Every wiki entry is scored on how complete it is, and the results feed a worklist that tells maintainers where the gaps are. It exists so that filling in the wiki is a queue you can work through rather than a guessing game about what is missing.

The tools live under Data health in the dashboard.

How the score works

Each category defines which fields matter and how much. Fields carry one of three weights:

ImportanceWeight
Crucial3
Important2
Optional1

The score is the weight you filled in divided by the total available weight, as a percentage. A creature scored on type, health, level (crucial, 3 each), plus description, damage and attacks (important, 2 each) has 15 points available. Filling everything except damage gives 13 of 15, or 87%.

A field counts as filled only if it holds real content. Empty strings and zeros count as missing, so a numeric field left at 0 does not earn its weight.

Grades, and the crucial-field rule

GradeWhen
Complete100%
Good80% or more, with every crucial field filled
Incomplete50% or more
StubBelow 50%

The part that surprises people: a missing crucial field caps the grade at Incomplete no matter how high the percentage is. An entry at 90% with one crucial field blank is Incomplete, not Good. Only a missing crucial field can hold back an otherwise well-filled entry like that.

This is deliberate. An entry that looks nearly finished but has no health value on a creature is not more useful than one that is obviously a stub, it is more misleading, because it looks trustworthy at a glance.

The practical consequence for maintainers: fix crucial gaps before topping up optional ones. Filling three optional fields on a 90% entry moves the number and changes nothing about its grade. Filling one crucial field can move it two bands.

The worklist

The worklist is every entry below 100%, worst first, filterable by category and by grade.

Working it grade-first is usually the most productive order. Stubs and entries with missing crucial fields are where the wiki is actually unusable; a Good entry missing one optional field is fine to leave.

Scores are recomputed by a scheduled job, not on save, so an entry you have just filled in will keep its old score for a while. There is a refresh action that queues a full recompute if you want to see the effect of a batch of edits immediately.

The uncategorized backlog

Separate from completeness, there is a triage list of items that have been seen in the game but do not yet belong to a wiki category. These typically come from loot and discovery data referencing something the wiki has never heard of.

Triage is two decisions:

  • Categorize it. The item belongs in an existing category, so create the entry there and mark the backlog row as categorized, which drops it from the queue.
  • Leave it. Not enough is known yet. It stays in the backlog.

The backlog is worth working regularly rather than in bursts. An item sitting uncategorized means anything referencing it links nowhere, so the cost is spread across every page that mentions it.

Marking entries obsolete

Content leaves the game. An entry can be flagged obsolete, which immediately drops it out of the completeness report.

This matters for the numbers: without it, removed items sit in the worklist forever as permanently incomplete entries nobody can ever finish, and they drag the category rollup down. If something no longer exists in game, flagging it obsolete is the correct action, not deleting it, since old globals and loot records still reference it.

The flag can be reversed if something returns.

Reading the summary

The summary gives headline numbers plus a per-category rollup. The rollup is the more useful of the two, because completeness is very uneven between categories: a category that was imported wholesale tends to sit at a uniform middling score, while a hand-built one is bimodal, with well-finished entries alongside untouched stubs.

Chase the categories where crucial fields are missing in bulk. That is where readers actually hit dead ends.