REALMARC_PATTERN_CANDIDATES_2026-07-27.md

supernatural/REALMARC_PATTERN_CANDIDATES_2026-07-27.md

The Final Ascent — pattern candidates for Ch 70-77 across all lenses

Read-only design analysis, 2026-07-27. Axis: PATTERN CANDIDATES. Produced for the realm-arc pattern

wave ordered by Josh's fourth-sitting ruling ("we need a different pattern for these mythological

realm arc leading into the final end arc; brainstorm this, deep dive across all our lenses").

Sibling axes (relief mechanism, hard-line-flag set) are referenced by NAMED MECHANISM only and are

not designed here. Nothing below is canon; every candidate is proposal-tier.

---

0. The reframe that governs every candidate (read this before the shapes)

0.1 The intensity column is a COMBAT-GAUNTLET column, not an importance column

This is not an interpretation. HL_0050's own live text says it: "'Relief' lowers the combat-gauntlet

intensity, never removes the fun," and it calls the three relief nodes "intensity valleys that frame

surrounding peaks — never action valleys." Verified verbatim from

registries/T0_Hard_Lines [ACTIVE v0.2]/Sheet1.csv @ HL_0050.

Three consequences follow immediately, and they decide the whole wave:

• "Build intensity continuously toward the fairy realm" CANNOT be executed on the intensity column

alone, because the arc's terminal node (Ch 77, the Waterfall) is canonically no-combat,

no-fail-state, player-paced dwell. On a combat-pressure column the game's most important beat is

correctly a LOW value. A pattern that reads the directive as "raise the intensity cells" produces

a string that peaks at Ch 76 and drops at the single most important chapter in the game — which is

exactly what the live string already does, so that reading changes nothing.

• The directive's real target is FELT ascent, and felt ascent is a composite. The realm charter

already separated the instruments (§1.4): POWER (ruled vril, non-negotiable), DEMAND (authored,

monotone), AWE (peaks at anchor density, never tuned as difficulty). The pattern's job is to say

which instrument carries the rise at each node and what the coarse column does around it.

• Therefore the deliverable is not a new intensity string. It is a NEW LAW with a new tooth set,

of which the intensity string is one output. Section 7 states it as five clauses.

0.2 The enum cannot express an 8-step monotone rise, by construction

harness/check_chapter_topography.py fixes INTENSITY = ("peak", "standard", "valley", "relief")

and R5 pins the relief set to exactly {CH_27, CH_52, CH_57}. So Ch 70-77 has THREE usable values for

EIGHT nodes. Any candidate claiming a smooth 8-step ascent on that column is claiming something the

column cannot hold. Every candidate below therefore names its CARRYING ORDINAL — the fine-grained

field that actually rises — and treats the enum as the coarse SHAPE around it.

The carrying ordinal already exists as a proposal: demand_ordinal in the charter's INTENSITY BLOCK

(§2.3), with the road's proposed values 1..9 at Ch 66-74, 10 at Ch 75, 11 at Ch 76, and Ch 77

EXEMPT via the empty-string sentinel (charter §1.2, and T-RLM-CURVE skips empty rows rather than

coercing them to zero). The pattern wave should adopt that column, not mint a rival.

0.3 The live state, verified

Read directly from registries/T0_Chapter_Index [ACTIVE v1.2]/T0_Chapter_Index DRAFT v1.1.csv:

Live Ch 70-77 string: standard · peak · peak · standard · standard · peak · peak · standard.

This matches the charter's §6.2 statement exactly, so the charter's reading is confirmed rather than

assumed.

0.4 What still binds (the envelope every candidate must survive)

• HL_0049, both clauses, universal and untouched by the ruling: no two consecutive nodes share

dominant_combat_mode; every 5-node window carries ≥3 distinct dominant modes. The live 70-78

mode run (vril_management · puzzle_routed · vril_management · direct · environmental · direct ·

puzzle_routed · astral · environmental) satisfies both with margin. No candidate below moves a

mode; if a later wave does, R1/R2 must be re-run.

• HL_0048's no-coasting clause survives the amendment ("Mode variety per chapter per Pillar 11 holds

across all 79 nodes"). A monotone-peak string is not a licence to coast.

• R6 / HL_0050 / HL_0051 / Josh's Ruling 1: combat=Y in EVERY chapter, never waivable. Even Ch 77

and the Epilogue carry the five-beat obligation.

• R5: relief as an enum VALUE stays locked to {27, 52, 57}. valley is not so restricted (Ch 1 and

the Epilogue both carry it), so a valley in the ascent zone is legal on the current tooth — which

matters for two candidates below.

• The ruled vril POWER curve, which the spine gates through ws.atmospheric_vril_density_modifier

seams and which no pacing pattern may contradict: plateau Ch 66-69, TROUGH Ch 70-71 (Hades low at

every site, Xibalba the floor), climb from Ch 72, high-to-anchor Ch 73-74, CEILING Ch 75.

• D-RR.5: Ch 70-74 showcase and extend existing disciplines and mint no new schools. School debuts

SCH_010..SCH_013 run Ch 66-69 and STOP. Mechanical novelty on the ascent must come from the

signature-law lane (charter §2.5/§2.6), not from new systems.

• Care: Ch 71 Xibalba LIVING tier, Ch 72 Naraka MAXIMUM tier. A pattern that raises pressure at

those nodes raises pressure on the two most care-loaded rows on the road. Escalation there must be

demand-shaped (the myth's own trials), never body-count-shaped.

• The tail is structurally fixed and is not this wave's to move: Ch 75 peak/direct, Ch 76 peak (the

double climax: Cassius named, then the father), Ch 77 standard/astral (the one sanctioned lift,

demand EXEMPT), Epilogue valley and path-invariant (HL_0045).

0.5 Two machine-layer ambiguities the ruling opens, flagged before they are inherited silently

• THE SEAM QUESTION. The amendment scopes the peaks clause to Ch 1-69. R3 currently walks a

3-window over the whole ordered node list. Does a window straddling the seam (68-69-70, 69-70-71)

count? Nobody has ruled it, and choosing silently is this repo's most-repeated defect class.

RECOMMENDED: windows must lie ENTIRELY inside Ch 1-69 to be judged by R3, and the new ascent tooth

owns Ch 70-77 exclusively. Every candidate below keeps Ch 70 at standard, which also means the

question never bites in practice — but the gate should still say which rule it applied.

• THE VALLEYS CLAUSE. The amended HL_0048 text scopes BOTH clauses with "WITHIN Ch 1-69" in its

first sentence, then names only the peaks clause as removed for Ch 70-77. Read strictly, the

valleys clause is simultaneously scoped out and not named as removed. RECOMMENDED: the Final

Ascent Law bans valleys in Ch 70-76 outright (clause 1, §7), which is strictly STRONGER than the

old clause and dissolves the ambiguity rather than arguing it.

0.6 The relief palette, named for the sibling analyst and not designed here

Seven mechanisms recur across the candidates. They are named so the pattern and relief waves cite

each other rather than duplicating:

• REGISTER ROTATION (L-R5's extracted slice) — elegiac / adversarial / moral-architectural /

seductive-hospitable / grave-vertical, one per node.

• POWER TROUGH AND RESTORATION — the ruled vril curve used as pacing, with Ch 75's ceiling-return as

the arc's largest exhale.

• BEAT-LEVEL TROUGH — the breath moves inside the chapter's 10-13 beat graph.
• SEAM SANCTUM — the inter-realm crossing as the rest point, reusing FORK C (playable navigable

transitions) and the ruled loadout-set-at-vril-sites doctrine.

• FACULTY ROTATION — the demand-KIND rotates (force / wit / release / restraint / commitment) so no

single faculty is taxed twice running.

• AWE-AS-REST — the vista and dwell beat, which the charter already forbids tuning as difficulty.
• EXPLOIT PAYOFF — the charter's per-law exploit field: an old constraint becomes a new tool, and

the mastery moment reads as relief without lowering demand.

---

1. THE TERRACED CRESCENDO — with the subito piano

Lens: presentation-craft-anim-audio (P1-C), specifically its adaptive-score half, which currently

owns R-A16 per-realm audio identity because T1_Audio_Spec is an empty stub. Secondary: L-R4

pilgrimage-traversal-and-awe.

What the lens actually teaches. Orchestral practice has known for three centuries that a literal

eight-step crescendo is unlistenable: the ear normalises, and by step five the rise is inaudible.

The working solutions are terraced dynamics (each step's FLOOR sits above the previous step's floor,

while the ceiling may be reused), and the subito piano — the sudden hush immediately before the

final fortissimo, which is the loudest instrument in the orchestra precisely because it is silence.

The Rossini crescendo adds the third device: the rise is carried by ORCHESTRATION DENSITY (more

instruments entering) rather than by volume, so the ear hears growth without hearing loudness.

Intensity string (Ch 70-77): standard · peak · peak · peak · standard · peak · peak · standard.

Carrying ordinal: demand 5,6,7,8,9,10,11,exempt — and critically, demand does NOT dip at Ch 74. The

hush is a REGISTER hush, not a demand drop; the door is quiet and still the hardest thing yet.

Relief carrier: REGISTER ROTATION plus the single authored hush.

Landing: Ch 75's ceiling-return becomes the largest dynamic leap in the game because it follows the

quietest node on the road. Ch 77 is the plagal cadence after the final chord; the Epilogue valley is

the resolution to tonic.

Strongest objection. Placing the hush at Ch 74 fights the content: Yggdrasil is a marquee dragon

fight plus a crossing-defense, and Annwn one node earlier is canonically "the last warm beat before

the crossing" with a single one-blow honour duel. The candidate is right about the DEVICE and wrong

about the SLOT. See §9 — this objection is what produces the recommendation.

Wiring note. If the arc's pattern is score-shaped, presentation-craft-anim-audio should carry the

arc-pacing question into its charter before the DR-2 freeze, not only the per-realm audio identity

it already claimed. That is a one-line charter amendment and it is time-critical for the same reason

the realm-ceiling amendment was.

---

2. THE PACT LADDER — boss-rush escalation with breath points

Lens: hades-death-run-realm-identity (proposed-new, P1-C) plus the god-of-war re-run widen

(charter §4.3: re-scope to realm adaptation — the nine realms as places, the world-tree hub,

per-realm identity).

What the lens actually teaches. Hades does not escalate by raising enemy numbers across Tartarus →

Asphodel → Elysium → Styx; it escalates by ACCUMULATING CONSTRAINT. The Pact of Punishment lets the

player stack clauses that raise demand while the build stays exactly where it was — which is,

precisely, our own ruled difficulty doctrine (the density dial moves four world-demand levers and

nothing on the enemy sheet). Its relief is not a lower-intensity level; it is the interstitial: the

chamber reward, the boon choice, the Styx hub between biomes. God of War's ladder teaches the same

at realm scale, with the boat rides and Tyr's temple doing the interstitial work.

The realm-road translation is unusually clean because the charter already built the machinery. Each

realm's SIGNATURE LAW is a pact clause: it is learned in the threshold zone (Law 5, legibility),

it constrains play for the rest of that realm, and from road position three onward each realm

reprises at least two laws already learned (§2.6, reprise_of). By Ch 74 the player carries eight

accumulated constraints. Demand rises monotonically with ZERO stat inflation, on a mechanism the

project has already ruled and can already measure off the registry (count signature laws per realm,

count reprise references per realm).

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard.

Carrying ordinal: demand 5..11, instrumented as accumulated active constraint-clauses.

Relief carrier: SEAM SANCTUM (the inter-realm crossing as the boon room — the ruled playable

transition plus the ruled loadout-swap-at-vril-sites is already a boon room in everything but name)

and EXPLOIT PAYOFF.

Landing: the fairy realm is the top rung, and the reason it reads sovereign is Law 0's prediction —

every law the road taught is simply the local weather there.

Strongest objection. Six consecutive peaks across five death-and-judgment realms is the fastest

route to the monotone-solemn slop mode the charter itself names (§4.2's register slice exists

because of exactly this risk), and it flattens nine cosmologies into a gauntlet — the SMT/Persona

failure mode L-R1 is chartered to teach us to avoid. The pact ladder is the best CARRIER in the set

and the weakest SHAPE.

---

3. THE PURGATORIAL TERRACES — the rung that transforms rather than escalates

Lens: L-R4 pilgrimage-traversal-and-awe (Journey dual-listed per the charter widen, Shadow of the

Colossus, Ico, Death Stranding), widened with a proposed literary cluster: Dante's Purgatorio and

the Divine Comedy's terrace structure. Proposing a literary comparator is inside the examples-not-

limits rule and the lens roster is explicitly LIVING; the roster has no non-game structural teacher

at all, which is itself a gap worth a row.

What the lens actually teaches. Purgatorio's mountain has seven terraces and the pilgrim gets

LIGHTER as he climbs — each terrace erases one mark from his brow, so the ascent's difficulty FALLS

while its significance rises. Dante is explicit that the last stretch is the easiest and the most

terrifying. Shadow of the Colossus teaches the same at game scale: sixteen fights, no stat

progression worth the name, escalation entirely in what each fight COSTS. Journey's summit is a

whiteout where the player's accumulated ability is stripped and the emotional peak arrives at the

moment of least power.

The realm-road translation writes itself off the built spine, because canon already authored the

sheddings: Hades strips the option of forgetting (she drinks the harder water); Xibalba strips force

(power is gone, wit is all that is left); Naraka strips the enclosing (the verb is RELEASE); Annwn

strips the second blow (the verb is RESTRAINT); Yggdrasil strips the open future (the carved debt is

a forward constraint the player writes on themself). She arrives at Ch 75 carrying nothing but the

choice.

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard, with the note that

this candidate cares least about the enum — its claim is that the rise is in SIGNIFICANCE and the

local gauntlet may legitimately fall.

Carrying ordinal: demand, split. This candidate's real contribution is that a single demand_ordinal

is under-specified: it argues for demand_combat and demand_moral as sub-axes, so total demand rises

monotonically while each faculty rests in turn.

Relief carrier: FACULTY ROTATION — and this is the strongest relief idea in the whole set, because

it is structurally identical to HL_0049 (no two consecutive nodes tax the same faculty; every

5-window carries ≥3 distinct primary faculties), which means the tooth is a copy of R1/R2 against a

new column rather than new machinery.

Landing: Ch 76's father encounter is the last terrace (she sheds the demand to be answered), and Ch

77 is the summit where nothing is left to shed — the standard/valley reading of the Waterfall stops

being a problem and becomes the point.

Strongest objection. "The ascent that gets easier" is a hard sell against a directive that says

BUILD intensity, and it can read as an intellectual defence of an anticlimax. It also risks the

player experiencing the last three realms as a decrescendo in the hands, whatever the columns say.

Mitigation: pair it with candidate 2's carrier so the hands stay loaded while the faculties rotate.

---

4. THE MEGADUNGEON — the road as one eight-realm dungeon, and the pacing unit drops a level

Lens: zelda-dungeon-tool-ladder and sm64-hub-star-space-reuse (charter widen: add per-world

ruleset and anthology escalation), plus L-R2 otherworld-ruleset-and-impossible-space, whose anchor

title is Control — a building whose physics ARE its lore.

What the lens actually teaches. A megadungeon's pacing is not per-floor intensity. It is the

key-and-lock ladder: each floor grants a tool that re-opens earlier floors, and the relief beat is

the SHORTCUT — the moment a door unlocks back toward the bonfire. Intensity per floor stays

comparatively flat; the escalation lives in the compounding toolset and in how far the return trip

now reaches. Crucially, dungeon pacing alternates at the ROOM level, not the floor level.

That last sentence is this candidate's real contribution, and it is the cleanest answer in the set

to "what carries relief" without touching a hard line at all: FOR CH 70-77 THE PACING UNIT MOVES

FROM THE CHAPTER TO THE BEAT. The chapter column goes monotone-high; the alternation obligation

moves into the beat graph, where the spine already carries 10-13 typed beats per node with role and

mode fields. Ch 70 already contains its own trough (the reverent nekyia, CH70_B04, a dialogue beat

between two low-threat traversal beats and the egress fight). Ch 73 has the feast; Ch 74 has the

wells; Ch 77 is nine-tenths trough by design.

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard, with a MANDATED

per-node authored low-pressure beat.

Carrying ordinal: demand, instrumented as the compounding law-count (identical to candidate 2).

Relief carrier: BEAT-LEVEL TROUGH plus SEAM SANCTUM (the shortcut home — each passed realm becomes

re-enterable from the road, which also gives the endgame free-play coda somewhere to live).

Landing: the fairy realm is the dungeon's final floor and the Waterfall is the shortcut that opens

back to the very beginning — the game's first mechanic returned (HL_0112), which canon already

authored as the ring closure.

Strongest objection. Moving a canonical column's granularity without a tooth that can SEE the new

level is an intention, not a law: check_chapter_topography.py reads chapter rows and would report

the arc as flat-green while the beats did whatever they liked. The candidate is only real if the

beat-trough clause ships with a gate that walks the beat graph. That gate is cheap (the beat rows

are already typed) but it is not free, and until it exists this candidate is the most likely of the

eight to rot into a nice paragraph.

---

5. ACT V — the law of rising action changes after the climax gate

Lens: witcher-3 and baldurs-gate-3 (both in flight, both act-structured endgames) plus Disco

Elysium from cell 19, whose whole subject is prose as the primary play surface — which is what the

last three nodes of this arc actually are.

What the lens actually teaches. A five-act structure does not run one pacing law throughout. Acts

1-4 alternate because an audience needs to survive three hours. Act 5 stops alternating and switches

to COMPRESSION: scenes shorten, the cast converges into one room, the locations contract, and every

surviving thread gets its single line. Shakespeare's relief in Act 5 is not a low-tension act; it is

one tonal-drop SCENE inside the rising action — the gravediggers before the duel, the porter after

the murder — placed to make the next beat unbearable rather than to rest the audience.

The realm-road translation gives the arc a monotone axis nobody has proposed and the repo can

already count: THREAD-CONVERGENCE DENSITY. Ch 70 advances three threads (6 CONVERGENCE, 8

CONFRONTATION, 1 ANCHOR) and explicitly carries no family/antagonist anchor because the road leaks

no pattern. As the arc closes, resolutions per node must rise — and T0_Thread_Grid already stores

this. A rise instrumented on an existing populated registry is worth more than a rise instrumented

on a column that does not exist yet.

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard.

Carrying ordinal: threads_resolving_per_node, rising 70→76, with Ch 77 resolving exactly ONE (the

thesis) — the theatrical coda, one line for the last thread.

Relief carrier: the porter scene — exactly one authored tonal-drop beat per node (this is the same

mechanism as candidate 4's beat trough, arrived at from the opposite direction, which is mild

evidence it is the right mechanism).

Landing: convergence density peaks at Ch 76 (the double climax, where the House thread and the

father thread land in the same hall), Ch 77 resolves the single remaining thread alone, and the

Epilogue holds zero — a frame, not a resolution, which is exactly what HL_0045 already requires.

Strongest objection. Convergence density is a strong AUDIT metric and a weak AUTHORING instrument:

you cannot make a chapter feel bigger by resolving more threads in it, and a node stuffed with

resolutions to satisfy an ordinal is the checklist failure mode. Best used as the tooth that

FALSIFIES a flat arc, not as the instruction that shapes one.

---

6. THE REGISTER PROCESSION — the ascent as emotional proximity

Lens: L-R5 underworld-grief-and-register-variety (Spiritfarer, Grim Fandango, Edith Finch), whose

REGISTER slice the charter already pulled forward into the doctrine wave precisely because five

consecutive death realms otherwise converge into monotone-solemn.

What the lens actually teaches. Grim Fandango's whole existence is proof that a death realm need not

be solemn, and that register is a per-realm CHOICE rather than a genre inheritance. Spiritfarer

proves grief can be an entire game's register without ever becoming grim, and its ferrying structure

is the closest analog in the medium to Ch 75's grief-crystallised-into-terrain.

The translation: the monotone axis is EMOTIONAL PROXIMITY — how close each realm's emotional content

sits to her own. Hades is a stranger's grief, handled as a traveller. Xibalba is a stranger's fear.

Naraka is a stranger's moral order. Annwn is a stranger's HOSPITALITY, which is the first one that

costs her something personal to accept. Yggdrasil is a world dying of the same depletion as her own.

Ch 75 is her father's grief, made of the land she was born in. Ch 76 is her father. Ch 77 is her own

face in the water. That is a monotone rise that is already fully authored in the built spine — this

candidate discovers the pattern rather than imposing one, which is its whole argument.

Intensity string: standard · peak · peak · standard · peak · peak · peak · standard — one cell

changed from live (Ch 74 standard → peak). This is deliberately the MINIMAL-CHANGE candidate, and it

is included as the steelmanned near-status-quo option the decision protocol requires.

Carrying ordinal: proximity_ordinal (proposed), 1..8 across Ch 70-77, no exemption needed — Ch 77 is

its maximum, which fixes the awkwardness every other candidate has with the exempt terminal node.

Relief carrier: REGISTER ROTATION, natively.

Landing: the arrival lands because the road has been rehearsing intimacy at increasing magnification

for eight nodes, so Ch 76's hall is not a difficulty spike but the first room where the distance is

zero.

Strongest objection. Proximity is invisible to the player as a system and therefore cannot be

FELT as escalation in the hands; it is a critic's axis. A player who is not moved by Annwn's

hospitality experiences node 73 as a lull no matter what the proximity column says. It also leaves

the directive's "build intensity" almost entirely unexecuted at the mechanical layer.

---

7. THE SOVEREIGN REHEARSAL — the ascent as a ladder of thrones

Lens: L-R1 mythic-cosmology-adaptation (God of War's realm-by-realm sovereign ladder is the shipped

precedent) plus L-R7 fading-court-realm (Twilight Princess's exiled sovereign; Hollow Knight's

ruined king).

What the lens teaches, and why it is not candidate 6 restated: God of War's nine realms are held

together not by escalating combat but by a repeating ENCOUNTER GRAMMAR with authority — each realm

has someone who rules it, and the protagonist's relationship to rule is the arc. The charter's

schema already reserves a field for this: sovereign_distance sits in the INTENSITY BLOCK at §2.3

and is currently un-designed.

The road's sovereigns are already authored and already form a ladder: the just implacable lord who

rules death by lot (impersonal authority, honoured, never fought); Louhi, a sovereign matriarch who

is RIGHT to fight and withdraws still sovereign; the Lords of Xibalba, grave order-keepers who must

be out-thought rather than deposed; Arawn, a host who becomes a FRIEND and whose kingdom she rules

in his place for a year; the Norns, who tend rather than rule; and then a King who is her father,

whose grief has replaced his governance. Every road sovereign is a rehearsal for the last one, and

Annwn's year-and-a-day stand-in charge is the most on-the-nose rehearsal in the corpus — she

literally holds someone else's throne before she walks into her father's hall.

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard.

Carrying ordinal: sovereign_distance, descending 8→0 (or inverted to an ascending

sovereign_proximity so all ordinals rise in the same direction — a schema-hygiene note for the

director: mixed-direction ordinals in one block are a persistent read-error source).

Relief carrier: AWE-AS-REST at each throne-room arrival, plus REGISTER ROTATION.

Landing: Ch 76's two courts are the payoff of the ladder — a guest seated in a grief-made door and

then a father who has stopped ruling. The pattern makes Ch 76 the structurally inevitable node

rather than merely the loudest one.

Strongest objection. It is a thematic pattern wearing a pacing pattern's clothes. It says nothing

about what the player's hands do at Ch 72 versus Ch 73, and a pacing law that cannot be violated by

a bad combat encounter is not a pacing law. Best adopted as the arc's L3 reading and as a

sovereign_distance population rule, not as the pattern itself.

---

8. THE STRIPPING ASCENT — erosion, then restoration

Lens: Journey (charter widen: dual-list into L-R4) plus sekiro-timing-defense-chain-meter, whose

subject is demand rising against a toolkit that does not grow.

What the lens teaches. Journey's final ascent removes the player's accumulated scarf, then removes

movement, then removes light — the emotional summit is the point of minimum capability, and the

flight that follows is overwhelming precisely because everything was taken first. Sekiro raises

demand for thirty hours while the player's numeric power barely moves.

The road already runs half of this, ruled and gated: the vril POWER curve troughs at Ch 70-71 and

Xibalba is canon's own floor, where the myth's heroes win by wit because power is gone. This

candidate extends the erosion from vril to the whole loadout: realms that do not admit the bonded

(familiars barred where the living-and-bound cannot pass), companions who cannot cross a given

threshold, realm deltas that close specific ability classes (ability_availability_delta is already

a schema field). By Ch 74 she is at her most constrained; Ch 75 restores the CEILING for the first

time since the Prologue, and that restoration is the largest single exhale available in the game.

Intensity string: standard · peak · peak · peak · peak · peak · peak · standard, with the erosion

running 70→74 and the restoration at 75.

Carrying ordinal: demand, with a companion constraint_count that rises 70→74 and drops to zero at 75.

Relief carrier: POWER TROUGH AND RESTORATION — this candidate's relief is the single biggest one in

the arc and it is already canon; nothing needs inventing.

Landing: the fairy realm arrives as power, which is the correct emotional shape for a HOMECOMING and

the only candidate whose arrival mechanism is joy rather than pressure.

Strongest objection. It collides with Law 6, THE FLOOR LAW, which states that a realm may close a

usual answer but NEVER confiscate the build — build, mastery tiers, gear, imprints and bonds cross

every threshold intact. Aggressive stripping is illegal under the charter's own floor. The candidate

survives only in its soft form: realms CLOSE answers (which the floor permits and the signature-law

lane already does), they do not take things away. Stated that way it is nearly identical to

candidate 2's constraint accumulation, which is a sign the two should merge rather than compete.

---

9. Cross-candidate comparison

Three cross-cutting dials, independent of which shape is adopted:

• THE FLOOR-SET. Ch 70 at standard (all candidates) versus VALLEY. Valley is legal — R5 pins only

the relief value, and Ch 70 is genuinely a low-gauntlet node (the descent is unopposed, the

honoured dead are never targets, combat routes only through the egress). A valley at Ch 70 gives

the ascent maximum runway and makes the law-change at the seam audible. It costs one thing: the

arc opens on its quietest node, and a player who reads valley as "filler" gets the wrong signal

from the road's most beautiful chapter. RECOMMENDED: standard, and let the vril trough carry the

floor — the power column already says "you are at the bottom" without the pacing column having to.

• THE HUSH SLOT. None (2,4,5,7,8) versus Ch 73 (6) versus Ch 74 (1). See §10.
• THE TAIL. Ch 77 standard (live, all candidates) versus valley. Under §0.1's semantics valley is

arguably more honest, but the awe column is the correct carrier for the Waterfall's magnitude and

a valley label on the game's terminal revelation invites every future reader to mis-tune it.

RECOMMENDED: keep standard, and let awe read CEILING.

---

10. Recommendation — THE FINAL ASCENT LAW (a composite, and the string it produces)

The eight candidates do not compete cleanly; they contribute different organs. The recommendation is

the composite, named so it can be cited as a replacement for mountain-range topography on this arc.

THE SHAPE comes from candidate 1's subito piano, relocated by candidate 6's content reading. The

crescendo lens is right that a rise needs one hush; the register lens is right about WHERE, and the

built spine has already authored it. Annwn is canonically "the last warm beat before the crossing,"

its fight is a single one-blow honour duel won by RESTRAINT, and its register is

seductive-hospitable. Yggdrasil, one node later, is a marquee dragon fight plus a crossing-defense.

The hush is already at Ch 73 and the door is already the loudest node on the road — the live

registry simply has the door mislabelled at standard because, until this ruling, lifting it would

have produced three consecutive peaks and broken HL_0048 outright. That constraint is exactly what

Josh removed. The ruling therefore unlocks precisely the one change the realm charter wanted at

§6.2 and could not take.

RECOMMENDED INTENSITY STRING, Ch 70-77:

standard · peak · peak · standard · PEAK · peak · peak · standard

One cell changes from live: Ch 74 Yggdrasil, standard → peak. Ch 74-75-76 becomes three consecutive

peaks, which is now legal and is the point. Epilogue stays valley. This also settles the charter's

C-1 fork on its merits rather than by workaround: lift the DOOR, not Annwn.

THE CARRIER comes from candidate 2, softened by candidate 8's Law-6 collision: demand rises

monotonically 5→11 and is INSTRUMENTED as accumulated signature-law constraint, never as stat

inflation. This reuses the ruled difficulty doctrine, the charter's compounding budget, and

reprise_of — all of which already exist. Nothing is minted.

THE RELIEF LAW comes from candidate 3, with candidates 4 and 5 supplying the granularity: faculty

rotation across nodes, plus one authored low-pressure beat inside every node. Both are shaped

exactly like clauses the repo already gates, so the teeth are copies rather than inventions.

THE READING comes from candidates 6 and 7 and costs nothing to adopt: the arc's L3 truth is that

proximity and sovereignty rise together, from a stranger's underworld to her own face. It is felt,

never declared — the same discipline the road already holds for the convergence thesis.

The law, as five checkable clauses

• CLAUSE 1 — THE RISING FLOOR. Across Ch 70-76 no node's intensity may fall more than ONE step below

the previous node's, and no node may be a valley or a relief. (peak=3, standard=2, valley=1.)

Strictly stronger than the removed clauses, so §0.5's valleys ambiguity dissolves.

• CLAUSE 2 — THE ONE HUSH. Exactly one step-down is permitted across Ch 70-76, and it must be

immediately followed by a peak. A hush is a preparation, never a rest. (Live at Ch 73 under the

recommended string; the clause is what makes it legal and what stops a second one appearing.)

• CLAUSE 3 — THE MONOTONE DEMAND. demand_ordinal strictly increases Ch 70→76. Ch 77 is EXEMPT via

the empty-string sentinel and the tooth skips empty rows explicitly rather than coercing them —

the charter's own construction, adopted unchanged.

• CLAUSE 4 — THE FACULTY ROTATION (the clause that replaces valleys). No two consecutive nodes take

the same primary demand-faculty, and every 5-node window carries ≥3 distinct primary faculties.

Structurally identical to HL_0049's two clauses, so the tooth is R1/R2 re-pointed at a new column.

• CLAUSE 5 — THE BEAT TROUGH. Every node in Ch 70-77 contains at least one authored low-pressure

beat in its beat graph. The breath moves inside the chapter; the gate walks the beat rows.

Plus the seam rule from §0.5: R3 judges only windows lying entirely within Ch 1-69; the Final Ascent

Law owns Ch 70-77 exclusively; Ch 69 stays peak and Ch 70 stays standard so the boundary is

unambiguous in practice as well as in the gate.

The seam is a feature, not an artifact

The ruling's range begins at Ch 70, which is node five of the nine-realm road (Ch 66-74) — the

ascent's law changes in the MIDDLE of the road, not at its mouth. That reads at first as a scoping

artifact worth extending back to Ch 66. It should not be extended, and the Act-V lens says why: a

law change is only felt if there is a visible seam. Ch 66-69 alternating under mountain-range law is

what makes Ch 70-77's continuity legible as a different world. Ch 70 is also the vril trough and the

katabasis — the ascent begins at the floor, which is the correct place for an ascent to begin. If

Josh later wants the whole road under the new law, that is a one-line extension; the recommendation

is to leave it at 70 and let the seam do its work.

The strongest objection to the recommendation

That it changes one cell and calls itself a new pattern. A reader could fairly say the intensity

string barely moved, so the "different pattern" Josh ordered has not been delivered.

The answer: the string is the smallest part of the deliverable and always was — §0.1 shows the

column cannot carry the directive at all. What actually changes is the LAW: mountain-range topography

is replaced on this arc by a rising-floor law with a single sanctioned hush, a monotone demand

column that did not previously exist, a relief obligation moved from the intensity column onto

faculty rotation and the beat graph, and four new teeth. The intensity string moving only one cell is

evidence that the built spine was already the right shape and was being held down by a rule written

for a different part of the game — which is precisely what Josh's ruling diagnosed.

---

11. What this hands to the other axes and to the director

• To the RELIEF analyst: seven named mechanisms in §0.6; the recommendation depends on faculty

rotation and the beat trough, and treats register rotation and power-restoration as already-canon.

If that analyst rules a different primary mechanism, clauses 4 and 5 are the swap points and the

rest of the law is unaffected.

• To the HARD-LINE-FLAG axis: the recommended string raises Ch 74, which is ALREADY flagged TRUE.

The five FALSE nodes (70, 71, 72, 73, 75) are untouched by this axis — the pattern does not need a

flag change and should not be read as implying one.

• To the DIRECTOR, decisions this analysis surfaces rather than settles:
• the seam rule and the valleys-clause ambiguity (§0.5) — both are gate-authoring decisions with

recommendations attached, and both are the silent-choice defect class if left to the tooth author;

• whether demand_ordinal mints now (it is the recommendation's load-bearing column and does not

yet exist as a live field);

• ordinal direction hygiene in the intensity block — sovereign_distance descends while every

other proposed ordinal ascends;

• the presentation-craft-anim-audio charter amendment carrying arc-pacing, not only per-realm

audio identity, before the DR-2 freeze;

• the lens-roster addition of a non-game structural comparator cluster (Dante, dramatic structure,

orchestral dynamics) — the roster currently has no structural teacher that is not a shipped game,

which is a roster gap by the roster's own governing principle.

12. Verification notes

• HL_0048 (amended), HL_0049, HL_0050 quoted from the live T0_Hard_Lines [ACTIVE v0.2]/Sheet1.csv,

not from memory or from the charter's paraphrase — the charter itself records that this exact

mis-attribution class was caught and corrected once before.

• The Ch 60-78 intensity/mode/flag table read directly from the live Chapter_Index CSV.
• The enum values, R1-R6 semantics and the relief-set pin read from

harness/check_chapter_topography.py (module docstring and evaluate).

• Ch 70-77 chapter content (crossing functions, bosses, winning verbs, registers, beat counts, the

Ch-77 approach fold) read from the live spine entries docs/spine/CH_70.md .. CH_77.md and

CH_EPILOGUE.md.

• The charter's intensity finding, the three instruments, the signature-law lane, the compounding

budget, Law 6's floor, the C-1 fork and the lens roster read from

docs/proposals/REALM_DESIGN_PROGRAM.md §1.1-§1.4, §2.1-§2.6, §4.1-§4.3, §6.2-§6.3 and

docs/proposals/COMPARATOR_LENS_PROGRAM.md.

• The ruling text read from docs/spine/DECISIONS_PENDING_JOSH.md fourth-sitting block.
• Absence claims made here and their controls: no lens on the 40-lens roster (nor among the seven

proposed realm lenses) is a non-game structural comparator — checked by reading the full

consolidated roster and the seven L-R charters, where the same read successfully identifies

Journey, Dante-adjacent absences and the presentation lens, so the search can match and the zero is

real. sovereign_distance appears exactly once in the corpus, in the charter's §2.3 intensity

block, with no design body attached — the same read finds demand_ordinal, awe_ordinal and

power_ordinal in the same list, so the matcher demonstrably matches.

Generated by harness/site/structure_site.py — the URL path is the repo path. review root