# Building a magical world for Meok and Sumbi

11 October 2026. A practical guide for building a game and manga world together. The proposed examples below are writing tests; they do not replace approved chapters or settle canon. The accompanying world workshop develops one connected coastal region rather than a whole planet.

## Begin with a place worth living in

A setting needs pleasures as well as threats. Choose something Sumbi wants to do on an ordinary afternoon, something Meok can enjoy being good at, and someone whose work lets that day happen. A shell-music festival, harbour reach contest or lantern-market meal can reveal customs without stopping the story. Later danger matters because it threatens a life we have already enjoyed.

Mashima describes writing desired friendships into *Fairy Tail*. Kikuchi distinguishes the game's overlapping sizes of goals from the manga's variable episode lengths. Our application is to build places that produce both enjoyable relationships and playable objectives, rather than treating lore as a substitute for either. [Creator and producer interview](https://blog.playstation.com/2020/08/06/fairy-tail-creator-talks-manga-beginnings-the-new-game-and-his-dream-party/)

**First exercise:** Write a short scene where no villain appears. What do they do, buy, eat, make, compete over or look forward to? If the world is only interesting during catastrophe, give it another source of appeal.

## Follow one magical possibility through ordinary life

Sanderson's Third Law encourages expanding existing magic and its connections to society before continually adding more. This is personal craft advice, not a rule every fantasy must obey. [Third Law of Magic](https://www.brandonsanderson.com/blogs/blog/sandersons-third-law-of-magic)

For this project, try a shared tide-and-road network. Its consequences can extend through several kinds of scene:

| Magical possibility | Ordinary dependency | Conflict | Playable consequence |
|---|---|---|---|
| A crossing opens on a schedule | Traders plan market days around it | A town delays another town's passage | Reach and hold the closing gate |
| Water can be routed between levels | Farmers and lower workshops share maintenance | One district survives by flooding another | Repair a second outlet before changing the flow |
| Lamps indicate an active route | Children and visitors can find their way | A saboteur points travellers into danger | Compare the lamp sequence with physical water marks |
| A ward arrests Ink movement | Watch crews can contain an attack | A harmless person is detained too | Break its exposed control without harming the watch crew |

The examples are our proposals. Freshness comes from the specific consequences, costs and human responses, rather than the number of invented terms.

## Make the geography usable before drawing it

In the *Fairy Tail* interview, Mashima acknowledges that Magnolia's appearance varied; Kikuchi describes months spent reconciling views to build the town in 3D. A game benefits from deciding where recurring places are before producing many illustrations. [Mashima and Kikuchi](https://blog.playstation.com/2020/08/06/fairy-tail-creator-talks-manga-beginnings-the-new-game-and-his-dream-party/)

Make a simple route diagram with an entrance, visible destination, safe return path, obstacle, optional branch and landmark. Record elevation, water direction and the materials that remain consistent between scenes. A stone town can have roots and gardens without becoming an unrelated forest in the next shot.

Nintendo's developers describe how new caves and vertical travel change exploration of a familiar landscape, and how dungeons carry regional character. [Tears of the Kingdom developer interview](https://www.nintendo.com/au/news-and-articles/ask-the-developer-vol-9-the-legend-of-zelda-tears-of-the-kingdom-chapter-3/)

Our test: revisit one crossing after an ability improves. Meok should notice a new reachable anchor, or Sumbi's scouting should reveal a route previously hidden from ground level. This gives growth a visible effect without requiring another continent.

## Keep mysteries separate from basic comprehension

Smith and Worch show how related environmental evidence lets players infer events and how a space can telegraph hazards. Their approach invites interpretation; it also gives designers ways to communicate practical information. [What Happened Here, GDC slides with notes](https://media.gdcvault.com/gdc10/slides/Smith_Harvey_WhatHappenedHereWeb_Notes.pdf)

For every important mystery, write five separate answers:

1. **Author truth:** What happened, or which part remains deliberately undecided?
2. **Public belief:** What do ordinary people think, and what experience makes that credible?
3. **Deliberate lie:** Who knows their claim is false, and what do they gain?
4. **Discoverable evidence:** What can Meok and Sumbi actually see, use or compare?
5. **Changed decision:** What becomes possible or harder after the discovery?

Example: a small monster emerges through a drainage opening. Its trail points toward misrouted water; an invasive mark disrupts its movement. Later, the same mark appears on a larger creature and a tampered valve. The reader can connect these events before knowing the corruptor's full identity. The immediate goal remains clear: get civilians across and stop the attack.

Justin Alexander recommends multiple clues for an important conclusion so a missed detail does not stall a scenario. His original context is tabletop play. Our adaptation is to support a major revelation through different channels, rather than repeating an identical speech. [Three Clue Rule](https://thealexandrian.net/wordpress/1118/roleplaying-games/three-clue-rule)

Former coexistence might appear in shared architecture, a travel permit and an ordinary family letter. Each clue adds meaning; none requires the characters to become historians before acting.

## Give factions services, arguments and disagreements

This faction checklist is our synthesis, not a quoted formula. Record what a group provides, why a decent person joins, what it controls, what its leaders conceal, what its methods cost and where its own members disagree.

A watch service can save lives while using a neutralizing device unjustly. A refuge for gifted children can offer real education while its inner circle prepares coercive service. A lower-town assembly can demand fair water while hiding an evacuation that will endanger another district. A villain can exploit a valid grievance and still choose deliberate cruelty.

Avoid making every kindness a trick. A healer can remain kind after rejecting the ritual leader. A guard can revise their view. An Ink citizen can oppose Meok's hope of restoration because they have already lost someone. These differences create conversation and action without pretending all positions are equally justified.

## Separate a power's limit, weakness and cost

Sanderson distinguishes a power's boundaries, exploitable vulnerabilities and consequences of use. These are useful design categories, rather than finished combat balance. [Second Law](https://www.brandonsanderson.com/blogs/blog/sandersons-second-law)

For each Sumbi or Meok ability, define the target, visible preparation, successful effect, interruption, recovery and counterplay. A scouting link might leave Sumbi less aware of her own surroundings; revealing an enemy opening still requires Meok to act at the right moment. Those are candidate constraints, not current canon.

Use the same discipline for destiny. An old binding reacting to Sumbi proves that a mechanism recognizes her; it does not automatically prove a god assigned her a moral duty. Record what happens if she refuses, who interprets the reaction, and what an alternative solution would require. Avoid resolving it with a new power introduced only at the final crisis.

## Build and test one small region

Use six connected places at most for the next workshop: home, watch road, spring settlements, lower market, routing dungeon and coastal archive. Give each a daily activity, a resource relationship, a navigable route, a danger and a clue.

Then test one ordinary scene, one minor fight and one discovery. Ask:

- Did we enjoy spending time here?
- Could a player tell where to go and why?
- Did the danger have a visible cause and response?
- Did Sumbi and Meok make different understandable choices?
- Did the discovery change an action or relationship?
- Which single unresolved question makes us want the next chapter?

Only expand the map when a character's need or the consequences of an existing choice lead somewhere new. Keep owner decisions, working proposals, in-world beliefs and unresolved questions visibly separate as the world develops.