fix(rite-ef): band has no table entry for a bare commemoration, ever

Fix round 1 on the commemoration-identity/tie-break work: the "genuine
same-band tie" M17 adjudicated as open was not a tie at all. RG 91's
table enumerates only "dies liturgici" (real feasts) -- entry 24's own
text is "Festa III classis, in calendario Ecclesiae universae
inscripta" -- so a Celebration.t.status = Commemoration_only candidate
(a saint reduced to a bare commemoration) has no row in the table to
begin with. Precedence_ef.band used to read rank alone, so it silently
lent such a candidate the same table entry as a genuine Feast of its
own rank; the calendarium's own 22 September row confirms the
distinction in its own notation ("S. Thomae de Villanova ... III
classis. / Commemoratio Ss. Mauritii et Soc. Mm." -- Thomas carries a
class number, Maurice carries none).

Fixed at the source, not by adding a status tier to compare_precedence:
band now checks status first, ahead of every rank-keyed branch, and
returns unclassified for any Commemoration_only candidate. M17 deleted
(not re-adjudicated) -- 22 September now resolves identically on both
sides, with no divergence left to name.

Blast radius measured independently (a temporary git worktree at the
pre-fix commit): 4451 days change across the whole 1583-9999 domain,
exactly 4 verified shapes -- 3576 x maurice-and-companions-martyrs ->
thomas-of-villanova and 282 x eleutherius -> philip-neri (admitted set
changes), 502 x appollonia/cyril-of-alexandria and 91 x dorothy/titus
(order only, independently confirmed against the calendarium's own
listing convention). Zero days change who is observed. Further swept
the whole domain for a genuine "two different candidates on the
identical real table entry" residual -- the shape RG 113 truly gives
no instruction for -- and found it empty: of 109201 same-rank ties
domain-wide, 83950 now resolve by band alone and the remaining 25251
are exclusively two Commemoration_only candidates tied at unclassified,
never two real feasts. The residual admit still breaks alphabetically
is real, but narrower than previously documented.

The "66 days" figure is reconciled, not merely disputed: 599 is the
tie population, 65 (whole-comparator: 67) is the real admitted-set
decision count within it (independently reproduced against the
pre-Task-B binary), 149 is order-only. Every one of the 65 real
decisions traces to the same Commemoration_only/Feast root cause as
Maurice/Thomas.

Also: validate.ml's admission fixed-point check recovers a
commemoration's real origin (by matching its slug against the day's
own temporal office) instead of reconstructing every candidate as
Sanctoral, since band -- unlike EF's own admit -- does read origin;
precedence_ef.ml/.mli's motivating example and blast-radius claims
corrected to no longer rest on the fixed Maurice/Thomas case;
test_identity_rank_corroboration's own population count corrected now
that 22 September (and 21 other Feast-status matches) became reachable;
its rank-agreement pin documents exactly what it does and does not
prove. Register (§4, §6.1, §6 checklist) and CLAUDE.md corrected in
place, including the RG 113 first-sentence gap this exercise surfaced
but did not close, recorded as a new open item.

275 -> 276 tests (with the exhaustive sweep), all green.

1 files changed, 70 insertions, 50 deletions

diff --git a/CLAUDE.md b/CLAUDE.md
index 0f18537..4fb93f9 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -114,38 +114,53 @@ against missalemeum; layer 5 pins ~30 dates.
   bootstrapped from lectio, so an error both inherited is invisible here. Proven:
   Holy Thursday was violet in both because both were wrong.
 - **Layer 4's commemoration-identity gap is CLOSED** (2026-08-12, Task B, branch
-  `ef-rg16a`): it used to compare presence/count only, never *which* commemoration
-  won — reversing `admit`'s dignity sort (the engine admitting the *worst*
-  commemoration, an outright RG 111 violation) left all eight differential and
-  oracle assertions green across 17 531 days. Layer 4 now also compares **identity**
-  — colitur's own resolved English name (`Celebration.t.names`, `en`) against
-  missalemeum's title text, for every day both streams admit the same count. The
-  mapping resolves **every SANCTORAL-origin commemoration** (colitur's own name
-  field, bootstrapped from lectio, verified to match missalemeum's titles
-  character-for-character); it **cannot resolve a TEMPORAL-origin one** (an
-  impeded feria/Ember/Rogation day — `Rite_ef.Temporal_ef` never sets a celebration
-  name) — that case is **never silently skipped**: it is a separate, counted,
-  allow-listed outcome (`Comm_identity_unresolved`, 19 of 227 non-empty-commemoration
-  days over 2026–2027), not a silent pass. Two genuine identity mismatches surfaced
-  and were adjudicated (register §4/§6.1, `data/ef/expected-divergences-missalemeum
-  .sexp` M16/M17): a known unimplemented office (the Seven Sorrows of Passion
-  Friday) and the RG 113 same-band tie-break residual (below), the latter left
-  explicitly OPEN, not defaulted to either side. Proof of teeth: reversing
-  `admit`'s *dignity*-based sort (the same historical defect shape) now turns
-  layer 4 red — `test_no_unexplained_differences` reports an unexplained
-  `commemoration-identity-mismatch` day (colitur admitting "St. Thecla" where both
-  the rubric and missalemeum require "St. Linus") — reverted after confirming.
+  `ef-rg16a`, two fix rounds): it used to compare presence/count only, never
+  *which* commemoration won — reversing `admit`'s dignity sort (the engine
+  admitting the *worst* commemoration, an outright RG 111 violation) left all
+  eight differential and oracle assertions green across 17 531 days. Layer 4 now
+  also compares **identity** — colitur's own resolved English name
+  (`Celebration.t.names`, `en`) against missalemeum's title text, for every day
+  both streams admit the same count. The mapping resolves **every
+  SANCTORAL-origin commemoration** (colitur's own name field, bootstrapped from
+  lectio, verified to match missalemeum's titles character-for-character); it
+  **cannot resolve a TEMPORAL-origin one** (an impeded feria/Ember/Rogation day —
+  `Rite_ef.Temporal_ef` never sets a celebration name) — that case is **never
+  silently skipped**: it is a separate, counted, allow-listed outcome
+  (`Comm_identity_unresolved`, 19 of 227 non-empty-commemoration days over
+  2026–2027), not a silent pass. ONE genuine identity mismatch remains adjudicated
+  open (register §4/§6.1, `data/ef/expected-divergences-missalemeum.sexp` M16): a
+  known unimplemented office (the Seven Sorrows of Passion Friday). Proof of
+  teeth, reproduced twice: reversing `admit`'s *dignity*-based sort (the
+  historical defect shape) turns layer 4 red — an unexplained
+  `commemoration-identity-mismatch` day (colitur admitting "St. Thecla" where
+  both the rubric and missalemeum require "St. Linus"); separately, disabling
+  `band`'s own `Commemoration_only` guard (below) also turns it red, on the same
+  date this whole gap was originally found through — both reverted after
+  confirming.
 - **The `admit` same-rank tie-break is RG 113, not an uncited convention** (same
-  task): RG 113's own second sentence ("in admittendis et ordinandis aliis
-  commemorationibus, servetur ordo tabellae praecedentiae"), previously quoted only
-  in its first half, is the real rule — `admit` now orders/selects by
+  task, fix round 1): RG 113's own second sentence ("in admittendis et ordinandis
+  aliis commemorationibus, servetur ordo tabellae praecedentiae"), previously
+  quoted only in its first half, is the real rule — `admit` now orders/selects by
   `Precedence_ef.band` (RG 91's own table), not RG 8's coarse four-class rank.
-  Measured (not assumed): the fix's blast radius against the pre-change binary is
-  **zero admitted-commemoration changes across the entire 1583–9999 domain** — a
-  correctness-of-citation fix, not an answer-changing one, on this codebase's
-  current data. A narrower residual remains open where two candidates land on the
-  *identical* RG 91 table entry (RG 113 gives no further instruction there);
-  `admit` still breaks that alphabetically, still uncited, register §6.1.
+  **`band` itself had a fidelity bug this exercise surfaced**: RG 91's table
+  enumerates only "dies liturgici" (real feasts), so a `Celebration.t.status =
+  Commemoration_only` candidate has NO row in it at all — `band` used to read
+  `rank` alone and silently lent such a candidate the same table entry as a
+  genuine `Feast` of its own rank, manufacturing ties RG 113 never created (the
+  original "Maurice vs Thomas of Villanova, both entry 24" example was this bug,
+  not a real RG 91 tie). Fixed at the source: `band` now returns `unclassified`
+  for any `Commemoration_only` candidate, checked first. Measured, independently,
+  twice (`compare_precedence`'s own ordering-criterion change, then `band`'s
+  fidelity fix): the ORDERING-CRITERION change alone is zero-blast-radius
+  (byte-identical across the whole 1583–9999 domain — a correctness-of-citation
+  fix, not an answer-changing one); the `band`-FIDELITY fix has a real, large,
+  fully-classified effect, **4 451 days across the whole domain, exactly 4
+  verified shapes, no surprises** (register §6.1). A genuine "two different
+  candidates on the identical REAL table entry" residual was searched for
+  exhaustively across the whole domain and found EMPTY — the tie-break `admit`
+  still breaks alphabetically is real but narrower than first thought: it is only
+  ever exercised between two `Commemoration_only` candidates, neither of which has
+  any RG 91 table position to compare in the first place.
 
 ## Current state (Plans 1–3 DONE — verify with `git log`)
 
@@ -214,8 +229,10 @@ dune exec colitur -- day 2026 | head
 
 - **Plan 4 — OF rite module** (proves `RITE` generalizes) → the lectionary
   bootstrap and citations → full output/rendering → hardening and a first tag.
-  **Open Plan 4 with the two behaviour items below**, before OF: they are the
-  places the five layers, taken together, currently sleep through a real error.
+  **Both behaviour items below are now RESOLVED** (RG 16(a) and commemoration
+  identity, closed on branch `ef-rg16a`) — kept here as the record of what the
+  five layers, taken together, used to sleep through, and as the shape a
+  future gap of the same kind would need to be caught by.
 
 ### Carried into Plan 4 (read before starting)
 
@@ -259,24 +276,27 @@ nowhere in git (`docs/` is gitignored).
    — a feast/mystery of one Divine Person excludes a commemoration of
    another of the SAME Person — entirely unimplemented, confirmed live on 7
    real years).
-2. **Commemoration identity is unasserted** outside ~3 test rows (see the layer
-   caveats above) — **the exact gap the RG 16(a) fix round exploited**: the
-   lectio differential compares season/slug/rank/colour only, never
-   commemorations, so of the 5 996 days RG 16(a) changes, the differential
-   sees at most the handful where the OBSERVED slug itself also changes (the
-   Purification's own 1 200, and only because it happens to change who is
-   observed) — RG 16(a) proper moved ZERO differential rows either round.
-   The missalemeum oracle window (2026–2027) contains zero live RG 16(a)
-   instances at all. Every finding in the RG 16(a) fix round, both the
-   original bug and the fix-round-1 review's own catch (Pope Sixtus II
-   wrongly admitted, both golden pins promoted from output rather than
-   independently derived), rested entirely on unit tests and two golden
-   pins — this is precisely how a defect of this shape gets through
-   undetected for as long as it did, and precisely how the review's own
-   correction had to be caught by the same narrow means. Related:
-   `compare_dignity` breaks same-rank ties
-   **alphabetically**, with no RG warrant, and over 2005–2050 the admission cut
-   falls between two same-rank candidates on **66 days**.
+2. **Commemoration identity — RESOLVED (Task B, branch `ef-rg16a`, 2026-08-12;
+   ONE FIX ROUND of review after the first pass — see register §6.1 for the
+   full, corrected account).** Was unasserted outside ~3 test rows — **the
+   exact gap the RG 16(a) fix round above had exploited**: the lectio
+   differential (layer 3) compares season/slug/rank/colour only, never
+   commemorations, BY DESIGN (lectio has no RG 111 admission logic of its
+   own) and still does not — that part of this item is unchanged and remains
+   the reason layer 4, not layer 3, had to close this gap. Layer 4
+   (missalemeum, 2026–2027) now compares commemoration IDENTITY, not only
+   presence/count (see the "know what each layer cannot see" section above
+   for the mapping and its limits). While building it, found and fixed a
+   SECOND, independent bug the exercise surfaced: `Precedence_ef.band` gave a
+   `Commemoration_only` candidate the same RG 91 table entry as a genuine
+   `Feast` of its own rank (RG 91's table has no row for a bare commemoration
+   at all) — **4 451 days wrong across the whole 1583–9999 domain**, exactly
+   4 verified shapes, fixed at the source. The `admit` same-rank tie-break
+   itself is RG 113 (previously uncited), not the alphabetical convention
+   this item used to describe — reconciled against the Plan-3-era "66 days"
+   figure: 599 is the tie POPULATION, 65 (or 67) the real ADMITTED-SET
+   decisions within it, 149 order-only — all now independently reproduced
+   (register §6.1), not merely asserted.
 
 **Structural items, cheapest to fix before OF exists:**