EscapistsReintroduced/docs/evidence/rescue-factions-2026-10-13/README.md

Rescue faction native evidence — 13 October 2026

This evidence qualifies faction-specific rescue assets and local native fleet mechanics. It does not qualify dedicated multiplayer, remote casualty ownership, twelve simultaneous players, escort combat effectiveness, or the full escape-boundary journey.

The full run-a03 fixture completed 112 assertions with no script or VM failures. Its six-faction matrix temporarily selected each real faction and its matching resolved choice, then checked the actual aircraft affiliation, original living crew prefabs, exact pilot/gunner seat occupancy, and free cargo capacity. Matrix aircraft were held still for inspection and deleted afterward. The later US flight used normal rescue flight control without that hold.

Selected faction	Transport	Free cargo per transport	Fleet capacity	Crew actually observed
US	Two unarmed Hueys	8	16	US helicopter pilot
USSR	One armed Mi-8	14	14	USSR helicopter pilot
FIA	One Mi-8 with Soviet flag removed	14	14	FIA squad leader
RHS US	Two unarmed Hueys	8	16	RHS USMC helicopter pilot
RHS Russia	One Russian Mi-8	14	14	RHS Russian helicopter pilot
ION	Two Hueys with civilian liveries	8	16	ION random squad leader

Armed escorts were checked separately for all six selections. The Western armed Hueys have four crew and six cargo seats; they are not counted toward the rescue transport capacity. FIA and ION use members of their selected native group because their faction choices do not provide a dedicated pilot in this mapping. Their actual prefab names appear in the matrix log.

The US flight dispatched two transports and two armed escorts through the production extraction service. Both transports physically landed on separate native-surveyed footprints approximately 90 metres apart. An already-unconscious living NPC boarded the second transport through the native cargo primitive, and the original controlled player boarded the first. The production all-survivors boarding timer accepted occupants distributed across both transports, then initiated outbound takeoff on both loaded helicopters. That first fixture checked departure state and loss of boarding readiness, rather than a minimum airborne height. The NPC was registered under an isolated fixture identity for the fleet departure check; this is not a second real networked player.

The legacy Core smoke fixture uses a hillside pickup at approximately 7012, 63.725, 7500, without space satisfying the second-aircraft survey. Run a03 explicitly logged that limitation and arranged an airport pickup only after both native footprint checks succeeded. Its pads were approximately 4788.55, 28.435, 11528.9 and 4788.55, 28.302, 11438.9. Later runs move only the isolated fixture's authored site input to those grounds before preparation; production assets and footprint rules are preserved. The fixture directly arranges the valid mission phase sequence and signal, so this run does not repeat smoke-grenade input acceptance.

The final run-a06 fixture completed 20 assertions with no script or VM failures, compiling the final source and qualifying the prepared pickup before dispatch. Its root compartment manager exposed ten compartments, including all eight accessible, audited cargo benches; the production physical feedback and casualty seat selector accepted the second transport. The final short-approach fixture adds a bounded departure poll: both engines running, no native ground contact, altitude more than ten metres above terrain, upward velocity over 0.5 m/s, and the original survivors retained in their separate aircraft. Physics activity, engine state, contact, velocity, altitude and flight phase are logged during the poll. Only the initial approach allocation is shortened; subsequent flight, landing, boarding and departure controls are unchanged.

The player, aircraft and original flight crew are protected from combat damage to isolate flight and boarding. Collision, terrain support, native compartments, unconscious state and the production fleet boarding timer remain active. The fixture moves the test player between objectives; it does not reposition an aircraft after its initial allocation.

Diagnostics and retained failed attempts

Passing behavioral assertions does not mean the strict native report passed. Runs a03 and a06 each retain 21 engine error occurrences, representing 16 distinct messages after timestamps are removed. Those messages also occur in the preceding Cessna boarding run-04; they include installed-addon localization resources, a missing AI goal class, inherited prefab properties, Steam statistics and authored world GUID diagnostics. The strict report remains failed. Additional warnings include repeated no-default-weapon-slot warnings, Mi-8 rotor collider warnings and a duplicate HUD class warning. Escort weapon effectiveness was not exercised.

Every attempted run is retained with complete available console, script and error logs. Run a01 could not start the engine because the sandbox blocked the local audio socket. Run a02 had an invalid fixture faction setup: changing the catalog field did not change the cached selected faction, so its all-US-crew matrix is unqualified; an immediate dispatch assertion also failed. The corrected a03 fixture checks the actual selected faction and independently resolves the expected crew. Run a04 exposed an Enforce compiler error in the new landing-pad helper's modulo expression; the production helper was corrected to use Math.Mod before the next run. Run a05 passed preparation, landing and wounded boarding, including all eight audited root-manager cargo benches and the production seat selector, but sampled the final departure only 6.3 seconds after lift-off was ordered. The controller intentionally holds zero climb for ten seconds; that fixed-time fixture check was premature. The final fixture waits for measured climbing instead. A small production correction also restores the base controller's TAKEOFF-only physics wake in the rescue override as prevention against sleep during this hold. Run a05 does not establish that either helicopter was permanently stuck.

Run summary preserves assertion counts and explicit qualification notes. Report binds qualified results to source hashes. Each run also retains its frozen fixture, fixture inputs, complete source manifest, external-addon input hashes, and runner cleanup result. The isolated runner mutes only its own audio, serializes access to the disposable Wine prefix, and does not contact or modify the live game server.