Preserve Tier2 candidate row family evidence

docs/rehost-queue.md

@@ -14,8 +14,10 @@ This file is the short active queue for the current runtime and reverse-engineer
   The checked `rt3_105/maps` compact-dispatch corpus is now exported directly: `41` maps scanned, `38` with dispatch-strip rows, `318` nondirect rows total, and the add-building subset is only `10` grouped occurrences across `7` descriptor keys, all still missing trigger kind. The active open question is therefore which ordinary loaded rows acquire or bypass the missing trigger-kind control lane before they can reach placed-structure mutation opcodes.
 - Keep the next static Tier-2 building pass focused on the earlier seed/projection seam into `0x00412d70`, not another broad `BuildingTypes` sweep.
   The grounded owner strip is `0x004196c0 -> 0x00414490 -> 0x00416ce0 -> 0x00419230`, but the active open question is which earlier seed/projection path lets the fixed candidate-table cluster at rows `35/43/45..66` reach `0x00412d70` with nonzero `[candidate+0xba/+0xbb]` before `0x00419230` clones or renames it.
+  Preserved checked row-family detail now lives in [Tier2 candidate row families](rehost-queue/tier2-candidate-row-families-2026-04-21.md).
 
 ## Preserved Detail
 
 - [Archive snapshot](rehost-queue/archive-2026-04-19.md)
+- [Tier2 candidate row families](rehost-queue/tier2-candidate-row-families-2026-04-21.md)
 - [Progress history](history/progress-history.md)

docs/rehost-queue/README.md

@@ -4,3 +4,5 @@ This directory preserves older queue snapshots and long-form implementation note
 useful as evidence, but should not stay in the short active queue file.
 
 - `archive-2026-04-19.md`: preserved detailed queue snapshot from the pre-index cleanup.
+- `tier2-candidate-row-families-2026-04-21.md`: checked candidate-table row-family split for the
+  active Tier-2 `0x00412d70` queue head.

docs/rehost-queue/tier2-candidate-row-families-2026-04-21.md

@@ -0,0 +1,46 @@
+# Tier2 Candidate Row Families (2026-04-21)
+
+This note preserves the concrete row-family evidence behind the current Tier-2 queue head, so the
+short queue can stay focused on the owner question rather than restating export details.
+
+## Checked Export
+
+- `artifacts/exports/rt3-1.06/candidate-table-named-runs.json`
+
+The export now keeps the row-family split explicit instead of burying it only inside per-map
+samples:
+
+- `files_with_port01_11_run_at_45_55_count = 37`
+- `files_with_warehouse01_11_run_at_56_66_count = 37`
+- `port00_warehouse00_row_pair_map_counts`
+  - `35/43 -> 30`
+  - `10/18 -> 7`
+- `numbered_port_warehouse_trailer_family_map_counts`
+  - `0x00000001 -> 28`
+  - `0x00000000 -> 9`
+
+## Off-Row Carrier Maps
+
+The earlier `10/18` family is narrow and fully enumerated in the checked export:
+
+- `Crossing the Alps.gmp`
+- `East Coast, USA.gmp`
+- `Germany.gmp`
+- `Go West!.gmp`
+- `Rhodes Unfinished.gmp`
+- `State of Germany.gmp`
+- `War Effort.gmp`
+
+Everything else in the current probe-bearing corpus keeps the `35/43` `Port00` / `Warehouse00`
+pair while still preserving the fixed numbered runs at `45..55` and `56..66`.
+
+## Immediate Reading
+
+This keeps the Tier-2 frontier narrower than a generic “candidate table varies by scenario” claim:
+
+- the `Port01..11` / `Warehouse01..11` numbered families are structurally fixed across all `37`
+  probe-bearing maps
+- only the `00` pair moves, and it does so in one small `7`-map family
+- the numbered trailer split is independent of the row-pair split, so the next owner question is
+  still the earlier seed or projection seam that feeds nonzero `[candidate+0xba/+0xbb]` into
+  `0x00412d70`, not the mere existence of the stable named rows themselves