Shared branch structure

03 A Harrowing Escape │ ├── escape_choice = 1 ──► 04a The Swamp of Esten ──┐ │ │ └── escape_choice = 2 ──► 04b The Midlands ─────────┤ ▼ 05 The Oldwood

Both branches use the same protagonist level, starting gold, default schedule, primary/extra gameplay music, 40% carryover, shared exile story panel, recalled companions, victory objective, winter deadline, and final strategic conclusion. Their scenario-specific systems are intentionally very different.

Direct comparison

Dimension04a — Swamp of Esten04b — Midlands
Map37 × 42; narrow/long wetland42 × 40; broad open region
Turns39 / 36 / 3336 / 33 / 30
VisibilityFog enabledNormal visibility
EnemiesThree Dark Sorcerer cult leadersOne Orcish Warrior and two Orcish Warlords
Primary tensionUncertainty, exploration, ambushesOpen fronts, reserves, night aggression
RandomizationTemple outcomes randomizedNo route-level random system
Scripted surpriseScorpions spawn near vision edgeReserve waves appear after chiefs are attacked
Optional allyMinister Edren, White MageNone during the battle
Victory knightSir RuddrySir Ladoc
Transition route markerlast_done="River"last_done="Midlands"
Next scenario05_The_Oldwood

What changes—and what does not

LayerSharedBranch-specific
NarrativeExile, winter pressure, road blocked, Oldwood convergenceHistorical geography, enemy faction, route knight
Player stateHaldric, Burin, optional Lady Outlaw, recruit list, gold carryoverMinister Edren only on river route; Ruddry versus Ladoc; last_done
Engine systemsCore AI, standard objectives, default schedule, journey/death utilitiesFog and visibility filtering versus attack-triggered nested reserve events
AssetsShared exile panel and music paletteOne route story panel and one knight portrait each

Branching-effort analysis

Design interpretation. These branches are expensive where the player notices—map, faction, encounters, dialogue, and tactical rhythm—but economical in campaign architecture. They share the entry point, core cast, balance frame, and convergence point.
Authoring strategyBenefit
Share campaign-state assumptionsBoth files can recall the same companions and inherit the same recruit list.
Use a common ending destinationLater scenarios do not need two complete versions.
Give each route one signature systemThe choice feels meaningful without requiring different core rules.
Record the completed route for the transitionTransition or journey logic can distinguish the completed branch before Oldwood clears the marker.
Introduce route-specific loyal unitsThe branch leaves a durable mechanical souvenir.

Which is the better learning example?

Learning goalBest routeReason
Fog and exploration designSwampVisibility edge, randomized discoveries, distributed objectives
Advanced WML event logicSwampNested events, location storage, search loop, delayed substitution
Multi-front combat pacingMidlandsThree visible leaders, reserve composition, survivor economy
Declarative AI tuningMidlandsTime-of-day aggression and difficulty-expanded rosters
Branch architectureBoth togetherMeaningful divergence followed by controlled reconvergence

Swamp WML · Midlands WML