SUCCESS	4.2s
Search-01-StateSpace.ipynb	✅ SUCCESS	3.2s
SL-1-LogicalLearning.ipynb	✅ SUCCESS	2.2s
rl_4_multi_armed_bandits.ipynb	✅ SUCCESS	19.3s
GameTheory-04c-NashExistence-Python.ipynb	✅ SUCCESS	3.2s

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

[github-actions]

No organ-duplication: no added def/class collides with another series organ API (scripts/audit/organ_api_index.yaml).

Detector: python scripts/audit/detect_organ_duplication.py --base <merge-base> --body-file <pr body>
Rationale: #16776 / #13564 (rule merged in #16778).

[github-actions]

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

[github-actions]

Notebook PR Validation: PASS

• Notebooks checked: 1
• Code cells validated: 3
• 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)

[jsboige]

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2026:CoursIA-3
pr: 18168
head: 4fbd990
complete: true
body: read
comments-reviewed: 5
reviews-reviewed: 0
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: b51a4c6dcaf600b22bf33a01d1d20d779e5aa0a216d31347456ab99c4a0b5655
diff-files: 1
diff-additions: 276
diff-deletions: 114
checks: latest-wins-green
b0: clear
scope: pass
domain: pass
verdict: READY
[/ADJOINT PREFLIGHT]

Dossier READY sur PR #18168 (lane myia-po-2026:CoursIA-3, cycle c.248, 28/09 07:52Z).

Substance : fix(notebook,#18124): research_commodity_term_structure -- real 21-futures universe replaces np.random roll simulation.

Vérif substance : +276/-114 sur 1 fichier MyIA.AI.Notebooks/Research-Executor/research_commodity_term_structure.ipynb. Grain MED/notebook-python.

Substance (item 11 crible de fond — PASS) :

• Le body énonce honnêtement : « the central cell was np.random.normal(0, 0.15, 21) dressed up as 'Simulated Roll Returns', the ending printed the author's numbers as results. »
• Le PR restaure l'exécution réelle : 21 commodités front-month via yfinance (19/21 dispo, 2 indisponibles sur Yahoo signalées dans le notebook), 1-year realized returns mesurés (Heating Oil +92.4%, Gasoline +60.7%, Crude +46.7%, ... Coffee -25.3%, Orange Juice ...).
• C'est un fix anti-SOTA-dégradé. Le PR nomme explicitement le workaround (np.random.normal maquillé en "Simulated") et ne le reproduit pas.
• Crible item 11 PASS : pas de marqueurs de dégradation persistants.

État actuel : mergeStateStatus: CLEAN, mergeable: MERGEABLE, reviewDecision: NONE, B.0 rc=0.

Vérif checks (latest-wins) : tous checks verts (PR gate success).

Grain: META/secretary -- lane myia-po-2026:CoursIA-3 -- prev: META/secretary c.247