Skip to content

Auto Warz — Progression Arc Design

Document Type: Design Document Version: 0.4 Last Updated: 2026-09-05 Vault destination: 01_Design Documents Phase: Spans both. Phase 1 ladder in §3, the War Update in §5 Related Documents: [13_Player Levelling System](<./13_Player Levelling System.md>) · [14_Win Conditions & Faction Allegiance](<./14_Win Conditions & Faction Allegiance.md>) · [35_Economy Design](<../Systems/35_Economy Design.md>) · [02_Scope & MVP Document](<../02_Scope & MVP Document.md>) · [28_Restoration - Third Endgame Path](<./28_Restoration - Third Endgame Path.md>) · [00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) Folder: 01_Design Documents

Purpose of This Document

This document owns the shape of a playthrough: what the player is trying to do at hour 1, hour 10 and hour 100, and what changes between them.

It does not own the reward schedule. [13_Player Levelling System](<./13_Player Levelling System.md>) owns XP and levels, [35_Economy Design](<../Systems/35_Economy Design.md>) owns the currencies, and [14_Win Conditions & Faction Allegiance](<./14_Win Conditions & Faction Allegiance.md>) owns factions and the endgame paths. This document decides the sequence those things are pacing, and the other three hang off it.

It exists because the arc was previously written entirely in the space and war layers. Every named milestone lived above the planet, which left the whole Early Access release with no beats at all.

1. The Principle

Progression is the ladder of bottlenecks, not the reward schedule.

What a factory player experiences as progress is not a number going up. It is the thing that was limiting me is no longer the thing limiting me. Levels, points, unlocks and milestones are feedback on that ladder. They are not the ladder.

This gives a hard test, applied to every beat in this document:

A milestone must change what you are optimising, not just how much.

A beat that makes you build more of what you were already building is not a milestone. A beat that makes yesterday's layout the wrong answer is.

1.1 Ceremony is scarce

Only beats that survive the test get ceremony: a cinematic, a large XP burst, an achievement, a moment where the game stops. Everything else announces itself in a notification and gets out of the way.

Five cinematics in a hundred hours cheapens all five. Two makes both of them matter.

1.2 The arc is not a win condition

Nothing here forces an ending. The ladder describes what the game offers next, never what it requires next. A player who ignores every beat and builds a decorative factory for two hundred hours has not failed at anything. See [14_Win Conditions & Faction Allegiance](<./14_Win Conditions & Faction Allegiance.md>) §1.

2. Two Phases, Three Releases

Decided 2026-09-05. The three-phase plan in [02_Scope & MVP Document](<../02_Scope & MVP Document.md>) (Planet, then Space, then War) is collapsed to two, and the boundary was corrected the same day.

Phase Promise Contains
Phase 1 Build the machine, and reach further with it. Everything except war: the planet layer, the ship and mothership editors, the gate, space, other planets, inter-planet supply chains
Phase 2 Win with the machine. War, and only war. Wars behind the gate, factions, fleet combat, war multiplayer, the Master Rill, mobile

2.1 The seam is war, not the planet surface

The first attempt at this put the planet layer in Phase 1 and everything above it in Phase 2. That was wrong, and the reason it was wrong is worth keeping.

Space belongs with the planet layer, because it is the same activity. You cannot travel between planets without building ships, and you cannot automate a supply chain between them without the editors. A freighter route to a second planet is the factory reaching further: played the same way, judged by the same numbers, solved with the same instincts. The editor's first customer is a hauler, not a warship ([34_Ship & Mothership Editor Design](<../Systems/34_Ship & Mothership Editor Design.md>)).

War is the only thing that is a different game. Shared instances, opponents, attrition, and a supply chain that is being consumed by someone shooting at it. Nothing in the planet or space layers needs any of that, and everything in war needs all of it.

So the boundary sits where the game actually changes character, which is at the war.

2.2 What "two phases" is a statement about

It is a design statement, not a release schedule:

Phase 1 must be complete and satisfying with no war existing at all. Phase 2 may assume every part of Phase 2 exists.

2.3 Three releases across the two phases

Phase 1 is roughly 1,500 to 2,200 hobby-pace hours, which is not a release, it is several years. So Phase 1 ships in two parts and Phase 2 in one:

Release Phase The player gets
The Launch release 1 The planet layer in full, ending in First Launch. Early Access begins
The Space Update 1 The editors, the gate, space, other planets, inter-planet supply chains
The War Update (1.0) 2 War

The release boundary must not become a design boundary. The editor is built once, complete, in the Space Update, including the frames and validation the War Update's combat classes need. That was the whole reason for collapsing the phases and shipping in parts must not undo it.

2.4 The consequence for the Launch release

It gains two things the old plan did not give it:

  1. The first ship launch, as its finale. See §3.5.
  2. Restoration, as its long tail. See §3.4.

It does not gain the ship editor. The Launch release ship is a single prescribed hull built from a recipe, like any other item in the game. The editor opens in the Space Update, and the first thing it lets you do is design something better than the one ship you were given.

3. The Launch Release Ladder

Five beats. Rough hours are for pacing intent, not targets to balance against.

Beat Rough hour The bottleneck What you start optimising Ceremony
First Assembly 1–2 You do not know the verbs Nothing yet. It runs without you No
First Ascent 8–15 Ground space. The answer is rebuilding, not extending Layout Yes
First Blackout 20–30 Power is a system with priorities, not a number Reliability No
First Restoration 30–60 Something outside your factory wants a bill of materials you did not plan for Flexibility No
First Launch 🚀 60–100 The hardest single bill of materials in the game so far All of it, at once Yes

3.1 First Assembly

Trigger: a machine completes a recipe with two or more distinct inputs, both delivered automatically.

Why here. The tutorial ([10_Tutorial & Onboarding Design](<../Interface/10_Tutorial & Onboarding Design.md>)) ends with a line: drill, belt, smelter, chest. A line has no ratio problem. The moment two chains have to arrive at the same machine, the player meets the first genuine factory question, which is how much of each.

What changes. Nothing unlocks. The beat is the realisation, and the notification names it so the player knows the game saw it. This is the first thing that is theirs rather than the tutorial's.

Not ceremonial. The tutorial completion already had a moment, and putting another one ninety seconds later devalues both.

3.2 First Ascent

Trigger: the first building placed on a level other than the starting one, underground or sky ([06_Vertical Layers & Core Power System](<../Systems/06_Vertical Layers & Core Power System.md>)).

Why here. By this point the player has extended outward until extending outward stopped helping. Going up or down is not more of the same: it needs Item Lifts, Fluid Risers and Power Conduits, and cross-level power is capacity-capped, so a level is a budget, not just more floor.

What changes. The player stops optimising throughput and starts optimising layout. The question turns from "can I make more" into "what belongs on which level". Most players rebuild something here, and that rebuild is the point.

Ceremonial. Camera pull showing the factory as a stack rather than a plane. First XP burst of real size. This is the beat where Auto Warz stops looking like every other factory game, and it should be sold as one.

Prerequisite: Branch 4 research (Basic Excavation, Vertical Lift, Pipe Pump). The tree already paces this correctly and needs no change.

3.3 First Blackout

Trigger: demand exceeds supply, the priority waterfall sheds a tier, and the player restores full supply within the same world session.

Why the recovery and not the failure. Running out of power is a mistake. Diagnosing it, seeing which tier was dropped and why, and fixing the cause is the skill. The beat rewards the second thing. Note the deliberate parallel with the war layer's core rule, that a loss must return a diagnosable failure to the factory ([31_War & Combat System Design](<../Systems/31_War & Combat System Design.md>)); this is the planet-side rehearsal of exactly that loop.

What changes. Power stops being a number in the HUD and becomes a system with structure: priorities, hysteresis, storage, and the knowledge that shedding is a designed behaviour rather than a bug. The player starts optimising reliability, which is a different discipline from throughput, and it is the one the war layer will later demand.

Not ceremonial, but it is the beat that most needs good feedback: the notification has to name the shed tier and the deficit, or the lesson does not land. Owned by [18_Notifications](<../Interface/18_Notifications.md>).

3.4 First Restoration

Trigger: the first derelict site in the player's own world brought back to working order ([28_Restoration - Third Endgame Path](<./28_Restoration - Third Endgame Path.md>)).

Why here, and why it matters more than it looks. Everything up to now has been a factory optimising itself. A wreck is the first consumer that is not the player: it names a bill of materials the player did not choose, on a schedule the player did not set, and it will not accept a substitute.

That is a rehearsal for the entire Phase 2 economy. The gate, the Forward Depot and the war all work this way, and Restoration teaches the shape of it with nothing at stake.

What changes. The player starts optimising flexibility: buffers, spare capacity, and the ability to produce something the factory was not built for without tearing it down. Factories built purely for one output hit a wall here, and that wall is instructive.

This is Phase 1's long tail. After the launch there is no space layer yet, so wrecks are what a Phase 1 player comes back for. Restoration is the answer to is Phase 1 enough without a war, and it is a better answer than adding sinks, because it adds demand rather than waste.

Not ceremonial for the first one, because there will be many. The site gets the drama: it arrives with a generated history ([28_Restoration - Third Endgame Path](<./28_Restoration - Third Endgame Path.md>) §3a), and the history is the reward.

3.5 First Launch 🚀

Trigger: the prescribed Phase 1 ship, built, fuelled and launched from the surface.

Why it is the finale. It is the hardest bill of materials in the Launch release by a wide margin, it consumes every chain the player has built, and it is the only object in the game so far that leaves. The player builds a door and watches something go through it. What is on the other side is the Space Update.

It ends a release, not a phase. In the Launch release you do not follow the ship. In the Space Update you do, and it is the same editor you will later use for warships.

This is deliberately Factorio's rocket, with one difference: the rocket is the door to the rest of the game, and the game says so at the moment it launches.

What changes. In the Launch release, nothing mechanical. That is the honest design, and it is fine: the launch is an ending, not an unlock. Pretending otherwise by bolting on a bonus would cheapen it. Once the Space Update exists, the same trigger opens the space layer and the editor.

Ceremonial, and the largest single moment in the game. Camera follows the ship out of atmosphere. Largest single XP award. Achievement. Credits are appropriate here for a player who wants an ending, with the factory still running behind them when they dismiss it.

Requires in the Launch release: a Shipyard building, one hull recipe, and a fuel requirement. Not the editor, not ship classes, not flight.

4. Hour 1, Hour 10, Hour 100

The same ladder, stated as the felt experience, for use in pitch and store copy.

You are trying to You stop when The game is asking
Hour 1 Make one thing happen without you Ore reaches a smelter and a chest fills while you watch Can you make it run?
Hour 10 Make it run fast enough You go up instead of out, and it works Can you make it fit?
Hour 100 Make it run for something else A wreck, then a ship, consumes everything you built Can you make it change?

The three questions are the arc. Run, fit, change.

5. The Ladders After Launch

Sketched, not paced. These systems are designed ([31_War & Combat System Design](<../Systems/31_War & Combat System Design.md>), [32_The Gate & War Logistics](<../Systems/32_The Gate & War Logistics.md>), [33_Mothership & Fleet Design](<../Systems/33_Mothership & Fleet Design.md>), [34_Ship & Mothership Editor Design](<../Systems/34_Ship & Mothership Editor Design.md>)) but their hour-by-hour pacing waits until the Launch release is real and its actual pacing is known.

The Space Update — still Phase 1

Beat The bottleneck What you start optimising
First Design The prescribed ship is the wrong shape for what you now want Trade-offs. The editor's first lesson is that everything costs hull space
First Passage Another planet has what this one does not Routes, and the fact that distance is a cost
First Site One planet cannot hold everything you want to build Where the factory goes, which the game prices and never prescribes ([39_Space Layer Design](<../Systems/39_Space Layer Design.md>) §4)

The War Update — Phase 2

Beat The bottleneck What you start optimising
First Gate Powered Your world is sealed. Power holds the gate open Committing capacity to something outside your world
First War Joined Your supply chain is now someone's ammunition Sustained rate under loss
Master Rill Launched 🛸 Nothing. This is a capstone Nothing new. It is the proof

The character change is visible in that split. Every Space Update beat is a factory problem in a new place. The War Update beats are the first time the problem is someone else.

5.1 Territory moves into the war

The old node-and-zone territory model, with persistent passive income, repair bases and AI factions pressing while the player is offline, does not survive the private-worlds architecture and is withdrawn as a progression beat.

Territory is a faction war thing. It exists inside a war instance, for the duration of that war, and it does not persist into anyone's world.

Holding a forward position during a war is a real mechanic with real value (staging, repair, the Forward Depot's reach). It is not a milestone and it is not a source of passive income between wars. Passive accrual has exactly one source, and it is the Outpost trickle ([32_The Gate & War Logistics](<../Systems/32_The Gate & War Logistics.md>), [04_Progression & Rewards](<../../06_Outpost/04_Progression & Rewards.md>)).

5.2 The Master Rill is the capstone of one path, not the door to the endgame

Previously the entire endgame was gated behind launching the Master Rill. It is now the end of the war path, not the beginning of everything. See §6.

6. Where the Three Endgame Paths Open

The three paths are three different appetites, and gating all of them behind the same object served none of them. They open when their prerequisite exists, which is at three different times.

Path Opens at Release Why there
Path C — Restoration First Restoration Launch release Wrecks are generated in the player's own world. Needs no space layer, no gate and no war
Path B — Supply Chain First Gate Powered War Update The gate is the ability to supply someone else. Before it there is no channel, and the gate is a war slot
Path A — War Effort First War Joined War Update Needs a mothership and a war to join

Two paths open in the War Update, since the gate is what both of them need ([39_Space Layer Design](<../Systems/39_Space Layer Design.md>) §2). The Space Update's open-ended draw is not a path but a place: settling further, which §4 of that document makes a genuinely open design problem rather than a checklist.

Consequences worth stating:

  • The Launch release has an endgame path. Restoration is genuinely open-ended, and it is available to a player who never launches anything.
  • The paths overlap and always did. A player can restore wrecks between wars, and supplying an ally is a way of fighting one.
  • The Master Rill sits at the far end of Path A, not in front of all three.

7. Levels Never Gate World Content

Rule, and it is absolute:

Player level gates nothing inside a world. Levels grant Engineering Points and cosmetics. They never unlock a building, a recipe, a research node, a layer, a beat or a path.

7.1 Why

Levels are account-scoped and permanent ([13_Player Levelling System](<./13_Player Levelling System.md>)). World content is world-scoped and resets. If a level gated world content, then:

  • A new world would not be a fresh start, it would be a partly pre-solved one.
  • A level 40 player and a level 3 player could not meaningfully play the same world, because they would not have the same game in front of them.
  • It would breach [00_Canonical Design Facts](<../../04_Project Management/00_Canonical Design Facts.md>) §20, capacity crosses, stock does not, by letting account progress cross as capability rather than capacity.

7.2 What gates what

Gated by Examples
Factory Research (world) Buildings, recipes, logistics tiers, vertical levels
Materials and time (world) Every beat in §3 and §5
Engineering Points (account) Rates and capacities on ship and mothership designs, and the mothership hulls themselves. Nothing else
Player level (account) Engineering Points, and cosmetics. Nothing else

The one honest exception is admission control: a war may set a minimum Fleet Power to join ([15_Ship & Army Points Design](<./15_Ship & Army Points Design.md>)). That gates entry to a shared instance, agreed between participants, not content inside anyone's world.

7.3 Why the mothership is on the account side

Amended 2026-09-15 (OI-391). The Engineering Points row previously read rates and capacities on ship and mothership designs. Nothing else, which made the mothership's own unlock a contradiction the moment [34_Ship & Mothership Editor Design](<../Systems/34_Ship & Mothership Editor Design.md>) folded motherships into the ordinary hull catalogue. Resolved rather than left standing:

A ship is world content. The mothership is how you show up.

Everything a player builds to use in their own world — freighters, miners, the scout, and the combat hulls they build at home — is gated in-world by Shipyard Mk. Planet-to-planet hauling and supply chains need nothing else.

A mothership exists only to take a player into somebody else's world. It is the vehicle by which they are present at a shared instance, which makes it the same shape as the admission-control exception above rather than a violation of the rule. It is not content inside anyone's world; it is the thing that arrives.

This does not loosen §7. A guest still brings no unlocked buildings, no recipes and no research into a host's world. What account progress buys is a better way to turn up to a war, not a head start on somebody's factory.

7a. Beats in Co-op

Co-op ships in Phase 1 ([04_Data Dictionary](<../../05_Reference/04_Data Dictionary.md>) §0b), so the ladder has to survive more than one player in a world. It does, without a new rule, because of §7's split.

A beat is world state. It fires once, for that world. Everyone present when it fires receives the XP, because XP is account state.

Consequences, all of which fall out of that one line:

  • No race. Two players cannot both earn First Ascent in one world. The world ascended.
  • A guest carries nothing in. A level 40 guest joining a level 3 host's world has Engineering Points, cosmetics and saved designs, and not one unlocked building, because levels never gate world content. The host's factory is exactly as far along as the host built it.
  • A guest carries the XP out. They were there, so they levelled. That is the whole reward for helping someone else build.
  • A player joining a finished world earns nothing retroactively. Beats already fired. This is correct: the ladder measures what a world has done, not what a person has seen.

Open: whether a guest who joins mid-session, after a beat fired earlier that day, should see any acknowledgement of it. Probably a line in the world summary rather than an award.

8. What the Arc Replaces

8.1 Host-set win conditions

[02_Scope & MVP Document](<../02_Scope & MVP Document.md>) lists a Phase 1 session goals system, where the host sets a win condition such as "produce X units of advanced components". It is a leftover from a match-based model that the game no longer is.

The ladder replaces it as the default. Sandbox mode and the host time limit survive as options for players who want them. A host-set production target may survive as a challenge preset, which is a Phase 1 nice-to-have, not a progression system.

8.2 The old Session Progression Arc

[14_Win Conditions & Faction Allegiance](<./14_Win Conditions & Faction Allegiance.md>)'s Session Progression Arc is superseded by §3 and §5. It was also misnamed: it describes the world arc, and no game state belongs to a session ([04_Data Dictionary](<../../05_Reference/04_Data Dictionary.md>) §0).

9. Implementation Notes

Beats are data, not code. One resource per beat, one manager watching for triggers.

# ProgressionBeat.gd — Resource
@export var id: StringName              # e.g. &"first_ascent"
@export var display_name: String
@export var phase: int                  # 1 or 2
@export var ceremonial: bool            # cinematic + large XP + achievement
@export var xp_award: int
@export var opens_path: StringName      # &"" | &"restoration" | &"supply" | &"war"
@export var notification_body: String

# ProgressionManager.gd — autoload
# Beats are recorded per WORLD, in the world save, never in PlayerProfile.
# Reason: a new world must re-earn every beat. See section 7.
# Listens to existing signals rather than polling:
#   recipe_completed(machine)      -> first_assembly
#   building_placed(level_index)   -> first_ascent
#   power_tier_shed / power_restored -> first_blackout
#   restoration_site_completed     -> first_restoration
#   ship_launched                  -> first_launch

Two rules for the implementation:

  1. Beats are world state. Recorded in the world save, not PlayerProfile. A new world starts with none of them.
  2. XP is account state. The award crosses; the beat does not. This is §20 again, and it is the line that keeps both true at once.

Open items for this document are tracked in docs/open-items.md, area progression.

Open Questions

  • Whether First Assembly should fire on any two-input recipe or only on a recipe whose inputs are themselves manufactured, which would push it later and make it a stronger beat
  • The XP values for each beat, once [13_Player Levelling System](<./13_Player Levelling System.md>)'s curve is rebased on real playtimes
  • Whether First Blackout should require recovery within a time window, and how long
  • What the prescribed Phase 1 ship is called, and its recipe
  • Whether the launch cinematic offers credits, and whether declining them is remembered
  • How Phase 2 is released to players: one drop, or space-then-war updates. Belongs to the phasing pass
  • Whether a host-set production target survives as a challenge preset

Revision History

Version Date Changes
0.4 2026-09-06 First Gate Powered and Path B moved to the War Update, since the gate is a war slot and another planet is reached by flying (39_Space Layer Design §2). New First Site beat in the Space Update, whose bottleneck is where the factory goes.
0.4 2026-09-15 §7.2 amended and §7.3 added (OI-391). The Engineering Points row said rates and capacities... nothing else, which contradicted the mothership's own unlock once 34_Ship & Mothership Editor Design v0.5 made motherships ordinary hulls. Mothership hulls are now explicitly on the account side, with the reason stated: a ship is world content, the mothership is how you show up. Everything built for use in your own world is gated in-world by Shipyard Mk; a mothership exists only to take you into somebody else's, which makes it the same shape as the admission-control exception rather than a breach of §7. §7 itself is unchanged and no looser: a guest still carries no buildings, recipes or research into a host's world.
0.3 2026-09-05 Phase boundary corrected: the seam is war, not the planet surface. §2 rewritten as Two Phases, Three Releases: space belongs with the planet layer because it is the same activity reaching further, and war is the only thing that is a different game. §3 renamed The Launch Release Ladder; First Launch ends a release, not a phase, and you follow the ship in the Space Update. §5 split into a Space Update ladder (still Phase 1) and a War Update ladder (Phase 2), which makes the character change visible: every Space Update beat is a factory problem in a new place, and the War Update beat is the first time the problem is another player. §6 paths keyed to releases, one opening in each.
0.2 2026-09-05 §7a Beats in Co-op added. Co-op ships in Phase 1, so the ladder had to answer it: a beat is world state and fires once for the world, everyone present takes the XP, a guest carries nothing in and the XP out. Falls out of §7 without a new rule.
0.1 2026-09-05 Initial document. The bottleneck-ladder principle and the milestone test. Two phases, not three, with the reason the editor cannot be split. The five Phase 1 beats: First Assembly, First Ascent, First Blackout, First Restoration, First Launch, with ceremony on Ascent and Launch only. Hour 1 / 10 / 100 as run, fit, change. Phase 2 beats sketched. Territory withdrawn as a milestone and confined to war instances. The three endgame paths staggered, with Restoration opening in Phase 1 and the Master Rill demoted to the capstone of the war path. Levels never gate world content, stated as an absolute rule with its §20 rationale. Supersedes the Session Progression Arc in 14_Win Conditions & Faction Allegiance.