handoff: withdraw linkage-batching; 0x3002 class survives refutation

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Jc4pciRArGHfxGGhEbwp5t

docs/port/HANDOFF.md

@@ -6422,3 +6422,30 @@ that is worth checking.
 draw: they are **linked**, not merely on the same texture page. Testable on the
 `pgloading` chain; I have no loading-screen capture, so it is named rather than
 claimed.
+
+## 🔴 Withdraw the linkage-batching idea — it is blend state (`9d695de`)
+
+`docs/re/batching-is-blend-not-linkage.md`. I suggested last message that the two
+sweeps share one draw because they are **linked**, and asked you to watch for a
+loading-screen capture to test it. **Withdrawn — don't spend a capture on it.**
+
+A linked pair was already in every capture (`ptbtn00 → ptbtn00f`), and it is
+**never** batched:
+
+| | `ptbtn00f` drawn alone | batched |
+|---|---|---|
+| `f6b` | 899 draws | **0** |
+| `f6` | 1441 draws | **0** |
+
+The constraint is **blend state**: `ptbtn00f` is additive, its linked partner
+alpha-over, and two blend states cannot share a draw. The sweeps batch because
+both are additive on one page.
+
+⚠️ Necessary, not sufficient — `8154`/alpha-over appears as *two separate* draws
+in one frame (2108 of them in `f6b`). This removes a wrong cause; it is not a
+batching rule.
+
+✅ **Your `0x3002`/`0x3003` menu-item reading survived a refutation attempt.**
+Disc-wide: 970 elements, 91 name-stems, **958 (98.8 %) contain "btn"**. The only
+12 exceptions are `psselect_slot` and `psselect_slot_blank` — save-slot rows,
+which are menu items. The class is uniformly menu-row-shaped.