Evidence notebook: facts, calculations and open questions

What an evidence label should tell you

A source-linked value means an editor checked a document for that claim, not that a player reproduced it in a controlled match. A calculated value follows disclosed inputs. A recommendation is an editorial choice under stated assumptions.

The source ledger retains links and review dates. When a community page mixes old and current values, the relevant official change takes priority. We do not average conflicting values: the average of two versions may be a value that neither used.

Worked audit: phase 2 ends at 13:00

Inputs are initial delay 90 seconds, phase 1 wait/shrink 240/270, and phase 2 wait/shrink 80/100. The sum is 90 + 240 + 270 + 80 + 100 = 780 seconds = 13:00. This is a derived chart position, not a measurement from landing.

Challenge the input, the arithmetic or the timeline origin separately. A screenshot taken 13 minutes after landing does not establish the same origin. A Ranked screenshot does not verify the Normal chart. The calculation is useful because those assumptions are visible.

Questions we leave open

A complete time-in-zone damage formula is not supplied by the cited official notes. We therefore do not generate survival estimates from a guessed curve. Actual damage, protection, health and interrupted healing matter.

A full probability distribution for every supply source is not established here. Naming possible contents is different from assigning odds. Availability across every map, event and custom preset is also not a verified universal list.

The route speed in the workbench is not benchmark data. For your own observation, record platform, movement state, route length, start/end definitions, repeated durations and interruptions. Keep observations separate when equipment or terrain changes.

A correction that can be checked

Include the URL and sentence, expected replacement, and an official source or reproducible observation. State platform, mode and date. For a tool issue supply the fields and displayed result; account details are unnecessary.

A patch can invalidate a previously correct row. Update scope and source, then recheck calculations that depend on it. The ledger and example expose that dependency instead of silently replacing a number everywhere.

Resolve one disagreement before copying a whole page

Create a claim ledger with columns for the exact claim, platform and mode, source date, source passage and confidence limit. If one source says an item has changed, identify which field changed: use duration, health endpoint, availability or another property. A later date alone is not enough when the newer document discusses a different mode or a different field.

For example, one document could describe a historical use duration while another changes the health effect. Replacing both fields just because one is newer would introduce an unsupported edit. Quote or summarize only what the relevant passage establishes. When no document resolves the conflict, preserve the disagreement in the note rather than choosing whichever number makes a cleaner table.

Confidence has a scope. An exact sum can be highly reliable given its inputs while applicability to an unobserved custom match remains unknown. Labeling the sum “uncertain” without explanation loses the useful part; calling the entire match prediction “verified” overstates it. Write separately that the arithmetic is reproducible and the live ruleset must still be checked.

Site revision log · 12 November 2026

This revision corrects page-specific language destinations in both header and footer, gives articles contextual follow-up links, and makes battlegrounds.click the primary identity. Original diagrams replace promotional lead images. The reading guide, vocabulary, decision-map and PC comparison pages receive worked analysis; the calculator keeps its assumptions visible. These are editorial and website changes, not a claim of new gameplay testing.

The release checks local files, repository-subdirectory preview and domain-root HTTP preview. A production-domain check and an inbox-delivery check have different evidence requirements. The public record does not convert a successful local test into a claim that either external service is configured correctly. Future corrections should preserve the same distinction.