Changelog

All notable changes to Grand Century are documented here.

[Unreleased]#

1.6.1 - 2026-08-24#

• Added local, era-styled flags for all thirteen polities introduced by the 1820 reconstruction. The nation browser and HUD no longer request missing flag assets.
• Added a release gate that requires every playable polity to have a local flag before the historical seed can ship.

1.6.0 - 2026-08-24#

The world of 1820#

Grand Century now begins in 1820, after the Napoleonic settlement and amid the Atlantic independence era. The date is its own premise, not an inherited 1836 starting line.

• Thirteen additional polities. The roster grows from 67 to 80 with Russian America, Hawaii, Finland, Congress Poland, Lombardy-Venetia, Algiers, Hejaz, Funj, Darfur, the Kazakh Steppe, Bukhara, Khiva, and Kokand.
• Imperial relationships are explicit. Constituent states, vassals, colonial administrations, tributaries, and decentralized polities no longer have to be flattened into either fully sovereign countries or blank land.
• The worst ownership blobs are broken up. Central Asia, the Ottoman sphere, North Africa, Russian America, Hawaii, and the post-colonial Americas now paint from checked-in 1820 assignments. National borders regenerate from that data.
• The nation browser speaks to the date. New and corrected polities carry an 1820 summary, primary culture, religion, and relationship instead of generic fallback text.
• Historical content is reproducible. A checked-in compiler validates dated source packets, province identity, capitals, owners, controllers, relationship chains, and historical anchors before producing the runtime seed.
• Save mismatches fail safely. The world fingerprint schema advances so saves from the former 67-country topology cannot silently load against the new map.

[1.5.0] — 2026-08-01#

The great game#

The century had stopped moving. On the default seeds no nation ever unified, the same eight great powers held the table for ninety years, and the economy quietly starved from year thirty onward. All four were mechanisms, not tuning.

• Unification wars. A formable candidate now holds a standing, discounted claim on every core state it lacks — the Risorgimento pretext. The AI presses it against the weakest beatable owner, and you will find it ready in the Foreign Office. Sphering the minors still works; it is just no longer the only road to a crown.
• Sphere defection. Client states can tear up their patronage. The AI does it when unification is one core away; you get a "Leave the sphere" button on the Formables panel. The patron's grip resets to zero and the insult is not forgotten. Together these two took seed 4711 from zero formations in a century to Germany and Italy both proclaimed and Germany seated among the great powers by 1870.
• Prestige fades. 0.5% a month, a half-life of about eleven years. Sphere and influence prestige only ever accrued to sitting great powers, so the top eight had become self-sealing — the Ottomans matched Spain on industry and military and sat outside at prestige 6 against 784. Old glory is now a flow, not a moat, and armament moves the table again.
• The land answers scarcity. Needs-met decayed from year 30 because promotion was one-way: workers left for the mills whenever openings existed, farm capacity was frozen at bootstrap, and by 1880 the countryside held 15% of the population and the world went short. Promotion now follows realized wages — factory pay must beat farm pay — a clear farm premium pulls craftsmen back, and a saturated farm whose good prices above base grows a level. Needs-met at 1880: 0.60 before, 0.98 after, with world fill back at 99.9%.
• Hoards become demand. Pops sitting on more than half a year of basket now buy multiples of their everyday and luxury needs instead of banking forever. This is the money sink that lets taxation reach consumption at the margin.
• Near-great-power crowns. The great-power gate on Germany, Italy and the North German Confederation also admits a nation within striking distance of the eighth seat. A Prussia that fought its way to every NGF core should be crowning an emperor, not filing paperwork.

1.4.0 — 2026-07-27#

The chancery#

Diplomacy was a wall of buttons. Forty eight countries, each with six identical parchment buttons stapled to it, sitting under a war form that filled the whole screen. On a phone you could not see a single country without scrolling.

It is now three bands, in the order a foreign minister actually works in.

• Your standing, at the top. Points, infamy, rivals and influence as four gauges. Every number opens to show where it came from. Fabrications you have in flight are listed here, so committed points are never hidden.
• The roster. One line per country: flag, name, a wax seal for the relationship, and a bar showing how they feel about you. No buttons. Filter by Relevant, Powers, Near or All, each with a live count, and search reaches every country on the map whatever filter you are on. That is the answer to a long list on a small screen. You type three letters instead of scrolling.
• The dossier. Pick a country and everything about that country is in one place, including the war console. The target dropdown is gone because the roster is the target picker. Buttons you cannot use now say why in plain words, for example "Alliance: Would refuse: score 20 of 70" or "Rivalry: Cap reached (4 of 4). End one first."

Great Powers got the same treatment and now links straight through to a country's dossier, so you are not holding two screens in your head.

Everywhere else#

• Panels are taller on phones. The old ceiling left a band of dead map above every sheet.
• Section headings stick to the top as you scroll, so you always know which ledger you are in.
• Every kind of notification now has its colour. Eight of them never had one, and two colours were defined for kinds that could never appear.

Under the floor#

Groundwork for the economy rework, switched off by default so nothing changes yet. The market has been paying producers for everything they make rather than everything they sell, which quietly created money out of nothing: by 1825 it was paying out seven and a half times what buyers spent. That is why taxes have never had any bite. The replacement is written and measured, and turning it on comes with the rebalance it needs.

1.3.0 — 2026-07-27#

The floor#

The game does the same things it did yesterday, using markedly less of your machine to do them. Nothing here changes how it plays — that was the rule, and it was checked rather than assumed.

• The map and panels load in pieces. Opening a panel used to pull every panel; the map used to arrive as one 1.3 MB block. Both are split now, and the offline cache went from 10.2 MB to 2.0 MB — which also means the world can get bigger later without quietly breaking offline play.
• Screens stop redrawing when nothing changed. Every panel, the HUD and the map used to redraw eight times a second whether or not anything they showed had moved. Most now watch only the parts they display.
• Multiplayer builds one world view per broadcast instead of one per player. With eight players that was twenty-seven full world rebuilds a second.
• The daily war pass, the snapshot builder and the culture ledger all got cheaper — the culture figures are recalculated when population actually changes rather than eight times a second.

Checked, not assumed#

Every change here was verified to produce a byte-identical world: ten sim-years across three seeds, hashed and compared against the previous release. Two things that verification caught, which would otherwise have shipped:

• A snapshot refactor appeared to change the world. It had not — the probe was comparing key order rather than content. The probe was fixed.
• The culture cache was refreshed monthly while the underlying figures change weekly, so the Cultures panel could show numbers up to a month old. Sampling mid-month rather than on year boundaries exposed it.

1.2.0 — 2026-07-27#

The instruments#

Nothing here changes how the game plays. It changes what we can see — the gates every later change will be measured against.

• Save fingerprint. Saves now record the world they were made against (province count, content schema version, and a hash of the seed's ownership topology). A save from a different world is refused with a message that says so, instead of loading and painting the map with someone else's provinces. Saves made before this still load — a missing fingerprint is accepted.
• Content lint + coverage. Every event has a choice, every decision chain resolves, and every formable's core states exist — checked automatically, so a future map rework cannot silently drop content that points at ids which no longer exist. It also counts what each nation has to do, and the answer is currently: 2 of 48 nations have a decision of their own, 3 have an event of their own, 20 can form something, and 28 have none of the three.
• Multiplayer conformance + bandwidth gates. The multiplayer server runs the only world, so any change to the simulation is a change to multiplayer. A test now proves the server and single-player produce identical worlds from the same seed and inputs, and pins the wire cost so it cannot grow unnoticed.
• Performance budgets, recorded. Snapshot building, the culture ledger, and the number of screens that redraw on every tick all have measured numbers now, so "we made it faster" becomes a claim with evidence.
• The steerability probe explains itself. When a lever looks like it does nothing, the report now says whether players were limited by money or by what the market had in stock — which turned "the tax slider is inert" into a specific, fixable finding about the economy.

What the new instruments immediately found#

The gates earned their keep on their first run. None of these are regressions — they are things that were always true and that nothing could see:

• Fiscal policy is decorative (#33). Cutting tax moves player pop cash by 2.5 million and how well fed those pops are by 0.0002. Nobody is short of money — pops sit on roughly 833 rounds of everything they need to buy — so taxing them cannot reach them, and the treasury has been pinned at its ceiling since year three of the century.
• No unification happens (#34). Across a full century on both balance seeds, not one of the six formable nations ever forms.
• The great powers never change (#35). After the opening decade the top table moves six times in ninety years on one seed and not once on the other.
• Prices sit at roughly twice their base for the whole century (#36), within 4% of tripping the inflation gate.

1.1.0 — 2026-07-27#

The audit release. A thirteen-system design pass asked three questions of every mechanic — is it balanced, is it visible, is it deep — and this is the answer to the first two. The game stops hiding its own arithmetic, and stops quietly lying about the parts that never worked.

The game stops lying#

Five reforms were decorative. Guarantees did not join the wars they guaranteed. Factory subsidies were debited from the treasury and never credited to the factory. Negative tariffs were free money. All of it is wired now.

• White peace is no longer a free exit — an empty peace offer used to end any war outright, whatever the war score said.
• Sphered cores are not conquered cores — unification no longer counts a sphere member's land as controlled, and no longer annexes it for nothing.
• Tariffs are billed both ways — a negative tariff is an import subsidy and the treasury pays for it.
• Factory employment is exclusive per state — N factories used to draw on the same craftsmen N times over.
• Scandinavia and Iberia required no action to form. They do now.
• Colonial claims from events persist; plant-spam restacking is blocked; research point grants respect the cap.

The economy you can finally see#

• The clerk and capitalist ladder is alive — the promotion paths existed in the data and moved nobody. Wages and profits now reach the classes that were written to receive them.
• A national stockpile with standing buy and sell orders — the Victoria trade lever that was missing entirely. Order a good into reserve or flood it onto the market at a daily rate; large orders visibly move the price.
• Honest build margins — production shows real cost and real profit, RGO weekly profit lands in the ledger, and factory P&L traces to its inputs.
• Budget breakdown as a chart — income and expense as hand-drawn plates, army and navy split in the upkeep trace, a live projected ledger while you drag the slider, bankruptcy recovery in the HUD.

A century that moves#

• Pop ideology drifts with consciousness, militancy, needs and literacy instead of standing frozen at 1820 for a hundred years.
• Reform fatigue — a nation that has just reformed resists the next one; support and cost both feel it, and it decays month by month.
• Irredentist claims on lost cultural cores — losing your people's land gives you the casus belli to want it back.
• Combat width soft-caps the stack, which kills the doomstack.
• Three new unification arcs — Gran Colombia, the Scandinavian Union, the Iberian Union. Nine nations now have something to become.

The hidden state, surfaced#

Roughly forty things the simulation always computed and never showed: war score by component, diplomatic points and fabrication cost, colonial points and claims, crisis showdown forecasts, upper-house composition, party standings and election maths, alliance acceptance previews, tech modifier stacks, invention odds, per-type militancy and needs in the province panel, migration and promotion ledgers, occupation percentage, supply badges, in-panel battle logs. Population and culture get real charts; every number traces to its inputs on tap or hover, on a phone as well as a desktop.

Faster#

• Whole-world scans removed from the daily war loop (#7); per-snapshot main-thread work cut at 8 Hz (#8); a single-pass snapshot builder with bit-identical output (#9); the in-game HUD split off first paint and a dead copy of the world seed dropped from the bundle (#10).

Also#

• Error tracking is real. GlitchTip was wired in 1.0 and never delivered a single event: the DSN pointed at a host that now redirects to another origin, and the release tag was the literal string dev because the deploy command never set it. Both fixed — the release is now <version>+<sha>, derived at build time so it cannot be forgotten again, and the build stamp is visible on the main menu so a bug report can name its build.
• War alerts name belligerents; unit markers are scoped to nations you can actually see; the province panel chrome shows the owner's flag; tooltips are touch-friendly and stay on screen.

1.0.0 — 2026-07-22#

The Unification Arc. A campaign now has a beginning, a story, and an end.

The arcs#

• Prussia — Found the Zollverein, raise the German Question, force the Brothers' War (free CBs vs Austria). North German Confederation as the historical stepping stone; Germany from 1848 (the era gate shows in the Formables panel). GERMANY cores healed to the actual German Confederation.
• Piedmont — Champion the Risorgimento, court the French entente, sail the Expedition of the Thousand, pose the Roman Question (France will not smile on it). ITALY cores now include Austrian Lombardy-Venetia: fight Vienna for it, or unite every last minor.
• AI nations play the arcs too — the AI now takes national decisions (conservatively, deterministically); AI Austria competes for grossdeutsch; balance-of-power pressure makes the great powers watch any near-unifier.

The visible economy (E1)#

• Build buttons carry production chains and LIVE per-unit margins at market prices, sorted most-profitable-first. 11 civilian industries from 1820 (pre-industrial crafts un-gated).

War readability (U4)#

• Battle reports name the WHY (dice, organization, leadership, technology, terrain, fortress) from the player's perspective.
• Nation flag chips on every army and fleet counter.
• War alerts and the war list name belligerents, not ids.

The finish line (U5)#

• A yearly campaign chronicle; at 1920 (or on elimination) it becomes a recap of atlas plates — territory, population, prestige curves, wars fought, best rank — with "keep playing" always available.

Pacing (U3)#

• Century-probe harness (npm run probe:pacing) encoding the pacing contract: no dead decades, anchored prices, era-gated unifications.
• Great-power score rebalanced (diminishing industrial returns): the GP table moved twice in 80 years before; rank churn is back.
• Prussia healed: real 1820 population weight and coal (Silesia, the Ruhr) instead of a uniform-weight timber lot.

Also#

• Living title screen: the camera drifts across the world behind the title card. Era-appropriate nation flags everywhere (U0). Tap a country name to open its diplomacy view. Mobile panel scrolling fixed (iOS dvh); tutorial coach never blocks the nav.

0.9.0 — 2026-07-21#

The visual overhaul: the game now looks like the atlas it always wanted to be.

The atlas plate (V1–V7)#

• Title screen — hero backdrop, searchable nation browser, resume row, seed behind Advanced.
• Panel & HUD chrome — engraved rules, title shields, event-kind borders, wax-seal close buttons, alternating rows; typography/focus/micro-interaction polish; mobile refinements.
• Map engraving — deeper nation fills, land aquatint, player border halo, settlement dots, real rivers (478) & lakes (324) from Natural Earth 10m, capital star glyphs.
• Real borders, done right (V7) — provinces rebuilt as a true Natural Earth 10m tessellation (every NE admin-1 unit assigned to exactly one province): 0 overlapping provinces worldwide (V6 had 150+ in Europe alone), 0 holes, real coastlines, national-border ink only on actual national borders. Topojson-style shared-arc simplification keeps both sides of every border vertex-identical.

Nation flags (1.0-U0)#

• Era-appropriate flags for all 48 nations + the German/Italian formables — hand-drawn plates in a muted atlas palette (Habsburg black-gold, 25-star US, pre-1910 blue-white Portugal, Bourbon Two Sicilies, Tokugawa mon) — shown in the HUD, diplomacy, great powers, event feed, panel chrome, and nation browser. Unknown tags (procedural nations) fall back to procedural shields.

Procedural map modes (PR #5)#

• Start on the historical map, or a seeded procedural remap — contiguous reshuffled realms using real nation identities or invented countries — wired through the menu, permalinks, and saves.

Fixes#

• Political fills / hover / occupation overlays render again (V6 had dropped the GeoJSON feature ids that feature-state needs; promoteId guards it now).
• Country labels anchor at each nation's CAPITAL (UK labeled over Bengal before — the seed's capitals were population-picked; 20 healed to real 1820 capitals) and no longer flash: label layout re-ran every sim tick, now only when ownership actually changes.
• Diplomacy/great-power rows keep flag and name together; seed reroll is a text button (the die glyph had no font coverage).

0.8.0 — 2026-07-21#

Nationalism, performance, and a mobile fix pass.

The Age of Nationalism (culture / national identity)#

• Cultures 8→32, religions 5→7. Multi-cultural empires now carry real minorities (Austria ~31% German / 22% Hungarian / 18% Italian / 16% South-Slav / 13% Czech; Russia's Poles/Finns/Balts/Ukrainians; British India; Dutch Indies; Ottoman Balkans/Arab vilayets).
• Accepted vs non-accepted cultures (non-accepted → more militancy, no crown recruitment), assimilation (isolated minorities melt toward the primary culture, conserving people; rate by isolation/literacy/policy), and national movements — a boiling non-accepted culture launches a separatist independence rebellion (through the existing rebellion caps). Player levers: cultural policy (exclusionary/ assimilationist/pluralist) + grant/revoke acceptance. New Cultures panel; province dossier shows cultural makeup.

Performance#

• O(n) hot loops fixed (politics.monthly pop-bucketing + reform map; war.daily indexes). GeoJSON quantized (~1.24MB → ~546KB raw). Map fill skips unchanged provinces. Behavior-preserving + deterministic.

Mobile UI#

• Fixed tap latency — buttons now register immediately (touch-action + instant :active feedback, no double-tap).
• Fixed notification spam — routine events (elections) collapse quietly into the outliner; only war / crisis / bankruptcy / your own election pop prominently.
• Mobile polish for the new Technology & Crisis panels as sheets.

Notes#

• 141 tests (test:all) green; balance envelope held; SP + MP intact; old saves self-heal. (One culture test is intermittently load-flaky — a test-infra item.)

0.7.0 — 2026-07-21#

Deeper gameplay: great-power crisis diplomacy + a much bigger tech tree.

The Concert of Europe (new headline system)#

• Great-power crisis diplomacy — the world accumulates tension; tension ignites flashpoint crises (sphere contests, containment, humiliations); great powers take sides; each crisis ends either at a peaceful congress (demand enforced, big prestige swings) or, when blocs are balanced and both press, a great war (bloc-vs-bloc, resolved by the existing war machinery).
• New Crisis panel inside "Great Powers & the Concert": tension meter, live crisis card with Back Attacker / Back Defender / Press Demand / Back Down, and a congress ledger. Emergent texture: ~18-22 crises per 60-year campaign, most resolved at congress, the occasional catastrophic great war.

Technology depth#

• The tech tree roughly doubled (~80 entries): railroads, chemistry, medicine, electricity, naval/army modernization and more, with more wired effects (movement/supply, pop health, tax/trade/profit, war stats) and new inventions + production chains.

Notes#

• Additive types (old saves self-heal); pure/DOM-free so both run in single-player and the multiplayer server. 120 tests (test:all) green; balance envelope held.

0.6.0 — 2026-07-21#

"The Inventive Century" — a real technology & invention system for the 1820-1920 industrial arc.

Technology & inventions (new)#

• A player-directed tech tree: 31 techs in 5 year-gated prereq columns + 12 inventions. Research points, selection, ETA, and an idle auto-pick so no one stalls. (Previously the player literally could not research — it was AI-only.)
• Tech effects now actually apply: factory throughput (economy), tax efficiency (budget, deducted from pops — no minted money), and army/navy quality (war). Formerly dead display strings.
• New tech-gated production chains fill previously-inert goods (fish, wine, furniture, machine parts, artillery) — no more permanent unmet needs.
• New Technology panel (5 period-styled columns, progress + invention ledger); Production offers only unlocked recipes.

Critical economy fix#

• Every input-consuming factory had produced zero output since the M2 economy (buyFromMarket clamped Infinity→0). Fixed — the industrial economy is now alive, with a factory recalibration routing value to worker pops while taxes stay the state's lever. Pops remain believably fed (balance envelope green).

Notes#

• Strictly additive types (old saves load unchanged); pure/DOM-free so it runs in both single-player and the multiplayer server. 110 tests (test:all) green.

0.5.0 — 2026-07-21#

The visual overhaul — the map now reads as a premium antique atlas.

Map rendering (0.5.1 + 0.5.2)#

• Engraved sea with wave-lines, graticule, coastal aquatint + waterline rings, and spaced-italic ocean lettering; opaque land underlay + SE plate-shadow depth; paper grain + mottle + vignette + page-light varnish; de-washed nation pigments; letterpress labels placed on-territory (pole of inaccessibility).
• Terrain mapmode — an 11-biome hand-tinted physical plate (deserts, plains, forest/jungle, slate mountains, arctic) with engraved terrain textures + legend.
• Heraldic map counters — engraved heater-shield army tokens vs round naval cartouches, owner pigment + strength ribbons, wax-seal battle/siege/blockade badges (replacing the old text pills).
• Engraved relief/hillshade for mountain provinces (NW light), mapmode cross-fade transitions, archipelago waterline cleanup.
• Fixed a race that could leave the map uncolored if paused before load.

0.4.0 — 2026-07-21#

The big one: live multiplayer, a full UI + visual overhaul, a topology-fixed map, and a timeline rescope. Play: https://lakesidegames.net/games/grand-century/

Multiplayer (new)#

• Session-based live multiplayer for 2-8 players — server-authoritative sessions on a Node WebSocket server; single-player is unchanged (local worker).
• Lobby: create/join sessions by clicking (invite links too), nation selection (competitive one-each) or teams (co-op), leader-controlled start/speed, presence.
• Snapshot diffing + compression + cadence cap — bandwidth for 8 clients dropped from ~94 MB/s to well under 1 MB/s; the sim no longer burns CPU while paused.
• Reconnect with nation-hold grace + resync, in-session chat, presence HUD.

UI overhaul#

• A cohesive premium design system (parchment/ink/wax palette, self-hosted EB Garamond + Source Serif, spacing/elevation tokens) applied across the HUD, every panel, menus, and event popups — desktop and mobile.
• Interactive tutorial coach, tooltips-that-trace, actionable alerts, Economy panel.
• Mobile controls fixed (all buttons were tap-dead — pointer-events).

Visual / map overhaul (0.5.1)#

• A premium antique-atlas map: engraved sea with wave-lines + graticule + ocean lettering, plate-shadow land depth, terrain-textured fills, paper grain + vignette, richer pigments, letterpress labels. Map labels now render on-territory (pole-of-inaccessibility placement).

Map data — borders fixed + rescope#

• Topology-preserving simplification (TopoJSON) — no more sliver gaps, overlaps, double lines, or missing national borders; German/Italian states and France de-boxed into organic regions. Fixed a gameplay bug (614/620 provinces were wrongly flagged coastal → now ~46%).
• Rescoped to 1820-1920 (from 1836) with a plausible 1820 political map.

Performance#

• Fast test gate (unit ~9s; test:balance/test:all for long sims). Perf audit drove the snapshot cadence + MP diffing work.

Notes#

• 93 tests (test:all) green. New MP server runs as a systemd service behind a Caddy WebSocket route.

0.3.0 — 2026-07-20#

Theme: deepen the war pillar, add narrative agency, pay down platform debt — plus a major map-data rework from playtest feedback.

War depth & UI (E5)#

• Peace-conference settlement (stack multiple war goals within the war-score budget; AI offers/accepts sensibly).
• Army/fleet stack counters, movement, occupation shading on the map.
• Unit composition on recruitment (infantry/cavalry/artillery/guard) gated by conscription/professionalism reforms, with distinct combat roles.
• Rebellions are real: rebel armies with demands; a victorious rebellion forces its demand (reform enacted or provinces flip).

Events & decisions (E4)#

• Data-driven events engine with a curated set: 1848 Springtime of Nations, economic panics, colonial scramble, succession, discoveries. Player decisions gated by prerequisites, with real trade-offs. Fired as event popups.

Balance & world feel (B1)#

• Deterministic season-report harness; tuned so pops are believably fed and pop growth is sane. Rebellion governance (thresholds, caps, cooldowns) — concurrent rebellions dropped from 1000+ to a handful.

Map data rework (playtest feedback)#

• Consolidated from 1450 to ~473 provinces using real Natural-Earth admin-1 units with real names (Gansu, Piedmont, California…). China 80 → 33 real provinces. Western Europe no longer boxy. Zero fake numbered names. Vivid, distinct nation colors.

Performance & platform (E6, #4)#

• Split the heavy multi-decade sims: npm run test (unit) now runs in ~9s; test:all / test:balance for the long-run envelope.
• MapLibre kept off the critical path; PWA (installable, offline-capable) with a service worker; shareable permalinks (#/new?seed&nation).

Fixes#

• Mobile: all HUD controls were tap-dead (pointer-events) — fixed.

Known limitations#

• Map labels are not rendering — deferred to the 0.4.0 UI overhaul.
• France came out heavily consolidated (few provinces); can be subdivided later.

0.2.0 — 2026-07-20#

Theme: make a full campaign fun and legible, and give the player a grand goal.

Balance & feel (E1)#

• Deterministic season-report harness (npm run season-report) runs 60/100-year AI games across seeds and emits economy/population/geopolitics metrics.
• Tuned the economy/war/pop constants against it. A 60-year, 3-seed campaign now holds healthy bands: inflation ~1.3%/yr, factories profitable in aggregate, ~6% peak bankruptcy, ~2.2 wars/yr all resolving, no runaway hegemon (largest nation ~21% at year 20), believable pop growth.
• New long-run balance test guards the bands + determinism; perf test made robust to machine load.

Legibility & onboarding (E2)#

• Interactive 7-step tutorial coach on first play (replayable), on desktop and mobile.
• Tooltips-that-trace on every important number (budget, prices, war score, great-power score, militancy, pop needs) — each breaks down into its inputs.
• Actionable alerts: outliner/election/war items now jump to the relevant panel and suggest the next action, and can be dismissed.
• Map polish: province name labels at high zoom, reduced world-zoom label crowding, and per-mapmode legends with real scales.

Formable nations (E3)#

• Cores/claims per nation and formation decisions: unify the German states → German Empire, the Italian states → Kingdom of Italy (plus cheap analogues).
• Forming transfers the core territory, merges pops, adopts a new tag/colors, awards prestige, and updates the great-power ranking. The AI pursues formation too — a rising Prussia can proclaim the German Empire.
• New Formables panel shows eligibility + a live requirements checklist; a Cores mapmode highlights your core states.

Notes#

• Test suite is heavier now (~225s) due to the long-run balance sims.
• Pops run ~40% needs-met in long sims (stable, low militancy) — a candidate for further tuning.

0.1.0 — 2026-07-20#

First playable release. A single-player, browser-based grand strategy game in the spirit of Victoria 2, on a historical 1836 Earth. Runs fully client-side; the simulation lives in a Web Worker so the map stays at 60fps.

Live: https://lakesidegames.net/games/grand-century/

Systems (the full Victoria-2 loop, broad-but-shallow)#

• Economy — pops with life/everyday/luxury needs, RGO + factory production, one global world market with supply/demand pricing and conservation, monthly budget with real bankruptcy.
• Population — growth, migration, promotion/demotion, militancy → unrest.
• Politics — reform tree (economic/political/social/military), ruling party + elections, tax/tariff policy; military reforms gate mobilization.
• Diplomacy — relations, alliances/rivalries, casus belli + war goals, infamy, great-power ranking, spheres of influence.
• War & expansion (the pillar) — recruit regiments from soldier pops, armies + generals, movement/supply/attrition, combat (organization then strength), sieges and occupation, navies + amphibious landings gated on naval supremacy, war score + exhaustion, peace with war-goal enforcement (annex/liberate/humiliate/sphere/ colony), colonization.
• AI — non-player nations run the economy + war loop: ally against real threats, target winnable wars with fitting goals, concentrate armies and siege, make peace by war score + exhaustion.

World & map#

• 1450 real Natural-Earth province polygons, 55 nations including the 1836 German and Italian states, painterly parchment paper-map aesthetic.
• Mapmodes: political, ruling ideology, unrest, population, economy, military, diplomatic.
• Zoom level-of-detail: country regions + national borders when zoomed out, province borders when zoomed in.
• Self-hosted atlas-serif labels, centered per country and staggered by zoom/prominence to avoid overlap.

Platform#

• Responsive/mobile layout (compact top bar, bottom nav, sheet panels, touch pan/zoom); desktop side-rail layout preserved.
• Save/load via IndexedDB + gzip with yearly autosave; deterministic (seeded RNG), so loads and replays reproduce exactly.
• Audio manager with mute toggle (synthesized tones).

Quality#

• 30 unit/integration tests including a 20-year AI-driven stability run (wars occur, no runaway hegemon, bounded prices/treasuries), a save/load round-trip, and a performance ceiling. End-to-end Playwright smoke.

Known limitations#

• Europe’s big central powers can still crowd at full world zoom; province name labels not yet printed (borders + hover tooltips only).
• Map JS chunk is MapLibre-heavy (code-split from the main bundle).
• Balance is tuned to stability bounds, not yet hand-playtested for a full campaign.