Chronicle map-data methodology notes
derived — Derived from our own records and analysis rather than an outside document; the method is described below.
Type: internal-recordArchived: 2026-09-03id: chronicle-map-methodology
Not a public web page — held only as this archived copy
The method used to distinguish real wars from free-for-all trading in the exchange log, and to infer alliance activity windows. Recorded with the draft change sets in data/chronicle/drafts/.
Text as archived (1,505 characters). This is the copy the wiki's citations were checked against; the original may since have changed.
# Chronicle map-data methodology notes
The reasoning behind alliance and event entries that were derived from the
territory-exchange log rather than from a written source, kept alongside the draft
change sets they justify.
## The core problem
Raw exchange counts do not distinguish a war from routine churn. Certain territory
clusters are contested continuously by guilds with no hostility between them —
free-for-alls that generate enormous volume. An early pass at this record mistook
that churn for warfare, and the correction is why the method below exists.
## The method
1. For a given window, rank territories by exchange volume and set aside the
highest-churn set as free-for-all activity.
2. Look for sustained fighting across the *remaining* quiet territories. A real war
shows up as many ordinarily-peaceful territories changing hands repeatedly
between the same two groups.
3. Distinguish war from alliance trading by symmetry and location: allies trade
high volume on hot territories in a near-even exchange; enemies fight across
quiet ground, often with one side losing net holdings over time.
4. Treat the resulting bloc boundaries as *inference*. Guild pairs can be
established this way; alliance names and rosters cannot, and are never invented
from map data.
## Standing caution
Where an article's claim rests on this method, the prose says so — "map-data
analysis places the war's start near..." — and it is never presented as though a
source had recorded it.
