docs: correct the "Act 1 idles beneath Act 2" premise; record the launch runway

docs/decisions.md

@@ -630,6 +630,20 @@ The Act 2 launch gate is simply **holding 250,000 prestige rank at the moment of
 
 **System**: vessel (`src/vessel/`, launch gate).
 
+### Vessel Launch Runway: ~828 Hours of Calendar Time — Arithmetic, Not a Simulation
+
+The 250,000 PR gate is reachable, and the answer is **closed-form rather than empirical**. The Loom's WR→PR conversion is the dominant endgame PR source: `wr_to_pr_per_hour(wr) = wr × (1 + wr/100)` (`src/loom/logic.rs:811`), documented peak **131 WR/hr → 302 PR/hr**. The grant (`src/core/tick_stages.rs:1235-1248`) computes `fill_secs = 3600 / pr_per_hour`, then adds `elapsed / fill_secs` PR, where `elapsed = (now - wr_pr_last_granted_at).min(604_800)` — **wall-clock seconds, capped at 7 days**.
+
+At peak: `250,000 ÷ 302 ≈ 828 hours ≈ 34.5 days` of *calendar* time, accruing with the game closed. The 7-day cap grants at most `302 × 168 ≈ 50,700` PR per catch-up, making the gate roughly **five weekly check-ins**. A player who leaves for a month and returns collects one week, not four — that cap is the only behavioral demand the runway makes.
+
+**Why this is not a simulator question**: there is no RNG in that path. `wr_to_pr_per_hour` is a pure function of WR; the grant is integer division on wall-clock elapsed. Nothing is emergent, so simulation can reveal nothing that division cannot. Worth recording because the headless simulator *structurally cannot* answer it: `src/bin/simulator/strategy.rs:361-363` notes the Loom's pattern sustain runs on wall clock "which doesn't advance meaningfully in the headless simulator" and injects pattern completions instead. Two 3,000-hour runs started at P50,000 and P250,000 with different seeds returned `pr_earned` of 17,995 and 17,994 — identical, because that PR came from injected challenge wins at ~6 PR/hr, not from the Loom. The simulator is not broken; it is answering a different question.
+
+**Scope of `balance-sim`, therefore**: it validates Act 1 progression to Z50 / Ascension X — the 28-pattern and Ascension X *prerequisites* of the gate. It does not and should not validate the PR runway. **Do not read a green balance-sim run as evidence about Act 2's entrance.**
+
+**Alternatives considered**: giving the simulator a virtual wall clock so the Loom economy runs headless (rejected for this purpose — it would spend real effort reproducing a closed-form constant; revisit only if the Loom gains RNG or cross-system feedback that makes its output genuinely emergent); leaving the runway unrecorded (rejected — it was mistaken for an open empirical question and briefly treated as a release blocker, which is what prompted this entry).
+
+**System**: vessel (launch gate) × loom (WR→PR) × simulator scope.
+
 ### Act 2 Souls: The Covenant — Nothing Harms a Soul While You Are Away
 
 A hard, CI-enforced invariant: **no tick-driven code path may reduce the roster.** Nights, weather, drift, hold-station, and offline resolution never touch souls; a property test simulates arbitrary offline windows and asserts roster count is invariant. Souls are lost **only in authored scenes** attached to named threats — the threat was on the junction card, the road was chosen, and the scene offered a priced alternative. There is no dice-based loss. A loss is memorialized: the soul's name is carved into the hull art for the rest of the game (Act 3 included), their arc becomes a memorial manifest line, and their counsel lines go silent.

docs/dossiers/act2-pilgrimage.md

@@ -283,7 +283,7 @@ Act 1 (everything)                Act 2: launch → crossing 1 → the Ferryman
   250,000 PR (burn)   ─┘        │               │             (Act 3 hook #1 —
   Zone 50 kill ─► signal ───────┘               │              spec-normative)
                                                 ▼                     │
-  Deep/Loom/etc. idle beneath ◄── untouched   Colony founded ◄────────┘
+  Deep/Loom/etc. STOP (frozen, no ticks)      Colony founded ◄────────┘
                                                 │
                           ┌───────────────────► Reckoning: Salvage
                           │                     ─► Drive / Shipwright / Ward
@@ -304,8 +304,14 @@ Act 1 (everything)                Act 2: launch → crossing 1 → the Ferryman
 - **In**: the launch gate deliberately braids *all* of Act 1 (Loom, Ascension,
   PR economy, Zone 50) into one burn — the act's strongest cross-system edge.
 - **During**: zero mechanical interaction with Act 1 — deliberate ("one-way
-  passage"); Act 1 idles untouched beneath. Narrative callbacks only
-  (Torvald is the Deep guild's captain).
+  passage"). Act 1 does not *idle* beneath: it **stops**. `main.rs:555` takes
+  the voyage branch and `continue`s past both `game_tick_with_context` call
+  sites, so no zone ticks, no PR accrues, no Haven builds — and permanently,
+  since `vessel_launched` is never unset. The hero the player spent hundreds of
+  hours building is inert for the whole era. Confirmed intended (2026-07-14);
+  noted here because the earlier "idles untouched" phrasing was read as "keeps
+  running" and became the (void) premise for the "Act 1 bridges" Phase 2 idea.
+  Narrative callbacks only (Torvald is the Deep guild's captain).
 - **The ferry loop is the act's second major cross-system braid**, but an
   internal one: Reckoning spend (Drive/Shipwright/Ward) feeds directly back
   into how much of the old world the next crossing can rescue before the

docs/explorations/2026-07-05-act2-systems-braiding.md

@@ -82,6 +82,30 @@ most of Act 1's growth systems are already maxed.** Ascension X is the
 level cap. 28/28 patterns is Loom's completion state. The two hardest gates
 in the game are, definitionally, *finished* the moment Act 2 opens.
 
+> **⚠️ CORRECTION (2026-07-14) — the paragraph below is factually wrong. Do
+> not build on it.**
+>
+> Act 1 does **not** keep running during Act 2. `main.rs:555` takes the voyage
+> branch on `vessel_launched && act2_enabled()` and `main.rs:1018` `continue`s,
+> while both `game_tick_with_context` call sites sit at `main.rs:1253` / `:1608`
+> — *after* it. Zones don't tick, Power Cores don't generate, Haven doesn't
+> build, nothing is gettable. Act 1 is **frozen**, and permanently so
+> (`vessel_launched` is only ever set `true`). Confirmed intended design,
+> 2026-07-14.
+>
+> **What this voids:** there is no stranded output to redirect, so the "second
+> outlet" framing below — and the "Act 1 bridges" Phase 2 item it justified
+> (#734) — has no premise. Any such bridge is *new content*, not reuse, and
+> must be re-argued from scratch on its own merits.
+>
+> **What survives:** the opening observation still holds — Ascension X and 28/28
+> patterns *are* maxed at launch. But they're not "stranded and still
+> producing"; they're switched off. The real question this exploration should
+> have asked: **is it right that the hero a player built over hundreds of hours
+> goes inert for a 3–7 month era?** That is a live design question, unanswered.
+>
+> Preserved verbatim below as the record of how the item came to be proposed.
+
 So a player deep in a 3–5-month Ferryman era who keeps playing Act 1 in
 parallel (zones still tick, Power Cores still generate PR, Haven rooms
 still build, Fishing/Challenges/Achievements are all still gettable) is

src/main.rs

@@ -549,9 +549,16 @@ fn main() -> io::Result<()> {
                     'game_loop: loop {
                         // ── Act 2: The Crossing ─────────────────────────────
                         // After the launch burn, the loop belongs to the
-                        // voyage: chart, holds, junctions. Act 1 systems idle
-                        // untouched beneath it; spec 6 (the Going-Dark) owns
-                        // their eventual wind-down.
+                        // voyage: chart, holds, junctions. This branch ends in
+                        // `continue`, so everything below it — including both
+                        // `game_tick_with_context` call sites — is skipped while
+                        // launched. Act 1 is *frozen*, not idling: no zones
+                        // tick, no PR accrues, no Haven builds. And because
+                        // `vessel_launched` is only ever set true (never unset),
+                        // that is permanent for this character. Do not write
+                        // "Act 1 idles beneath" here — that phrasing has
+                        // misled design docs into assuming Act 1 output keeps
+                        // accumulating during the era. It does not.
                         if state.vessel_launched && vessel::act2_enabled() {
                             // The 5-beat launch transition (spec 4) plays once,
                             // between the burn and the first Voyage frame. It