Investigation Architecture
The full skeleton of your scenario — locations, clue web, puzzles, checks and NPC roster — before any prose gets polished.
Due: 23 April 2027 · Weight: 25%
The brief#
Submit the complete skeleton of your scenario: every scene, every clue, where puzzles and checks sit, and who your NPCs are and why.
This is the studio work from weeks 4–9 assembled into one document: your committed structure, your scene map, your clue web tagged against both taxonomies from week 6, your puzzle placements, your reworked checks, and your cut-and-sharpened NPC roster. Nothing needs finished prose yet — this is graded as engineering, not writing. What matters is whether the pieces connect: whether a player who fails a roll or misses a clue still has a way forward.
What you submit#
A single document (or linked set of documents) covering: scene map, clue web with taxonomy tags, puzzle list, check list with failure text, and NPC roster with one-sentence purposes. This is also the document your playtest in week 11 will be run from, so it should be usable at a table, not just readable.
See the clue web explorer for a worked example of a tagged clue web, including a redundant path to a core clue.
The spec
The brief poses the problem and leaves room for a response. The spec is the fixed part: what a marker checks. Some lines can be checked mechanically; the rest call for judgement.
- your scene map, clue web, puzzle placement, check design and NPC roster are submitted as one connected document
- every load-bearing clue has at least one redundant path to it
- every NPC's one-sentence purpose is stated explicitly
How it is marked
| Criterion | Weight |
|---|---|
| Clue web has redundant paths to every load-bearing clue | 40% |
| Structure and scene map are internally consistent | 25% |
| Checks are designed so failure changes cost, not access | 20% |
| Every NPC has a stated one-sentence purpose | 15% |