- Update display name in registry, scene title, README, and tooling references
- Load ENABLE word list at runtime; fall back to unsteered random on failure
- Introduce makeSteeredGrid/refillSteering helpers for class/vowel-aware letter placement
- Export LETTER_WEIGHTS from BookworkLogic for reuse by the steering module
- Add `steer` tuning block (wordBoostK, classThreshold/damp/boost, vowelBand) to bookwork.json
- Rename card back themes (Cthulhu, Cyberpunk, Nautical, etc.) and update
artwork assets to match
- Move player portraits just left of each seat's tiles so they read as
belonging to that player without covering any tile
- Reposition the "Hands" reference button clear of seat 2's melds/bonus
tiles, and hide seat 1's portrait video while the panel covers it
- Auto-close the open reference panel whenever claim or win/kong buttons
appear so nothing clickable is hidden
- Add ~25 new steps: nearby stars, pan/zoom lesson, star info, fleet movement, home star/system view, colony management (allocation, focus), end turn, and the full Empire menu tour (Research, Diplomacy, Leaders, Colonies)
- Introduce hotspotAction, starPick, advanceOn modes (starPick/fleetInFlight/external), and advanceAction fields in tutorial data schema with validation
- Add VegaStarMap.fitToWorldBounds and extend pickFitZoomIndex to support a narrowed viewport area so the reachable-stars frame avoids the docked panel
- Bypass openModal gating for System View, Colony View, and Empire screens by calling their open functions directly from VegaTutorial; wire onClose/onColonyOpen callbacks to advance()
- Add tutorialFreePan flag so pan/zoom work during the free-pan step while all other map interactions stay frozen
- Add multi-target resolution (nearbyStars returns an array), invisible-shape holes, and per-star click hotspots in the tutorial overlay
- Update verification script to allow buttonless modal steps that advance via external callbacks
- New data-driven script (data/mastervega-tutorial.json) with intro modal and three callout steps over the fleet marker and side-panel ship rows/count controls.
- Split into Phaser-free VegaTutorialData.js (schema validation, interpolation, target-id registry) and Phaser VegaTutorial.js (darken overlay, accent rings, callouts, hotspot, skip confirm, state machine).
- Tutorial auto-opens on every brand-new game and is re-triggerable from the menu when the starting fleet is still in orbit; malformed JSON disables the feature with a warning instead of crashing.
- Add D.tutorial depth (just below detail) so an invited ship-detail window layers above the overlay, and expose VegaSidePanel.tutorialRegion() for screen-space panel regions.
- Register the tutorial asset in the manifest and add verifier section 12 covering schema, targets, voice clips, placeholders, and interpolation.
Replace the flat dark rectangle status bar with a baked riveted-metal "blue steel plate" background styled after Wolfenstein 3D's original HUD:
- Add `paintHudBar` to WolfensteinArt.js generating a 1920x140 texture with vertical steel gradient, bevel edges, recessed readout panels, divider strips with rivets, and a centered portrait socket sized around the new `HUD_PORTRAIT_SIZE` constant
- Add `paintProfile` fallback face and wire up `profile.png` (128x128, 3 health frames) in wolfenstein-artwork.json
- Rebuild `_buildHud` in WolfensteinGame.js: left panel shows episode/mission label + HEALTH value, center shows health-reactive portrait (frame swaps at 66/33 HP thresholds), right panel shows INVENTORY key icons + AMMO count with equipped weapon icon
- Derive weapon/key icon frames from `rules.items` data instead of hardcoded values so HUD stays in sync with wolfenstein-rules.json
- Add `_hudEpisodeMissionText()` helper for static per-level label text (campaign episode/mission or test level name)
- Apply subdued CRT overlay (`applyArcadeCRTOverlay`) tuned for a tactical-display feel, destroyed on teardown
- Switch HUD fonts to shared `m6x11` pixel font with white/ice-blue lettering; remove old KEY_COLORS constant and rectangular key indicators
- Add fists weapon viewmodel with idle/hit animation states (220ms hit pose)
- Implement mouse-wheel weapon cycling that skips unowned weapons
- Add 33% chance for killed guards to drop ammo-clip pickups at death location
- Introduce campaign carry-over system preserving weapons/ammo/health between missions
- Fix stale 'attack' state clearing when enemies lose sight of player
- Extend save/load serialization to persist nextPickupId counter
- Update level-e1m1 with pistol pickup at spawn point
- Add comprehensive test coverage for all new features in verifyWolfenstein.js
- Add shotgun, machine gun, gatling gun, and plasma rifle weapons with
distinct fire modes (semi-auto, auto, burst) and damage profiles
- Refactor ammo from per-weapon pools to shared pools keyed by ammo type
(9mm, shells, plasma), allowing multiple weapons to share the same pool
- Add pickup system with 11 item types: 5 weapons, 4 ammo variants, and
2 health packs, all with spritesheet frames for rendering
- Update level editor with dynamic pickup dropdown populated from rules.json
- Generalize weapon viewmodel rendering to support multiple weapon images
with per-weapon bob/sway animation
- Add cosmetic pickup animations: vertical bobbing and orbiting sparkle
effects drawn in screen space around each pickup billboard
- Bump save version to v3 to invalidate old saves with incompatible ammo
structure
- Update keyboard bindings for weapon switching (keys 1-6)
- Add comprehensive test coverage for new weapons, burst fire mechanics,
ammo pooling, and pickup behavior
- Rename weapon_gattling.png to weapon_gatling.png to fix spelling
Introduces a new "Object" entity type that blocks movement but not sight/bullets, distinct from walls and pickups:
- **Data**: New `objects` array in level JSON (`{x, y, frame}`), with `data/wolfenstein-objects.json` as the frame registry (Tall Bush, Gold Eagle)
- **Artwork**: New `objects.png` spritesheet registered in `wolfenstein-artwork.json`; placeholder texture via `paintObject()` in WolfensteinArt.js
- **Logic**: Objects stored in `state.objects`, with a derived `map.objectBlocked` Set checked only by movement collision (`isWallCell`) — raycaster/sight/bullets never read it, so objects are see-through and shoot-through by construction
- **View**: Rendered as billboard sprites in `_drawSprites` with ground-anchored positioning (bottom edge on floor) and asymptotic depth sorting that stays below the weapon viewmodel
- **Editor**: New "Object" tool category with dynamic dropdown; click open floor to place/remove (toggle), click wall is a no-op; teal square indicator on board
- **Validation**: `validateLevel` checks objects sit on open floor; save/load round-trips with graceful degradation for pre-existing saves
- **Tests**: Verifies object blocks player movement but bullets pass through
- Add new `wallArt` level data array with per-cell decal frames
- Integrate wall art rendering into WolfensteinView using same texture sampling as walls
- Add Wall Art tool to WolfensteinEditor with dynamic dropdown from JSON registry
- Create wolfenstein-wallart.json frame registry for easy decal management
- Add validation ensuring wall art sits on actual wall cells
- Support wall art in save/load system with backward compatibility
- Alert idle enemies when shot, even outside vision cone
This refactors weapon/enemy combat systems to support per-weapon cooldowns,
fire modes (auto/semi), and data-driven damage rolls via ammo types. Guards
now use the pistol's stats for ranged attacks and can be stunned by non-lethal
hits. The level editor now includes a directional facing picker for newly placed
enemies and displays enemy orientation as arrows instead of dots.
Key changes:
- Logic: Replaced single weaponCooldownMs with per-weapon cooldowns map; added
prevFireHeld for semi-auto trigger-edge detection; guards now reference real
weapon definitions via rangedWeapon; projectiles carry ammoType and resolve
damage at hit time using rollDamage(); non-lethal ranged hits apply stunMs to
enemies, freezing their AI until it expires.
- Rules: Added ammoTypes array with damageMin/damageMax; weapons now specify
fireMode and ammoType (projectiles) or damage (melee); enemy definitions
include stunMs and reference weapon IDs instead of hardcoded projectile stats.
- Editor: Added FACING_DIRS constants and drawFacingPicker() overlay that lets
users click N/S/E/W neighbor cells to set a just-placed enemy's facing;
enemies now render as directional arrows showing their spawn orientation;
pendingFacingEnemy state tracks the in-progress pick across tool switches,
undo/redo, and level resets.
- View: Added GUARD_FRAME.stunned (frame 9) which takes visual priority over
all other states when e.stunMs > 0, ensuring the flinch pose displays even if
the enemy's internal state still reads 'attack' or 'chase'.
- Data: Expanded level-e1m1 to 19x21 grid with repositioned doors/enemies/exit;
bumped SAVE_VERSION to 2 since player.cooldowns and prevFireHeld are new
required fields absent from v1 saves.
- Tests: Added coverage for semi-auto trigger-edge behavior, stun application
and duration, melee non-stun behavior, damage roll range bounds, and guard
projectile ammo type/cooldown spacing; updated existing tests to use new
cooldowns/ammoType fields and toggle fireHeld for repeated shots.
Introduce a first-person weapon viewmodel for the pistol that peeks up
from behind the HUD, completing the classic FPS look. The viewmodel
features a subtle, smoothed bob and sway animation that responds to
player movement (forward/strafe), ramping in and out to avoid
snapping.
- Add `paintWeaponPistol` to `WolfensteinArt` as a small procedural
placeholder (360x260) for when real art is absent.
- Update `WolfensteinView` to handle both real and procedural weapon art
with appropriate origins and base positions.
- Implement `_drawWeapon` with bob/sway logic and smooth strength
transitions.
- Register the new sprite key in `wolfenstein-artwork.json`.
- Document the new `wolfenstein-weapon-pistol` asset in `sprites.md`,
including its size, depth, and behavior.
- Add directional billboarding for guards: front/side/back idle & walk
frames selected by relative angle to camera, with flipX for left side
- Add shooting pose (always shown when guard is in attack state)
- Add 2-frame walk cycle (300ms) for moving guards (chase/alert/patrol)
- Add death sequence: fall (350ms) → ground (2s) → fade (800ms),
view-only state in WolfensteinView, pooled sprite destroyed on completion
- Wire guard sheet texture from artwork JSON (enemies.png, 9×7 grid,
frames 0-8 used)
- Add E key as alternate interact key for opening doors
- Update door hint text to show [SPACE/E]
- Update sprites.md with full animation/facing/death documentation
The raycaster's door geometry now takes the live `doors[]` state instead of a boolean flag, so partially-open door cells let render rays pass through the already-slid portion to show what's genuinely beyond, rather than painting a flat "socket" fill. `intersectDoorMidplane` returns a slide-shifted `textureX` so the door texture visibly translates into the wall as it opens.
`WolfensteinView` samples the new `wolfenstein-doors` sheet the same way it does walls (source-texel column, vertical crop, multiply shading), falling back to the flat tan placeholder. Added the doors.png/.psd assets and a `sheets.doors` entry in the art manifest. Updated sprites.md to document the new sheet, the see-through behavior, and that this stays render-only (gameplay raycasts are unaffected).
Editor (WolfensteinEditor.js):
- Decouple the board from grid size: fixed 900px viewport with independent
pan (right-drag / WASD / arrows) and zoom (wheel / +/- / F to fit)
instead of squeezing the whole grid in; add a whole-level map overview
minimap with click-to-jump.
- Grids up to 1000 cells/side (was 48); painting past the edge auto-grows
the grid (preserving content, re-sealing borders) instead of being refused.
- Rebuild the toolbar as a categorized DOM radio + dropdown panel (Wall /
Door / Pickup / Enemy / Patrol / Zone / Erase) instead of a flat list.
- Add a Patrol tool: select a guard, then click tiles to add/remove
waypoints; routes are drawn (dim for all, highlighted for the selection).
- Throttle the reachability validate during paint-drag; flush it before
Test Play / Export; queue game assets in preload() like ZumaEditor.
Doors (WolfensteinLogic.js, WolfensteinRaycaster.js, WolfensteinView.js):
- Player now opens doors by pressing Space (HUD "[SPACE] Open" prompt when
one is in range); enemies still shove them open. Doors animate open/closed
via a slide value (400ms) instead of popping, and stay solid for
gameplay until fully open.
- Render doors as recessed mid-cell geometry (real depth, producing the "H"
doorway shape) via a door-aware render-only raycast — collision/LOS are
unchanged; the door visibly slides sideways into a dark socket as it opens.
Enemy AI (WolfensteinLogic.js):
- Idle enemies with an authored patrol route ping-pong along
[home, ...nodes] until they spot the player; validate patrol nodes.
Rendering (WolfensteinView.js, assets, data):
- Add textured wall rendering: per-column 1px source-texel sampling from
the new walls.png spritesheet, darkened with a 'multiply' grey to match
the existing side/fog shading, with per-wall-type flat-color fallback.
- Wire in assets/images/wolfenstein/walls.png and point the art manifest at it.
Misc:
- level-e1m1.json: reworked to a taller (20-row) map exercising the new
wall/door types; playerRadius 0.25 -> 0.35.
- sprites.md: document exact wall-sheet size, the sliding/recessed door
behavior, guard sheet usage, and standalone image sizes/proportions.
Movement & collision:
- stepSeparation (TALogic) replaces the hard "snap unit to nearest clear
tile centre" correction with a circle-vs-blocked-tile push proportional
to overlap, damped by separationStiffness — no more teleport-yank,
corner-bouncing, or stalling in exact-width corridors (caught as an
AI-vs-AI win-rate regression in testing).
- servicePathQueue now routes at a unit's exact required clearance instead
of padding +1, so traffic spreads across all passable tiles rather than
funneling into a few wide chokepoint corridors.
- TAArt's structure painter falls back to a generic spinning glyph for
topperFrame buildings.
Upgrades (new rule: upgradesFrom):
- A building may be placed directly on top of a friendly building it
names at the same tile/footprint, consuming it for a 50% refund of the
replaced building's OWN cost (not a price discount) — implemented via
TALogic.findUpgradeTarget and an army-aware canPlaceAt.
- New definitions: nuclearplant (upgrades energygen, frame 22) and
advancedmassgen (upgrades massgen, frame 24).
- compileRules (TARules) validates upgradesFrom references and topperFrame;
sprites.md documents the semantics.
- New map m06 "Annihilation" (medium, snowfields, Klaxon skill 5, seed
90123 — pre-vetted from the original sweep for early-economy balance).
Tests (tools/verifyTotalAnnihilation.js):
- Section 4d: pathfinding routes around obstacles and a unit threads a
corridor exactly its own width.
- Section 5b: obstacle push is a bounded per-tick nudge (not a
teleport), settles clear without oscillation, a unit ordered past an
obstacle arrives, and no stuck-give-up flagging.
- Upgrade suite: exact 50% refund credited in place, fresh-ground builds
still work, enemy-owned and mismatched-type targets rejected, reverse
nuclearplant/energygen pairing, and canPlaceAt without an army never
grants upgrade placements.
- New `upgradesFrom` rule: a building def (Nuclear Power Plant on the
Energy Generator, Advanced Metal Generator on the Mass Generator) may
be placed exactly on a friendly, finished building of the named type,
consuming it for a 50% refund of its build cost instead of needing
clear ground (TALogic findUpgradeTarget / consumeUpgrade / buildCommand)
- New `topperFrame`: a second sprite stacked on the finished structure
that spins continuously (the Advanced Metal Generator's centrifuge,
frame 26), hidden while under construction; TAArt sizes the sheet for
it and paints a generic spinner-glyph fallback
- New rules: Nuclear Power Plant (frames 22-23) and Advanced Metal
Generator (frames 24-26) in totalannihilation-rules.json
- TARules: validate topperFrame and upgradesFrom (real building with a
cost); TAWorldView renders/spins the topper; verifyTotalAnnihilation
checks topper frames fit and were painted
- Test section covering upgrade placement (ownership, type match, exact
footprint, refund amount, army-less fallback) and sprites.md updated
for frames 22-26
- Introduce Megatank: medium commander with Twin 150mm Cannon and Side Rocket Pods
- Implement multi-barrel weapon system (barrels + barrelSpacing) in TALogic and TARules
- Lower DEFENCE_MIN_SKILL from 4 to 3; make first Laser Tower proactive, not reactive
- Rebalance unit triangle: Tank now counters Rocket Tank; rockets weak vs vehicle armour
- Re-seed campaign missions m05 (67890) and m06 (80808) with updated maps and briefings
- Skirmish default theme uses rules default (tropics) instead of hardcoded grasslands
- Add intro video before main menu
- Update sprite docs and verification tests for new balance
- Introduce Megatank: medium commander with Twin 150mm Cannon and Side Rocket Pods
- Implement multi-barrel weapon system (barrels + barrelSpacing) in TALogic and TARules
- Lower DEFENCE_MIN_SKILL from 4 to 3; make first Laser Tower proactive, not reactive
- Rebalance unit triangle: Tank now counters Rocket Tank; rockets weak vs vehicle armour
- Re-seed campaign missions m05 (67890) and m06 (80808) with updated maps and briefings
- Skirmish default theme uses rules default (tropics) instead of hardcoded grasslands
- Add intro video before main menu
- Update sprite docs and verification tests for new balance
- Add morale bonus from buildings (e.g., Holo Simulator): flat +N% to total
colony production, uniformly boosting construction, defense, industry,
ecology, and research
- Add espionage restraint (offDrift): setting Intelligence to 'off' toward
an empire with neutral-or-better opinion drifts their attitude upward by
1/turn, giving a way to improve relations without gifting (gated on not
being at war and their opinion of you being >= 0)
- Display morale bonus in colony view when applicable
- Reduce planetShotDamageMult from 10 to 6
- Add verification tests for both new mechanics
Bumps the planetShotDamageMult config value in mastervega-rules.json
to significantly increase damage dealt during planetary bombardment
or planet defense scenarios.
- defenseHp is now a pure per-battle ammo gauge (one full-strength
shot per planetDefensePerShot points, remainder as fractional shot)
instead of in-battle damage tracking; no longer written back after
battle outcome
- Planet now has a real range limit (planetRange, 1200) — previously
the only entity with no range check at all
- Per-shot damage scales with the defending empire's best known weapon
tech (weaponAvgDmg) instead of a flat HP-proportion curve
- defenseHp is only wiped to 0 when the planet is actually DESTROYED,
fixing the bug where a defeated planet could be re-attacked multiple
times in one turn with full strength each time
- Added combat config: planetDefensePerShot (5), planetShotDamageMult (1)
- Added combatV2 config: planetRange (1200)
- Comprehensive tests for range gating, shot formula, ammo exhaustion,
multi-battle-same-turn fix, and tech-scaled damage
Add siegeFreezesDefense mechanic (default true) that stops a colony's
passive defenseHp trickle and blocks defensive building completion at the
front of the build queue while a hostile warship fleet is present.
This fixes the recurring "Under Attack" notification loop reported by Brian
where defenseHp would regenerate each turn, re-arming the pendingBattlesFor
gate and triggering fresh popup notices for an already-beaten colony.
Also adds an AI dispatch throttle in manageFleets: idle fleets no longer
open redundant attack orders against colonies the empire already holds orbit
over with defenseHp beaten to 0. Excluded fleets fall through to rally logic
and naturally join the existing siege or find a fresh front.
New helpers: isDefensiveBuildingId, isHostileWarshipFleet, hasHostileWarshipFleet
— shared across processColony, pendingBattlesFor, and VegaAI to ensure a
single definition of "under active siege".
Includes tests (sections 5d, 5e) covering regen freeze, build queue freeze,
siegeFreezesDefense=false toggle, and AI throttle.
- **Visual Feedback**: Introduce VegaBombardScreen.js — a two-stage popup for
bombard/invasion results featuring JIT-loaded videos (per planet type or
invading species) and animated stat bars showing population, factory, and
building losses.
- **Expanded Consequences**: Factory and building destruction now scale with
the population kill ratio. On colony wipeout, weapon Mark gates planetary
ruin/ruination (toxic/radiated/dead/asteroids) via a weighted probability table.
- **AI & Tactics**: Ursaal now queue troop transports alongside warships for
invasion fleets. Added `bombardHoldFireChance` trait (0–1) to let species
defer bombing when invasion troops are staged, preserving intact colonies
without creating fortress-world stalemates.
- **Battle Logic Fixes**: Correct `prepareBattleAt` and `pendingBattlesFor` to
properly handle stars with multiple colonies, ignore transport-only fleets,
and suppress pending battles when planetary defenses are depleted.
- **UI/Reporting**: Wire player-initiated and AI-initiated attacks into the
new popup queue. Add `bombard` and `invasionFailed` to notable turn report
events with detailed outcome descriptions.
- **Testing & Tooling**: Add comprehensive verification tests for planet
destruction distribution, building damage rates, hold-fire chances, and
battle edge cases. Expose `window.game` for devtools debugging.
- Replace flat planetDefenseBase (20) with proportional planetDefenseScale
(0.06), so colony defense damage scales purely with current defenseHp.
A near-depleted colony now barely scratches attackers, while a fully
tech'd one still tops out ~110 damage — fixing "attacked with the
strength of a much-better-defended planet" (Brian, 2026-08-14).
- Add AI war memory: track consecutive losses per enemy and escalate
attackMultiplier up to 3x so repeated losing attacks stop dribbling
small fleets. A win resets the streak. Prevents permanent phoney wars
where both sides build to equal strength and nothing moves.
- Show "Under Attack" notice (one-button, no formation picker) when AI
attacks the player, since formation choice is meaningless for defense.
- Add planetary beam fire SFX for combat view.
- Add event `seq` cursor for war memory to survive event list trimming.
- Full test coverage for planet damage formula and AI escalation behavior.
- Replace □/■ glyphs with channel-specific icons inside the lock toggle boxes
- Introduce `allocation.png` (5 frames) and register it in the artwork config
- Add hover tooltips explaining each channel's production role and MOO1-style
overflow behavior
- Expose slider `zone` for tooltip attachment and adjust icon/layout positioning
- Update sprite documentation and include minor audio asset update
- Add leader detail pop-over (VegaLeaderDetail.js) with portrait, bio, and
current posting bonuses displayed in fleet and colony views
- Add leader portrait click-to-zoom in the Leaders screen
- Add leader skill summary formatting (leaderSkillSummary) for readable
bonus descriptions across UI components
- Integrate leader skills into espionage power/defense and invasion attack
calculations in VegaLogic.js
- Cap bombard() at once per (attacker, colony) per turn via lastBombardTurn
tracking; fixes UI double-click and AI double-bombard bugs
- Contact-gate GNN breakthrough stories and ranking charts on first contact,
matching existing rankingRows behavior
- Round population displays throughout the UI (toFixed(1) → Math.round)
for cleaner presentation
- Bump colony ship and transport base costs in rules data
- Add leaders.png artwork reference to mastervega-artwork.json
- Add tests for bombard cap and GNN contact-gating in verifyMasterOfVega.js
- Add "peace request" diplomacy: player can ask an AI to end a war with a
third empire, and AIs can ask the player the same. Refusing costs attitude.
Implemented in VegaDiplomacy.js (wouldAcceptPeaceRequest/requestPeace/offerPeaceRequest),
VegaAudience.js (Demand Peace button + war picker), and full chat lines for
all species in VegaChat.js.
- Extract openFormationPicker to VegaScreens.js and wire it into
MasterOfVegaGame.js so the player chooses a formation before every battle.
AI opponents still pick silently.
- Fix Combat V2 planet positioning (fixed at worldWidth * 0.75 instead of
formation-relative), pass typeId for real planet art, and apply the
Planetary Shield building's shieldBonus (was hardcoded to 0).
- Add icon images to research screen tech field headers.
- Add comprehensive tests for peace requests and prepareBattleAt fixes.
Introduce a full-screen, click-advanced ceremony (VegaCouncilSession.js)
that replays runCouncil()'s result delegate-by-delegate instead of hiding
the tally behind the turn-report. Three fixed stages — Roster, Voting,
and Verdict — with a persistent 2:3 portrait anchor video, green-screen
ticker, live scoreboard with animated bars, and sting-then-loop audio.
Key changes:
- VegaLogic.js: expose per-voter `voters` breakdown and `pendingSession`
flag so the UI can replay the session before the victory check.
- MasterOfVegaGame.js: consume `pendingSession` in `runToHumanTurn()`
before `state.over`, ensuring a decisive council still gets its ceremony.
- VegaTurnReport.js: remove `council` from NOTABLE_TYPES (ceremony handles
it); keep `councilRefused` as a distinct war-declaration row.
- assetManifest.js / Sounds.js: register video + sting/loop audio cues.
- VegaScreens.js: add `D.council` depth value.
- verifyMasterOfVega.js: assert voter breakdown consistency and pending
session lifecycle.
Introduce GNN, a galaxy-wide news broadcast feature that reports major
events to the player via a full-screen takeover UI with anchor desk
video, story pages, MOO1-style animated rankings, and a diplomatic
relations reference.
Key changes:
- VegaGnn.js: headless event classification, anchor copywriting, and
ranking logic (Phaser-free, testable in Node)
- VegaGnnScreen.js: full-screen UI with looping anchor video,
green-screen ticker terminal, story renderers (diplomacy/tech/
territory/espionage), animated bar-chart rankings, and relations page
- HUD: new GNN button with unread bullet cue; auto-opens after turns
when stories are pending, on-demand via button
- VegaLogic.js: new `lastColony` event (reduced to single world),
`attacker` field on elimination events, `gnn` state with history
buffer, export empireFleetPower for multi-empire rankings
- VegaColoniesScreen.js: export and generalize createAdvisorTerminal
for reuse by GNN's ticker
- mastervega-rules.json: flag ~16 headline-worthy techs with
gnnHeadline (research-sourced only, not espionage-stolen)
- Assets: eager-load GNN anchor video, sting/loop audio, and populate
all 10 species colony advisor videos in mastervega-artwork.json
- verifyMasterOfVega.js: comprehensive 7b test section covering
classification, consumption, describeGnnStory, anchorLine, attacker
threading, ranking metrics, relationsRows, serialization, and AI soak
Introduce VegaColoniesScreen.js, a spreadsheet-style screen that displays
and manages all of the player's colonies at once, grouped by star and
sorted by population. Features include:
- JIT-loaded advisor video panel (one 1:1 muted loop per species, falling
back to the species portrait when no clip is recorded)
- Retro terminal with typing animation, scanlines, and glitch effects
that displays advisorColonyReport commentary for the selected colony
- Inline Colony Focus and Allocation Focus pills with dropdown pickers
showing the advisor's live recommendation outlined in yellow
- Population transfer (arm-and-target idiom matching fleet orders)
- Manage button delegating to the existing single-colony flyout
Refactor VegaColonyView.js to use shared applyColonyFocus and
applyAllocationFocus setters from VegaLogic.js so both the flyout and
inline pills apply identical side effects. Add empireColoniesByStar
grouping helper for the spreadsheet view. Wire a "Colonies" entry into
the Empire dropdown menu in MasterOfVegaGame.js. Add advisorVideos
entries to mastervega-artwork.json with JIT-loading helpers in VegaArt.js.
Update verifyMasterOfVega.js to validate the new assets, lazy-loading
behavior, and all new logic functions.
- Hulls: frigate turnRateBase 150→340, brakeSeconds 11→1.0 (X-wing agility);
destroyer turnRateBase 100→230, brakeSeconds 2.8→1.35 (midpoint leaning frigate).
- Fix battleship (and all slow hulls) no longer holding position: added
anticipatory retrograde pre-turning in computeShipMove, comparing
worstTurnTime vs timeToRange so the nose begins flipping before braking
range, with urgency-blended facing to avoid "running away" appearance.
(Was accidentally crutched by strong avoidAccel collision avoidance.)
- Fix parked ships coasting indefinitely: velocity now starts from 0 when
hasTask is false instead of carrying residual velocity forever.
- V2 damageMultiplier (2.8): scales weapon raw output before shield
mitigation, V2-only, doesn't touch live engine balance. Shortens battle
duration meaningfully across test scenarios.
- V2 maxDurationSeconds (240) and disengageFraction (0.9): decoupled from
shared rules.combat.maxRounds. Counterintuitively, the old disengage
timer was an implicit duration cap; extending it let genuine fights play
out while damageMultiplier handles the shortening.
- Species-colored beams: VegaCombatViewV2.js now renders beams in a
vibrant, saturation-boosted version of the firing ship's species color,
memoized. Missiles keep fixed orange.
- Mirror-match bias tolerance widened 0.25→0.30 (expected variance cost
of lower TTK from higher damage multiplier).
- Frigate stationary-target check inverted: now asserts it CAN hold
position (alongside battleship/cruiser/destroyer). Added agility ordering
data-integrity checks on hulls block.
Implement continuous velocity vectors (vx/vy) replacing the old
kinematic position clamp. Each hull gets brakeSeconds-derived
linearAccel so stopping distance naturally falls under or over beamRange,
producing battleship hold-position vs frigate strafe-run behaviour
without hull-specific tactic branches.
Key mechanics:
- Thrust magnitude scaled by facing alignment (cos of angle to task
direction), honouring "turn and engage thrusters" without locking
thrust to facing (which caused battleships to never stop).
- GOLDEN_ANGLE-spaced approach points fan out ships sharing a target.
- avoidAccel (60) is a gentle, hull-independent collision avoidance
budget — ships can overlap when objectives require it.
- Target avoidance uses reduced radius (0.35×) so engagement isn't
fought by personal-space calculations.
Data: add brakeSeconds to all hulls, avoidAccel to combatV2 constants.
Verifier: rewrite turn-rate check to time-to-range metric, add hold-vs-
strafe gradient and max-delta-speed checks, relax cloak/singularity to
"doesn't hurt" invariant.
Introduce `VegaCombatV2.js`, a parallel headless combat engine that simulates
individual ships (not stack aggregates) in continuous time rather than discrete
rounds. Ships have per-hull `sizeScale`, `sizeSpeedMult`, and `turnRateBase`
fields; every entity tracks its own cooldown, position, facing, and angular
momentum. Damage and movement are banked per-tick to preserve fairness.
Accompanying files:
- `VegaCombatViewV2.js` — zoomable/pannable Phaser view with parallax starfield
- `VegaCombatCamera.js` — cursor-anchored zoom + drag-to-pan (listener cleanup
on destroy to avoid stacking across battles)
- `VegaFormations.js` — formation strategy plumbing (chosen per side, not yet
read by movement/targeting)
- Hull-specific destroy audio cues and vega buildings art
Wired behind `?movsim`'s Live/V2 toggle in `VegaCombatSim.js` — the live engine
and all real player battles are untouched.
Also fixes several latent bugs discovered during the rewrite:
- Trap 26: simultaneous wipe-out now checks draw before single-side wins
- Trap 27: movement is banked (two-phase) like damage
- Trap 28: weighted-random targeting tames chaotic nearest-neighbour sensitivity
- Trap 29: fleet gap clamped to 85% world width to prevent starvation
- Trap 30: `RANGE_EPS` tolerance prevents steering-induced standoffs
See `docs/mastervega-build-plan.md` for full design notes.
Move vega audio files into assets/fx/vega/ subdirectory and register them via
assetManifest.js. Add weapon and missile sound cues banded by ship Mark, with
staggered playback in VegaCombatView to prevent audio clipping during barrages.
Wire up two previously inert tech effects in combat:
- Stealth Field (cloaked): applies a flat accuracy malus via cloakEvasion rule.
- Black Hole Generator (singularity): adds passive shield pierce via
singularityShieldPierce rule on every weapon fired by the design.
Update tech descriptions, ship detail display, and designFor() to reflect
these effects. Fix a bug where singularity was never copied into the design
object.
Add vega-endturn cue to End Turn / Next Round buttons (no-op when busy),
remove hardcoded audio loads from PreloadScene, and add verification tests
that isolate each flag with A/B battle comparisons.
Introduces VegaCombatSim, a standalone dev tool for tuning and
testing VegaCombat. It allows configuring attacker/defender fleets
(or fleet vs planet), watching animated battles on the real tactical
screen, and running batch trials for win-rate statistics. Both paths
use the exact same VegaCombat stepper as the live game.
Also fixes a transport unit amount default (1 → 0) in
VegaColonyView and corrects a music file reference in the JSON
data file.
- Introduce Colony Focus: a per-colony autopilot that auto-queues one
building/ship per turn when the queue is empty. Supports seven modes
(manual, improvement, research, fleet, growth, trade, defense) with
cost-aware affordability checks matching AI behaviour.
- Add Allocation Focus: one-time slider presets (default, research,
growth, production, military) that bulk-overwrite channel allocations.
- Refactor colony view flyout into a shared drawer supporting queue,
colony focus, and allocation focus modes with in-place swapping.
- Replace "Built here" text with hoverable building icon row and
tooltips on the colony screen.
- Add per-species audience music tracks and diplomacy music ducking when
seeking/planning an audience.
- Add sound effects for star selection, zoom in/out, and colony view
opening.
- Back-fill `focus: 'manual'` on old saves for forwards compatibility.
- Add verification tests exercising each focus mode in isolation.
- Implement 10 manual save slots with metadata (turn, year, empire, species, date)
- Add ☰ game menu button with Save, Load, Return to Main Menu, and Quit to Arcade
- Replace generic sounds with 7 custom Vega sound effects (select, build, close,
newturn, unit, view, warp)
- Introduce uiClick() helper to standardize click sound across all modal screens
- Update research video paths for umbrix and rrashaa species
- Refactor button handlers in all Vega screens to use uiClick() pattern
- Save/load uses scene restart via pendingSavedState for clean HUD/map teardown
- Add scaling leader hire costs based on number of leaders already owned (costScalePerOwned: 0.5), making it progressively more expensive to hire multiple leaders and preventing cheap roster stockpiling
- Add leaderHireCost() function in VegaLogic.js to calculate scaled hire costs
- Add unassignLeader() function to allow removing leader assignments
- Update AI in VegaLeaders.js to use dynamic pricing when deciding to hire leaders
- Add research videos for ursaal, kkrix, and mekhan factions
- Enhance leader screen UI with:
- Dynamic cost display reflecting current scaling
- Fleet labeling showing ship composition and location
- Assign/Reassign buttons with list picker for colonies (admins) and fleets (captains)
- Unassign button for posted leaders
- Better location display for assigned leaders
- Add full diplomacy subsystem: tiered gifts (BC payments with diminishing returns), trade agreements (passive BC income, coexists with treaties, cancelled on war), and espionage mission toggles (Off/Steal/Sabotage) in the Audience screen.
- Implement whole-galaxy diplomacy passes: border and fleet proximity tension, fleet intrusion tracking with escalating penalties and force-opened Audience complaints, and "enemy of my enemy" attitude triangulation.
- Update Audience screen UI with new action buttons, gift picker, intelligence toggle, and species-specific chat dialogue for all diplomacy events.
- Add research choice prompts for ambiguous tech frontier picks, surfaced before the turn report.
- Improve star map: Delaunay triangulation for controlled-space range darkness, staggered docked fleet markers, expanded pan slack to clear UI chrome, and zoom-aware colony info labels.
- Backward-compatible save serialization for new diplomacy fields.
- Add comprehensive test coverage for all new logic, chat completeness, and AI soak validation.
- Add 21 new alternate techs across all 6 fields (computers, construction,
forcefields, planetology, propulsion, weapons), using icon frames 78-96
- Set research video paths for human and kestrelli species
- Expand tech sprite sheet from 8 to 10 rows to accommodate new frames
- Add fieldTechLevel() to VegaLogic: displays per-field level as highest
known tier × 5 plus extra known techs below the frontier
- Display field level in research screen headers
- Add verification tests for 3-way branching rungs, fieldTechLevel calculations,
and refine availability roll tests to measure unbranched rungs only
- Replace vertical "research ladder" list with a visual tech tree rendered as
connected node badges showing prerequisites and branch structure
- Add `initialField` parameter to `openResearchScreen` to open on a specific
category instead of always defaulting to the first one
- Add "View Research" shortcut button on turn report when a tech completes,
jumping directly to that field
- Add Cerebrai species research video (research-cerebrai.mp4)
- Add branching tech alternatives across all six fields (computers,
construction, forcefields, planetology, propulsion, weapons). Each
branched tier has 2+ techs sharing the same rung; researching any one
clears the rung while leaving siblings researchable as side-picks.
- Correct racial skill to affect tech COST only (Poor 125%, Average 100%,
Good 80%, Excellent 60%), not availability. Availability is now a flat
50% roll per tech (75% for Cerebrai/psilon analogue), with a safety
net that no branched rung can end up with zero available options.
- Add techRungsByField data structure and rungCleared/canResearch/setResearchTarget
logic so the research screen lets players pick among open rung alternatives
without losing banked beakers.
- Update sprite sheet layout (+1 row, frames 66-77 for new tech icons).
- Add comprehensive verification tests for cost buckets, rung gating,
setResearchTarget behaviour, and statistical availability rates.
Introduce colony naming with a 50-name bank per species, shuffled
per-empire and displayed as defaults in naming prompts. Colony names
now appear across all UI layers (star map, side panel, colony view,
system view) with the astronomical designation as fallback.
Add three pure-flavor conversation topics to diplomacy: "Their People"
(neutral+), "Their Homeworld" (friendly), and "A Story" (friendly).
Each diplomacy-capable species receives unique lore-rich responses
(3 variants each) reflecting their culture, homeworld, and history.
Other changes:
- Switch ursaal videos from generic "human" assets to species-specific
- Widen diplomacy buttons to 200px; add auto-shrink for long labels
- Add ±5 population transport buttons to colony view
- Validate colonyNames (50+ entries, no duplicates)
- Enforce minimum lore variant counts in verification
Brian Fertig