The North-Star Vision (Three Modes, Bounded Campaign, Living Factions)

Status: VISION LOCKED 2026-07-17 (owner goal, session 634b4cf5). This is the target we build toward.

Last updated: July 18, 2026

Blood Lineage — The North-Star Vision (Three Modes, Bounded Campaign, Living Factions)

Status: VISION LOCKED 2026-07-17 (owner goal, session 634b4cf5). This is the target we build toward.

Parent: index.md (current mechanics). Builds ON, does not replace: progression.md (turns/victory/dynasty), empire-4x-layer.md (post-victory depth, the seven voice-memory rules), return-briefing.md (entry-time push), diplomacy.md (AI personalities → factions), military.md (combat → the war loop). Voice standard: ../../philosophy/player-experience.md.

Why this doc exists

Blood Lineage is the most mechanically ambitious game in the catalog — a genuine turn-based 4X in voice: seven resources, twelve units, fifteen buildings, sixteen spells, five victory conditions, a generational dynasty (aging, succession, heirs, cadet branches, inheritable traits), diplomacy with seven AI archetypes, a twenty-three-tech tree across five eras with prerequisite chains, religion, thirteen random events, seasons, AND a shipped post-victory Empire / NG+ / Chronicle layer. The depth is real and mostly works in isolation.

It has also not had focused love in a while, and the cost shows in one sentence: the systems are powerful but the experience is not yet meaningful. Power without shape is noise. This doc locks in the shape.

The evidence base (validated 2026-07-17)

This vision is not a hunch. It is grounded in the live data and a firsthand play session.

Where players actually are (lineage.lineage_game_map, active single-player games):

Turn reached	Games	Completed
1–5	103	0
6–15	11	0
16–30	11	0
31–60	18	3
61–100	10	4
100+	18	13

Two facts jump out. ~60% of single-player games die at turns 1–5 — a brutal onboarding cliff. And the dedicated core who survive finish around turn 100+ (13 of 18 games past turn 100 completed). The owner's instinct — "maybe 100 rounds?" — is exactly right: ~100 turns is already the natural length of a satisfying Blood Lineage game. We are not inventing a target; we are naming one the data already drew.

What the engaged players say (25 distinct players, 452 feedback rows; one 204-row whale plus ~7 players with 15–56 rows each — a small, devoted audience). The recurring themes, verbatim:

War does not resolve — "I captured the enemy's capital but they did not surrender"; "the AI kingdoms never attacked back"; "I keep marching to these capitals... it's like they're moving, I can't reach them"; "it won't tell me where this kingdom's capital is so I can attack it."
Army control is opaque — "I cannot recruit units into army 2" (many times); "can't control more than 12 army"; "I don't know how to garrison armies on restless territories."
Diplomacy is hollow — "they made alliance but aren't sending military to help"; "proposed trade but no trades come"; "declared war and it went on the wrong kingdom."
The dynasty feels mechanical, not alive — "can the patterns be erased?"; "only like 20 names in the whole system"; "3rd generation aren't having children, I'm going to be stuck"; "the brothers and sisters live forever."
They want the vision — "is there a way to get into a world with real kingdoms and AI with other real players' kingdoms?"

Firsthand play confirms it (live prod session, 2026-07-17):

Onboarding says "recruit units," but recruit 10 warriors → "You need to create an army first" — an undiscoverable prerequisite. That single step is the turn-1–5 cliff.
declare war → "War is declared on Shadowmere" (blind-targeted, no choice), then where is the enemy capital → NLU fallback. You can declare a war you can neither aim, find, reach, nor end.
what should I do next → "That command is not available"; recommend an action → NLU fallback. No advisor exists.
Turns run indefinitely toward a threshold. There is no arc, no target, no decision-driven ending.

The diagnosis writes itself: each system runs; the loops between them are broken, and the whole has no shape. Fixing that — not adding more systems — is the work.

The one-line vision

Blood Lineage is a voice-first grand-strategy game of dynasty and realm offered in three modes that share one engine: a bounded, decision-resolved Campaign you return to and finish; an Eternal Realm you build, grow, and rank against others forever; and War of the Realms scenarios where storied factions fight a board with a clear objective. In every mode the world is alive with named factions that want things, and every player — expert or newcomer — can either drive the story themselves or lean on a steward's counsel and still finish with honor.

Three design commitments hold it together:

Meaningful over merely powerful. Every system must close a loop the player can feel — declare war → find the foe → march → take the capital → they sue for peace. Depth that does not resolve is removed or fixed, never shipped as "technically works."
Shape over sandbox. Play has a beginning, a rising struggle, and a resolution the player's choices author. A game you can finish is a game worth restarting.
Two on-ramps, one game. The complexity is structured and dynamic. A steward can recommend a solid next move to anyone; a player who reads the factions and times their wars earns far more. Both are first-class. This is how a genuinely deep 4X survives on a device with no screen.

All three sit on top of the existing engine and the seven voice-memory-load rules already codified in empire-4x-layer.md. We add uses of numbers players already know; we do not mint parallel systems. That restraint is the whole reason a voice 4X can get deeper without getting heavier to hold in the head.

The three modes

The three modes are the three answers to "what kind of session do I want?" They reuse the same kingdom, combat, dynasty, diplomacy, magic, and research code. They differ in shape, pacing, and stakes — the same way Total War offers a campaign, a custom battle, and a persistent multiplayer.

	The Campaign	The Eternal Realm	War of the Realms
Fantasy	"Tell the story of my dynasty"	"Build a kingdom I keep and grow"	"Win a fight on a set board"
Reference	Total War campaign	Clash of Clans	Total War custom / MP battle
Shape	Bounded, ~100 turns, 3 acts, resolves	Persistent, never ends, ranked	Scenario with a clear win objective
Players	Solo vs AI factions	Solo build + async vs real kingdoms	Solo vs AI, or MP vs real players
Existing mode it grows from	Local (single-player)	Global (1000×1000 persistent)	Local + Multiplayer
Win state	Chosen victory path OR turn-~100 reckoning	No win — power ranking + longevity	Bespoke per-scenario objective
Current state	Shapeless sandbox	7 kingdoms, no growth/compare loop	Does not exist as a framing

Mode 1 — The Campaign (the heart): a bounded, decision-resolved dynasty

The single-player game becomes a Campaign: a set interactive struggle with a beginning, a rising conflict, and an ending the player's choices author — something you return to across sessions and finish, then start again differently. The return briefing already handles "keep coming back"; this mode gives the thing you come back to a shape and a destination.

1a — Length target: The Century (~100 turns), soft not hard

The campaign targets ~100 turns — validated as the natural length by the completion data. It is a target, not a wall:

Reaching a chosen victory path (below) ends the campaign early, in triumph, at any turn.
If turn ~100 arrives with no decisive winner, a Twilight of the Age reckoning resolves the campaign by scoring every realm on the player's accumulated choices (territory, gold, alliances, tech, dynasty prestige, faction standing) and narrating which dynasty history remembers as ascendant.
Difficulty and scenario length can scale the target (a "Short Reign" ≈ 50 turns, an "Age of Legends" ≈ 150) — but ~100 is the default the whole arc is tuned around.

1b — Three acts (a spine over the existing turn pipeline)

The acts are pacing and narrative framing over the same loop — not new mechanics. Each act shifts faction pressure, advisor prominence, and the beats that fire.

Act	Turns	Feel	What changes
I — Founding	1–25	Guided, forgiving	Advisor is prominent; factions are cordial; the arc ends when "the realm takes notice" — one rival faction marks your dynasty as a threat. This act is the fix for the turn-1–5 cliff.
II — Ascendance	26–70	The open struggle	Factions actively pursue their own agendas (war, alliance, expansion, tech races). A mid-act Realm Crisis (a shared threat, a dominant rival, or a dynastic turning point) forces a strategic pivot. This is the meat.
III — Reckoning	71–~100	Escalating, decisive	The campaign drives to resolution: secure your victory path before the strongest rival secures theirs. Ends in triumph, defeat, or the Twilight reckoning.

Acts are re-readable on demand ("what act am I in / what is my situation?") and never a wall the player must remember — the game holds the state; the player asks.

1c — Resolution authored by decisions (the "story-driven" ask)

The campaign silently tracks the player's path from their actions — Conqueror, Diplomat, Steward, Archmage, Dynast — the same signals the five victory conditions already read. The ending is not a bare "you won / you lost"; it is an epilogue that narrates how the dynasty is remembered, shaped by that path and by the pivotal choices logged in the Chronicle (empire-4x-layer.md SE-3 already persists beats). Winning by conquest, by gold, by alliance, or by the Twilight reckoning each tells a different story. That is what makes the world feel alive and the ending feel earned.

1d — Difficulty

Selectable at campaign start and re-readable: number of rival factions, their aggression and competence (the existing AIPersonality.difficulty knob), the player's starting position and headstart, and the length target. NG+ (already shipped) makes each subsequent campaign a chosen legacy, not a reset.

Mode 2 — The Eternal Realm (Clash of Clans): build, grow, compare, forever

The Global world becomes The Eternal Realm: a persistent kingdom you nurture across weeks — build it up, reshape it, defend it, and see where you rank among every other realm. It has no ending; the reward is growth and standing. Today it holds 7 kingdoms and has no growth-and-compare loop — that loop is the whole point of this mode.

The mode grows from the existing Global infrastructure (weekly turns, action economy, protected base territory, 1000×1000 map) and adds the three things that make a builder-ranker sing:

A kingdom worth returning to. Your realm persists and visibly grows — more territory, taller buildings, a mightier standing army, an older and more storied dynasty. The return briefing narrates what grew since you were away.
Comparison — the Realm Ledger. A speakable power ranking: "You are the 14th realm of the age. The Iron Compact holds the throne of power. You rose two ranks this week." One number (Realm Power = territory + buildings + army + prestige, numbers the player already has), a rank, and a trend. The single most important missing piece — Clash of Clans is nothing without the leaderboard.
Living stakes. Raid and be raided on expansion territory (base stays protected — already true); defend, retaliate, ally. Other realms are real players and long-lived AI factions, so the world is populated even when the human count is low.

Voice-memory: everything is pull-on-demand ("realm ledger," "who raided me," "my rank") plus the weekly return-briefing push. No new currency; Realm Power is a rollup of existing numbers.

Mode 3 — War of the Realms (Total War III): factions, a board, a clear objective

A Scenario is a pre-set board — fixed starting positions, a chosen faction, a specific map — with a clear, stated win objective and a difficulty. It is the "I want a focused fight with a point" mode, single-player against AI factions or multiplayer against real players, all on our existing turn/combat/ diplomacy engine. This is the framing that does not exist today.

Pick a faction (below) — each with its own identity, bias, and flavor.
A scenario names its objective up front and re-reads it on demand. Examples, all expressible in existing mechanics: "Unite the Broken Coast — control the three crown territories by turn 40," "Break the Iron Compact — eliminate their dynasty," "Hold the Ashen Gate — survive twenty turns of siege," "The Merchant's Gambit — reach 60,000 gold before the League does."
Difficulty — enemy faction count, aggression, competence, and player handicap.
Multiplayer reuses the existing shared-map multiplayer service; AI factions fill empty seats so a 1-human scenario is still a full board.

A scenario is a Campaign with the sandbox cut away and a single sharp objective bolted on — same engine, different shape. It is also the cleanest home for the "real kingdoms and other real players" the engaged player explicitly asked for.

Living factions (the story flesh that spans all three modes)

Today the world's rivals are seven generic AI archetypes (AGGRESSIVE, PEACEFUL, DIPLOMATIC, SCHOLAR, MERCHANT, MYSTIC, BALANCED) wearing procedurally-generated names. They compute, but they do not mean anything — and players feel it ("the patterns are boring"). The vision replaces anonymity with a roster of named, storied factions that recur across every mode and give the world its life.

Each faction carries:

Identity — name, sigil, motto, homeland, a one-paragraph lore and agenda (what they are trying to achieve — so the world is full of competing ambitions, not idle AI).
Archetype + bias — mapped onto an existing AIPersonality archetype, plus a signature leaning (a favored unit line, tech track, or diplomatic posture) that makes fighting or allying them feel distinct.
Stance toward the player — ancient enemy, wary rival, or courtable ally — seeded per mode.

An illustrative starting roster (names are placeholders for a design pass, but the shape is the point):

Faction	Archetype	Homeland & agenda
The Iron Compact	Aggressive	Mountain war-clans; conquest and iron tribute. The default rival.
The Gilded League	Merchant	Coastal trade-cities; wealth and monopoly over war.
The Ashen Covenant	Mystic	An arcane order; legendary magic and the Time-Stop victory.
The Verdant Throne	Balanced/Peaceful	Agrarian heartland; endure, grow, out-last.
The Pale Court	Diplomatic	Schemers; win the realm through alliances and treachery.
The Sunfire Dynasty	Scholar	Old blood and older libraries; the technological ascendancy.

The same factions are the Campaign's rivals, the Scenario's playable and enemy sides, and the long-lived AI realms in the Eternal Realm — so investing in faction identity pays off three times. This is also the concrete answer to the dynasty-feels-mechanical feedback: a wider name pool, distinct faction voices, and agendas that visibly differ turn to turn make the world read as alive and random instead of patterned.

The advisor: structured complexity for everyone (the two on-ramps)

The single highest-leverage addition. A Steward the player can consult at any time — "what should I do," "advise me," "counsel," "what do you recommend" — who reads the current state and recommends the highest-value next action, bound by the voice-memory rules (≤3 recommendations, pull-on-demand, each short, ends with an actionable handoff).

It is the structured on-ramp: a newcomer or a returning player can lean on the steward and play a solid game without mastering the vocabulary — the direct fix for the turn-1–5 churn cliff and the "what should I do next → not available" dead-end.
Driving the story yourself — reading the factions, timing wars, exploiting rivalries, min-maxing the dynasty — is the dynamic on-ramp and earns far more. The steward plays safe and solid; the player who ignores it and plays sharp wins bigger and faster. Both are legitimate ways to finish.
It is context-aware: in Act I it teaches the loop ("raise your first army, then claim land"); in a war it points at the reachable enemy capital; on a food deficit it says fix it first. The advisor is where "structured and dynamic complexity" becomes real instead of a slogan.

Fixing the loops (the meaningful-over-powerful backlog)

The vision cannot be reached on top of broken loops. These are the loops the evidence proves are broken today; they are prerequisites, not polish. (Each becomes a tracked fix through the standard find→validate→fix pipeline — this doc names them; it does not fix them.)

Loop	Broken symptom (evidence)	What "meaningful" requires
Onboarding → first action	"recruit units" → "create an army first"; 60% churn at turn 1–5	The guided path (and the advisor) leads a new ruler from naming → workers → first army → first territory without dead vocabulary.
War → resolution	Can't aim war, can't find/reach the enemy capital, capitals "move," no surrender	Declare war names a real target; the map speaks where the foe is; armies can march to and assault a capital; taking it resolves (surrender / vassalage / annexation).
Army command	"can't recruit into army 2," "can't control >12," "can't garrison"	Raising, naming, reinforcing, garrisoning, and moving multiple armies is discoverable and speakable.
Diplomacy → follow-through	Alliances send no aid; trades never arrive; war hits the wrong kingdom	Treaties do what they promise; allies send force; trades deliver; every diplomatic verb names its target.
Dynasty → living continuity	Patterned births, ~20 names, gen-3 dead-ends, immortal siblings	Wider name pool, less-patterned births, cadet lines that continue, mortality for all — the dynasty reads as alive.

Difficulty & win conditions across modes (summary)

Campaign — chosen victory path (the five existing conditions) or the turn-~100 Twilight reckoning; difficulty = faction count / aggression / competence / start / length.
Eternal Realm — no win state; Realm Power ranking + longevity; "difficulty" is the live competition of other realms.
War of the Realms — bespoke per-scenario objective, re-read on demand; difficulty = enemy count / aggression / competence / handicap.

Every win/objective surface obeys the no-dead-end and re-readable rules: the objective can always be spoken back, and every screen ends with an actionable handoff.

Build path (phased, ADD vs FIX)

Sequenced so each phase is playable and the highest-leverage, lowest-risk work comes first. This is the roadmap; each item lands as its own tracked epic/fix under the normal deploy gate — this doc does not ship code.

Phase 0 — Fix the loops (prerequisite). War resolution, army command, diplomacy follow-through, onboarding path, dynasty continuity. Nothing below feels good until these do. Highest priority.

Phase 1 — The Advisor. The steward / recommended-action layer. Highest leverage: it rescues onboarding and delivers the "structured vs dynamic" promise, and it is largely a read-and-recommend layer over existing state.

Phase 2 — Campaign shape. The Century length target, the three acts, and the decision-authored Twilight resolution + epilogue — a framing layer over the existing turn pipeline and Chronicle.

Phase 3 — Living factions. The named-faction roster (identity, agenda, bias) replacing anonymous archetypes across all three modes; the wider name pool and less-patterned dynasty.

Phase 4 — The Eternal Realm loop. The Realm Ledger power ranking + the build-grow-compare loop on the existing Global infrastructure.

Phase 5 — War of the Realms scenarios. The scenario framing (faction pick, stated objective, difficulty) single-player, then multiplayer.

Phases 0–1 are the difference between "a deep game nobody can get into" and "a deep game anyone can enjoy," so they come first. Everything after is turning a good game into a great one.

What this vision is NOT

Not a rewrite. Every mode reuses the existing engine; we add shape and story, we do not replace mechanics. The owner was explicit: build on top of OUR system, ADD to what we have.
Not a new set of systems to remember. Bound by the seven voice-memory rules — depth comes from new uses of numbers players already know, surfaced pull-on-demand.
Not a spec for any single feature. Each phase gets its own detailed design + epic. This is the north star they aim at, and the bar they are measured against.

eternal-realm.md — Phase 4, IMPLEMENTED: the Eternal Realm build-grow-compare loop, Realm Power, the Realm Ledger, and AI-faction board population.
index.md — current mechanics (authoritative for WHAT exists).
progression.md — turn pipeline, the five victory conditions, dynasty, events.
empire-4x-layer.md — the seven voice-memory rules; post-victory Empire / NG+ / Chronicle depth this vision's endings and factions compose with.
return-briefing.md — the entry-time "keep coming back" surface every mode uses.
diplomacy.md — AI personalities that become the living factions.
military.md — combat + movement, the war loop Phase 0 must make resolve.
../../philosophy/player-experience.md — voice-first, never-dead-end.