# Review Note - LLY Deep Dive 2026-05-22 Round 4

## 1. Missing or shallow sections

- 2026-05-21 official close is still not available from Lilly/LSEG at the 2026-05-22 KST check. P1 is treated as resolved by using the latest official price basis, not by inventing a 5/21 close.
- Peer valuation was refreshed from public StockAnalysis/Fiscal.ai pages, not a paid Bloomberg/FactSet terminal. It is terminal-style, source-dated, and ratio-complete, but not a true paid terminal export.
- Drug-level FY28/FY30/FY35 sales paths are model envelopes, not company guidance or analyst consensus.
- Pipeline distribution chart remains a high-relevance investment subset, not Lilly's full official pipeline count.
- Browser QA used Chromium on the local `file://` URL. CDN Chart.js/font loads were denied by the sandbox, but the new local `chart_fallback.js` rendered all 17 canvases nonblank.

## 2. Unsupported or risky claims

- FY26-FY35 drug revenue, policy EPS impacts, scenario probabilities, and price zones are analyst-style inference.
- 12m/24m targets are model outputs, not consensus or company guidance.
- Policy stress values translate CMS/Novo facts into margin/EPS/multiple sensitivity; they must be rebuilt after any CMS/HHS/White House action.
- Kisunla ramp is based on Medicare coverage mechanics and Leqembi-like adoption logic, not terminal prescription data.
- Humira/Lipitor analogs are structural comparisons, not controlled event studies.

## 3. Exemplar / hub comparison

- Applied standards:
  - `C:\Users\robotext\.codex\memories\artifact_quality_standard_20260518.md`
  - `C:\Users\robotext\.codex\docs\artifact-quality\README.md`
  - `C:\Users\robotext\.codex\docs\artifact-quality\reference_manifest.json`
- Checked market exemplars:
  - `C:\Users\robotext\.codex\docs\market\nintendo_deep_dive_20260521\index.html`
  - `C:\Users\robotext\.codex\docs\market\healthcare_sector_deep_dive_20260522\index.html`
  - `C:\Users\robotext\.codex\docs\market\btmd_deep_dive_20260522\index.html`
  - `C:\Users\robotext\.codex\docs\market\jerrynews_lly_retatrutide_20260521\index.html`
- Reused patterns: conclusion-first hero, price dashboard, sticky TOC, chart cards, decision matrix, policy mechanism table, bottom evidence/limitations, gap-first review.
- Target stats after Round 4: 105,550 chars, 20 h2 sections, 17 canvas tags, 17 `new Chart(...)` calls, `maintainAspectRatio` count 0.

## 4. Critic Round 2 repair status

| Item | Status | Evidence |
|---|---|---|
| P1 official/latest price | Resolved with latest official basis | Lilly/LSEG still latest May 20 close $1,018.87; returns and zones recalculated from it |
| P2 drug peak-sales | Resolved | FY28/FY30/FY35 drug-level sales path table added |
| P3 policy quantification | Resolved | monthly net price, volume, gross margin, EPS and multiple sensitivity table added |
| P4 Novo comparison | Resolved | CagriSema, Wegovy HD, oral sema and LLY assets compared in one table |
| P5 Kisunla coverage/ramp | Resolved | Medicare/commercial/infusion/diagnostic/Leqembi ramp table added |
| P6 peer valuation + render QA | Resolved with caveat | public ratio pages refreshed; Chromium render QA passed via local fallback |

## 5. Score caps triggered

- Latest official price rather than 5/21 official close: evidence correctness capped below 94.
- Public ratio pages rather than paid terminal: peer valuation capped below 93.
- Full official pipeline count not refreshed: pipeline completeness capped below 91.
- No global market hub update: integration quality capped below 90.

## 6. Coverage skeleton audit

| Expected first-class section | Status | Notes |
|---|---|---|
| One-line conclusion | Present | Watch / Evidence-based Accumulate |
| Current price and basis | Present | latest official $1,018.87 with refresh limitation |
| Expected return / zones | Present | $1,168 probability-weighted value, +14.6%; $920 risk line |
| GLP-1 product cycle | Present | Mounjaro/Zepbound quarterly curve and FY35 drug path |
| Pipeline lifecycle | Present | Foundayo, retatrutide, Kisunla, oncology/Jaypirca |
| Novo competition | Present | efficacy, tolerability, price, coverage, timing table |
| Policy economics | Present | Bridge/IRA/MFN/EU reference pricing to EPS/multiple |
| Kisunla/Alzheimer ramp | Present | coverage and workflow bottleneck table |
| Valuation / DCF | Present | multiple model, DCF formula, terminal FCF basis |
| Peer valuation | Present | LLY/NVO/MRK/PFE/ABBV/AZN refreshed table |
| Strategy and risk | Present | price action matrix, KPI thresholds, bull/bear debate |
| Evidence/limitations | Present | HTML bottom section, `research.md`, `source_log.md` |

## 7. Verification

- Static check: 20 h2 sections, 17 canvas tags, 17 chart calls, index size 117,479 bytes, `maintainAspectRatio` 0.
- Browser QA: Playwright Chromium opened `file:///C:/Users/robotext/.codex/docs/market/lly_deep_dive_20260522/index.html`.
- Render result: `sections=20`, `h2=20`, `canvases=17`, `chartDefined=True`, `blankCanvases=none`, `smallCanvases=none`.
- Caveat: sandbox blocked CDN resources with `ERR_NETWORK_ACCESS_DENIED`; local `chart_fallback.js` handled rendering. Screenshot saved as `qa_round4_screenshot.png`.
- Backup created before edits: `backup_round4_20260522\`.

## 8. Self-score

| Dimension | Score | Reason |
|---|---:|---|
| Scope coverage | 95 | All P1-P6 items now have concrete tables or recalculated values |
| Evidence correctness | 91 | official Lilly/FDA/CMS/Novo sources used; 5/21 official close unavailable |
| Practical actionability | 94 | price zones, policy triggers, peak-sales envelopes and peer premium are decision-ready |
| Korean readability | 90 | dense but bottom-source structure keeps top decision path readable |
| Integration quality | 89 | local folder and manifest updated; no global hub update |
| Remaining risk disclosure | 93 | paid-terminal, official-close and model limitations visible |

Final conservative self-score: 91/100.

Verdict: ship_candidate_with_caveats. The artifact is usable for review and decision discussion, but live trading still needs a fresh official quote and paid-terminal peer refresh.