FTL Endgame - Heliomance/Uncharted-Horizons GitHub Wiki

_{Design document §4}

Rewritten in v3.112 from the FTL endgame redesign (the direction document's W1–W19). Two agreed rules that belong to the discovery model — Grand Endeavours priced at 2× as insurance, and proximity revelation (a 10% chance each turn that an eligible idea which fails to occur is glimpsed) — are written into §1.2 and §1.5 with that model, a deferred task; this section relies on them.

To win a Science Victory a civilisation builds and launches a ship with an FTL Drive. Exactly one way of outrunning light works in any world, chosen secretly at world creation from the five its ontology offers (FTL in Elemental Worlds, below). The player finds out which by investigation: each candidate is a hypothesis whose conditions Space-era experiments confirm or rule out, and the Drive can be built only from a successful prototype of the viable method.

There are no false positives: a gamble's risk comes only from what the player has not yet tested, never from evidence that lies. Certainty is information, not a gate (W2): a civilisation may commit to a prototype before it is certain — the Hail Mary — and the finale is decided by judgement, never by dice.

The evidence comes only from experiments humanity has not yet done (W1): every condition is a Space-era result, while Atomic-era rows keep their real content and find what ours found.

• The puzzle has its own screen (W6), separate from the research screen: the hypotheses, a running list of results, hints of where to look, what has been disproved, and commitment to a prototype.
• The puzzle opens with RC01 Relativity, which establishes the light barrier; before it, going faster than light is no special achievement, and nothing is proposed. A hypothesis whose proposing row comes first waits for it.
• Each hypothesis is proposed by a CORE theory row, as each was in real history long before anything could test it: Warp, Wormhole and Tachyonic by RC01 Relativity (special relativity admits the faster-than-light particle, general relativity the warp and the wormhole; both theories have been one breakthrough since v3.124, so the three are proposed together), Hyperspace by MA13 Topology, Tunnelling by QP02 Quantum Mechanics Formalism (atomic worlds), Plenum by QM-E04 Unified Field Theory (elemental worlds). Each world proposes exactly its own five. Proposing a hypothesis Reveals it — the point from which commitment is possible.
• Hypotheses are named for what they would do, not in physics terms (the player-facing names are spec work). The game may say how many there are and name them; it never says which is true.
• The unified theory closes the list — QP08 Quantum Gravity Theory, or QM-E27 Plenum Gravitation, at the start of the Space era: every consistent completion of the theory admits exactly one way to outrun light, the candidates are these five, and which holds depends on properties not yet measured. Before it, the board makes no claim that any proposal is true or that the list is complete. Each unified theory is built on its ontology's great instrument (QP04 Particle Accelerator / QM-E08 Boundary Furnace, both CORE), so it arrives with its first tests nearly ready (W17); MA13 Topology is among QM-E27's prerequisites, as it is in QP08's chain, so every proposal arrives before the list is closed.
• Every change to the screen is announced. A discovery that changes anything on the FTL screen — a proposal, a condition's result, a build need met, the closing statement, a hint moving — carries a line in its notification, such as "This bears on the question of how to exceed the speed of light." It is computed from the screen's state before and after the discovery, and says only that something changed, never what it means.

• A checklist of questions and needs, deliberately. Each hypothesis lists its open questions and its build needs in plain words; the puzzle lies in their order and in when to commit (W12), not in finding them.
• Every hint points to where progress is possible now. For each open question or build need, the screen follows the answering breakthrough's prerequisites back to the first missing step whose own prerequisites are met, and names that step's focus — each focus where several steps can be worked on at once, each route's foci where there are alternatives. As research advances the hint moves, so following a hint in good faith is never a dead end. The research screen complements it: each focus says, vaguely, what it tends to help with (W19; mechanism and wording open).
• The breakthroughs, and how far away each answer lies, stay hidden. No breakthrough is named, nor its prerequisites, until it is Revealed or discovered in the normal way; a hint naming an unexpected focus shows only that the answer is more than one step away, never how many. The one exception is an instrument (Conditions and Their Tests): once a test's breakthrough is discovered, the screen names the building that will settle its question.
• No odds, and no reasoning beyond one step. The screen applies only direct consequences: a failed condition disproves every hypothesis that needs it; a hypothesis with all its conditions confirmed is confirmed; the last one standing is marked as the only one left.

(1) Show facts, not advice: the screen and the reports state what is known and its direct consequences, and where each open question can be investigated; they never rate likelihood, recommend a test, or say which hypothesis to pursue — standing is factual, never "promising" or "likely". (2) Use the player's vocabulary. (3) Negative results are clearer than positive ones: a dead end is signposted plainly. (4) Several signals are needed to be sure — met by the conditions' structure, not by noise in reports. (5) No outside knowledge is required: a thoughtful player who follows the screen can always reach a victory.

A method is viable only if all its conditions hold, each a Space-era result (W14). A single condition can hold in worlds where its method fails, so a positive raises that method's standing without settling it; any failed condition rules the method out. Suggestive evidence (W5) and the need for several signals come from this structure, not from randomness in reports. Conditions marked merged are ingredients of the method's own build, tested by building them.

Method	Atomic conditions	Elemental conditions
Warp	Exotic matter holds (merged); extra dimensions exist; the bubble is stable (merged)	Exotic matter holds (merged); higher strata or a higher-order cycle; the bubble of shaped Aether holds (merged)
Wormhole	Exotic matter holds (merged); closed timelike curves are stable; the throat is stable (merged)	Exotic matter holds (merged); closed circuits of the medium are stable; the throat is stable (merged)
Hyperspace	Extra dimensions exist; gravitational leakage shows they are open; dark matter is a hidden sector	Higher strata or a higher-order cycle; leakage into the strata shows they are open; dark matter is a hidden sector
Tunnelling	Macro-scale coherence (merged); spacetime is discrete; tunnelling carries information faster than light — the Hartman effect (merged)	—
Tachyonic	Dark energy is a tachyonic field; closed timelike curves are stable; a tachyonic sector exists	The medium carries a superluminal mode; closed circuits of the medium are stable; the catalogue contains a superluminal form
Plenum	—	The currents run faster than light (merged); the medium carries a superluminal mode; the currents can be ridden (merged)

Exotic matter — QP11 Exotic Matter Synthesis / QM-E06 Exotic Matter (Elemental); extra dimensions or strata — QP09 String Theory Validation / QM-E12 Higher Strata Theory; discreteness — QP10 Loop Quantum Gravity, which excludes extra dimensions; closed curves or circuits — RC07 Closed Timelike Curves; leakage — RC11 Gravitational Leakage Survey, which can hold only where extra dimensions or strata exist; dark matter — RC05 Dark Matter Identification; the tachyonic field or superluminal mode — RC06 Dark Energy Understanding; coherence — CP21 Macroscopic Quantum Control, or GN10 Bio-Quantum Coherence only where FQ6 (Biological Depth) makes the biological route real; the currents' speed — QM-E19 Celestial Current Survey; bubble and throat stability — MS08 Exotic Matter Engineering (only where exotic matter exists); the Hartman effect — QP14 Quantum Communication; a tachyonic sector or superluminal form — QP15 Beyond the Standard Model, or QM-E28 Beyond the Catalogue in elemental worlds; riding the currents — AE20 Current Rigging. QP04 Particle Accelerator and QM-E08 Boundary Furnace settle no condition: they are the great instruments the unified theories are built on.

Each deciding row's prerequisites include the missing piece that explains why the question is open in our world, which keeps every decider beyond our present for a real reason and makes the order of testing a real choice. The missing pieces are of two kinds:

• Theoretical at root — the missing piece is the unified theory, which gates them: RC07 Closed Timelike Curves (whether they survive depends on quantum effects in curved spacetime — chronology protection); QP11 Exotic Matter Synthesis (the Casimir effect gives tiny negative energy densities, and quantum inequalities constrain anything more; it also needs QP12 Casimir Effect Amplification); QP09 String Theory Validation and QP10 Loop Quantum Gravity (neither theory has yet made a prediction we can reach). For these, the breakthrough is the answer.
• Blocked by measurement — the missing piece is an instrument. The breakthrough is the design; the construction is the measurement. Each such test unlocks its instrument, a national wonder every civilisation may build for itself (construction about 3–6 turns), and the answer arrives when it is completed:

Test (what it settles)	Instrument it unlocks	Real-world model
CP21 Macroscopic Quantum Control (coherence survives at large scales)	Orbital Quantum Laboratory — massive superpositions in free fall	ESA's MAQRO proposal
RC11 Gravitational Leakage Survey (extra dimensions are open)	Space Interferometer — a gravitational-wave detector in orbit	LISA, planned for the 2030s
RC05 Dark Matter Identification (dark matter is a hidden sector)	Survey Telescope — lensing maps and the expansion history	Euclid, which measures both
RC06 Dark Energy Understanding (dark energy is a tachyonic field; the medium carries a superluminal mode)	the same Survey Telescope	DESI; Euclid
QP15 Beyond the Standard Model (a tachyonic sector exists); elemental QM-E28 Beyond the Catalogue (a superluminal form)	Very Large Particle Collider — a next-generation collider; elemental: the Very Large Boundary Furnace (+6 Science, 2 Scientist slots each)	the proposed Future Circular Collider

Either RC05 or RC06 unlocks the Survey Telescope, and once built it answers each of the two questions whose breakthrough the civilisation holds; RC05 and RC06 are researched alongside each other, not one behind the other. Once a test's breakthrough is discovered, while its question is open and the instrument not yet built, the FTL screen says so — "To settle this unanswered question, we'll need to build a Space Interferometer" — naming only the building the civilisation has just learned: the puzzle stays suggestive about what to research, but is explicit about what to build, since building is a deliberate act rather than a discovery. Instruments are visible, so rivals can see what a civilisation is testing; a civilisation that cannot build its instrument cannot settle that question.

For tests that are theoretical or merged into a build, the answer comes with the breakthrough however it arrives — discovery, diffusion, trade or theft. For measurement tests, the receiver must still build its own instrument to learn the answer. Knowledge diffusion thus acts as catch-up for laggards and never helps the leader.

The deciding rows cost 1× their era's price, so theory is the cheaper phase; so do the build rows that host a condition (CP21, MS08, AE20). Engineering rows in the signature fields cost 2.5× and their capstones 4× (the multipliers are marked on each row, §1.2); the Space era's base price is set in §1.2.

A method's build list is the theory and engineering to construct it if it is real; its prove list is its conditions outside the build (W4). Some ingredients are their own tests (merged). Balance is counted in Space-era rows, since earlier rows are already researched when the investigation begins.

Method	Build (atomic / elemental)	Prove	Space-era rows
Warp	RC01 Relativity; QP08 / QM-E27; QP11 / QM-E06 (merged); MS08 Exotic Matter Engineering (merged); power; ES18 Superconducting Energy Storage (after EL11 Superconductivity and EL07 Power Grid)	QP09 / QM-E12	6
Wormhole	As Warp without power or ES18, plus RC04 Black Hole Thermodynamics, QP12 Casimir Effect Amplification / QM-E13 Aetheric Vacuum, MA13 Topology, and MS17 Transformation Media	RC07	6
Hyperspace	QP08 / QM-E27; MA13 Topology; power; MA19 Higher-Dimensional Geometry (after MA15 Optimisation Theory)	QP09 / QM-E12; RC11; RC05	6
Tunnelling (atomic)	QP02 Quantum Mechanics Formalism; CP10 Quantum Computing (after CP16 Planetary Data Networks); MS02 Nanotechnology or GN05 Synthetic Biology; QP14 Quantum Communication (merged; needs QP07 or GN10); CP21 Macroscopic Quantum Control (merged)	QP10	6
Tachyonic	QP06 / QM-E17; QP08 / QM-E27; QM-E05 in elemental worlds; SE02 Orbital Assembly; power; NP15 Superconducting Storage Rings in atomic worlds (after EL11 and a power plant), TH20 Boundary Ring / Cycle-Limit Ring in elemental ones	RC06; RC07; QP15 / QM-E28	5 atomic / 6 elemental
Plenum (elemental)	QM-E27; MS02 Fine Structuring; QM-E19 Celestial Current Survey (merged); QM-E23 Current Sail (after AE14 Aerospace Industry); AE20 Current Rigging (merged)	RC06	6

A fully biological route to Tunnelling stays open but rare: QP14 accepts GN10 Bio-Quantum Coherence, MS02 may be replaced by GN05, and CP10 accepts GN10 — through GN08 Bio-Computing, a RARE gateway, and only where FQ6 allows.

Where a method needs a great deal of power the game measures it, without a new resource. Only the brute-force methods — Warp, Hyperspace and Tachyonic — need it ("this requires considerable amounts of energy"): the prototype needs generating capacity worth one advanced fission plant, counted from the grid plants the civilisation owns, weighted by era through Unciv's existing countables, and shown only on the FTL screen. Modern-era plants count 1 point; Atomic-era plants (ES06 Nuclear Fission Plant, ES07 Solar Panel, ES15 Quintessic Plant) 3; ES14 Advanced Fission 8 — the requirement; ES08 Fusion Reactor, QM-E05 Re-Tempering / Self-Feeding Cycle Reactor and the RARE compact and cascade reactors 10. Power is flavour, not a balance lever.

Each method's engineering lies in one field's Space-era pool: Warp in Energy Systems, Wormhole in Materials Science, Hyperspace in Mathematics, Tunnelling in Computing, Tachyonic in Nuclear Physics (atomic worlds) or Thermodynamics (elemental), Plenum in Aeronautics. The pools are optional content: a science player meets the FTL requirements, which stay light, and is then free to explore any field; incidental discoveries along the way are part of the race, and the pools stay open during it. The evidence fields (Quantum Physics or Quintessic Physics, and Relativity & Cosmology) and the ship fields (Rocketry, Space Engineering) are shared by every method.

(1) no single result settles a method; (2) every method has at least one condition outside its build list, so a Hail Mary is always possible; (3) conditions are shared wherever the physics allows, so choosing a test is a real decision; (4) shapes vary, and match each method's lane (W13: the Engineering, Renaissance, Theorist's, Information Age, Big Science and Elemental paths); (5) total time to victory is balanced across methods and ontologies, through the build and prove lists — to the standard that no method gives an advantage or disadvantage larger than ordinary game-to-game randomness, so a world whose method is Warp does not make it markedly harder to beat a diplomatic or cultural player; the prototype's cost is flat, and power is flavour; (6) confirming is the normal path — waiting to confirm is the better bet when no rival is close, while a right gamble skips enough confirmation to win a close race; (7) the theory phase is judged by how it plays — how soon the first answer comes, whether the narrowing is steady, how long the silences are — not by a length target.

A STANDARD idea whose moment passed comes back when something waits on it: a breakthrough it alone blocks, an FTL prototype, or an open question on the FTL screen (rediscovery on demand, §1.2). Whoever needs a row gets it, without the row being CORE. An idea nothing waits on stays lost, and UNCOMMON, RARE and UNIQUE alternatives keep their odds. Grand Endeavours (2×) are the deliberate route to any revealed goal, its idea occurred or not. Knowledge diffusion recovers old rows long before the race begins. With rediscovery every measured game finishes, isolated civilisations included. (W16, which held a Space-era row's window open while its focus was funded, was dropped in v3.120: with rediscovery it added only speed, about 8 turns for the expert, and its written form did nothing measurable for the simulated players.)

Every science victory needs a working prototype drive (W2): a Grand Endeavour on the FTL screen, committable once its hypothesis is Revealed and its build prerequisites are discovered, whether or not the method has been verified. Committing before verification is the Hail Mary, a risk a player chooses — typically one who expects to lose the race otherwise.

• It costs 5× the Space era's price (about 5 turns on a good economy, 6 on an average one) — the same for every method — and runs its full length, then reports whether it works. A failed prototype proves its method invalid.
• It costs more than any experiment that could disprove its method, so a prototype is never the cheaper way to test a hypothesis — only a bet on finishing sooner.
• The ship's FTL Drive requires a successful prototype: one condition, the viable method's prototype researched. No failed prototype can satisfy it.
• Six prototype Endeavours exist, one per method (Tunnelling's and Plenum's in their own worlds only), and live on the FTL screen, not in the research tree.
• The active/spread split holds. A prototype takes the active focus's share (80% once SO05 Specialization of Labor is discovered, 50% before), and the spread still funds every other focus with work under study, so evidence keeps arriving during a build — possibly the experiment that disproves its method. Committing early gives up directed testing, not all testing.
• A failed prototype returns nothing but its proof: a refund would erode the rule that failure costs more than the disproving experiment.
• A prototype halts if an experiment disproves its method mid-build, and its progress is lost — being overtaken by evidence is part of the gamble's risk. A disproved hypothesis cannot be committed.
• Completing the prototype, and later the SS FTL Drive, is announced to every civilisation, as base Civ V announces spaceship parts; committing to a prototype stays private.
• The AI commits when certain, and gambles only as a Hail Mary: when a rival's ship looks close to launch and its best hypothesis has decent odds. (The AI's detailed policy belongs to the AI-automation task.)
• The finale is decided by judgement: wait, and win by commitment; commit early, and take a risk of your own choosing — never dice.

At world creation (W14): (1) the viable method is chosen uniformly from the ontology's five. (2) Its conditions hold; every other condition holds or fails at even odds, subject to extra dimensions and discreteness being exclusive, leakage holding only where extra dimensions or higher strata exist, and every other method failing at least one of its conditions — in practice, a uniform draw from the valid world truths. (3) The guarantee, checked as a battery: exactly one viable method; no method's conditions inside another's; and the viable method's build and prove rows always reachable, whatever the world's other truths — QP12 and QM-E13 always work as science (FQ4 is read at ES12 and QM-E07); fusion-free worlds keep advanced fission; GN10 is only ever an alternative; no Removal Truth touches these rows. MS08's two results exist only where exotic matter does, which matters only to methods exotic matter has already ruled out.

Reliability: MS02 Nanotechnology / Fine Structuring is CORE, QM-E06 accepts MS02 as well as its chemistry routes, and the rows that host a condition inside a build cost 1×, so a lost idea rarely leaves a world unwinnable; Grand Endeavours and proximity revelation recover the rest (§1.2, §1.5).

Section 4 continues on FTL in Elemental Worlds and How the Victories Are Built. The decisions behind it are summarised in design/notes/ftl_decisions_digest.md in the repository.