This is a choice between two ways to carry the same SMALL campaign promises, not a choice between a short and a long game. Neither count demonstrates the owner’s 30/90/100–200-hour ambitions. Both keep four regions, discovery first, expressive connected builds second, meaningful ordinary hordes, two alien languages, short cooperative rescues, five people aboard, and the optional task of physically joining every playable room that remains in that option.

In the **40-room reference**, an ordinary expedition moves through places with separate jobs: R05 teaches voluntary motion and may later yield the navigation coupler through R07; R28 lets the player operate the regulator racks while R30 can supply a distinct before/after observation angle; R33’s fixed conversion can be compared with R34’s reversible converter; and the core approach gives the first and second clamps their own R37 and R38 recovery spaces. The later Moving Scan in R15 turns an ordinary route, horde and growing Storm Choir build into a bridge discovery at R28. The player comes home with parts, evidence or a person, then eventually returns to connect forty remembered places.

In the **36-room concentration**, the expedition objectives and payoffs remain, but four of those spatial pauses disappear. The coupler is recovered in R05; the regulator is operated and observed inside R28; fixed and reversible transformation are both read in R34; and R36 holds the first clamp before the player reaches the second in R38. The scan, both setpieces, every build part, all five rescues and all eight investigation chains remain. The deepest construction joins thirty-six rooms. This option is therefore not “the same game for ten percent less”: it is a denser version whose retained discoveries compete for more attention in fewer hosts.

| Scope unit or player consequence | 40-room reference | 36-room concentration | Honest production meaning |

|---|---|---|---|

| Regions; objectives; ship assemblies | 4; 12; 4 | 4; 12; 4 | No reduction in campaign dependency or reward logic. |

| Reusable modules; placed rooms | 24; 40 | 22; 36 | Removes two module definitions and four independently identified placements; reduces unique layout, room-state and join-graph obligations. |

| People; primary scenes; home projects | 5; 15; 5 | 5; 15; 5 | No cast, rescue, dialogue-state or home-project cut. |

| Equipment definitions | 12 | 12 | No reduction in connected-build rules, acquisition or compatibility work. |

| Encounters; setpieces; enemy roles | 10; 2; 6 ordinary + 3 special | Same | Meaningful hordes and both climaxes remain; no population, source or staging work is silently removed. |

| Mechanism families; major placed mechanisms | 5; **14 proposed** | Same | P14 Moving Scan remains in R15 with its R28 receiver. Current Production Scope and Whole Game Outline still say 13; that is reference lag, not a cut. |

| Languages; investigation chains; evidence index | 2; 4 language + 4 story; 26 entries | Same | All comparisons, corrections and consequences remain; rehosting does not erase their writing or state cost. |

| Deepest joining payoff | Connect all 40 placed rooms | Connect all 36 placed rooms | A smaller physical achievement and graph, but still the entire chosen station. Neither option selects what completion does or an ending. |

| Further detailed-campaign documentation | Existing SMALL allowance: 350–590 role-hours | No defensible adjusted total yet | The allowance excludes the current first pass. Only part of its 90–150-hour room/objective/arrangement stream is reduced; the other streams and rehost reviews remain. |

**Removed as independent content.** Delete R07 Navigation Service Room, R30 Reply Observation Niche, R33 Imaging Exchange and R37 East Clamp Refuge. Remove module definition M21 Fixed Conversion Exchange. Absorb M03 Glazed Control Cell into an expanded M02 opening/control module, with R03 becoming M02’s second placement. These are four fewer room identities and two fewer reusable spatial grammars; they are not four deleted objectives or discoveries.

**Moved, with a price.** Move R07’s navigation-coupler release to R05’s existing side service housing. The part and carrying choice survive, but the empty assembly-housing revisit does not. Move R30’s before/after capture angle into a protected recess in R28. The reply chain survives, but operating and observing from different rooms no longer teaches the distinction; R28 must now separate racks, P09 Grid pressure, P14’s receiver landing and the observation cue. Move R33’s fixed A→R exchanger into a separately labelled channel in R34 beside its reversible A↔B mode. The network can still be completed, but the clean room-to-room transformation comparison becomes a crowded same-room comparison. Move R37’s first clamp into R36’s east wall while R38 keeps the second. Both physical alternatives remain, but one full withdrawal refuge disappears and setpiece B becomes denser.

**Retained.** Keep objectives 1–12; R14–R16’s builder-run chain; R21–R23’s visitor-transfer chain; the R28 reply procedure and R36 inlet reply; P14’s moving scan and bridge; all twelve equipment definitions; five people and fifteen primary scenes; ten encounter placements, two setpieces and all nine enemy roles; fourteen mechanisms; and the R36/R38–R40 surviving core assembly with two-clamp access. Keep the ordinary join vocabulary and the obligation that every remaining room can join a ship-rooted network. The recent Rescue Under Pressure crossing/death alternatives and Completing The Station outcomes are unselected proposals and are excluded from both counts.

**New obligations created by the move.** Enlarge M02 so R03 still has readable fastener work and recovery. Add a distinct coupler housing and release to R05 without compromising its gait comparison, belt, Carousel or crowd lane. Fit a protected observation recess and cue hierarchy into R28. Give R34 two visibly different transformation channels, refusal/recovery states and enough space to compare them. Give R36 an accessible first-clamp work face and recovery line, then recheck the compact core assembly, setpiece B and the full connection graph. No new room, objective, scene, source or mechanism is introduced, but this geometry and state work is real.

The published 350–590 role-hours cover **future detailed-campaign documentation and campaign content only**: exhaustive variants for the current first-pass briefs, exact geometry and encounter combinations, clue consequences, cast branches, information details, and persistence/integration review. They do not include the first pass already on the wiki. They also exclude implementation, tools, assets, animation, audio, localisation production, recruitment, performance work, accessibility testing and player sessions.

A numeric 36-room total is unavailable without inventing throughput. Four of forty placements and two of twenty-four definitions disappear, but all objectives, arrangements, mechanisms, clues, people, builds and encounters remain. Even the 90–150-hour rooms/objectives/arrangements stream is not linear by room: moving content into R28, R34 and R36 adds congestion and cross-state review. The other 260–440 hours of the published allowance are not directly reduced by these deletions. No like-for-like completed detailed packages or recorded session timings exist from which to estimate the offset. Until three representative packages—a repeatable mechanism room, a two-language investigation with correction, and a short rescue/home arc—are written and independently checked, planning should retain the 350–590 envelope for either option rather than claim a false saving. The concentration still offers a legitimate directional reduction in unique layouts, placed-room briefs and graph nodes; its exact hours, future art savings and implementation savings are unknown.

Provisionally keep **24 modules / 40 rooms** as the SMALL planning reference. Its four extra rooms buy exactly the kind of separated observation, reinterpretation and physical recovery that makes discovery feel spatial rather than administrative. That matters more after P14: R15 and R28 already carry a rail, scan bay, bridge relation, shared horde trigger and receiver, so moving R30 into R28 spends attention where the current design is already densest. Keeping R37 also protects a staged withdrawal before the deepest core payoff. The accepted price is four more placed-room briefs, two more module grammars, a larger full-station graph and later production work whose feasibility is not yet demonstrated.

The strongest contrary case is the 36-room option. It retains every headline promise while reducing unique layout and connection obligations, and it may be the safer production choice if capacity is tight. Choose it if the owner values a smaller, denser station more than the four lost spatial pleasures and accepts the rehost risks explicitly.

Reopen the recommendation when representative detailed packages narrow the allowance; when an exact paper layout shows whether R05, R28, R34 and R36 can carry their added obligations readably; when a valid complete 40-room versus 36-room connection graph exists; when staffing or budget is known; or when later authorised player observation shows that the extra rooms feel repetitive or that the concentrated hosts overload comprehension. Scope selection authorises no development, simulation, playtest, purchase or art, and neither option chooses the first survivor, mortality, the hospital truth, full-station behaviour or an ending.