signals/periphery
00:00:00
SIGNAL
How this site works

Methodology.

This page explains how the things on this site are made: the instruments, the briefings, and the reading room beside them. It is deliberately short. If something here is unclear, or you think the method has failed somewhere specific, the Corrections log is open.

How an instrument is built

The core of this site is a bench of interactive instruments. Each one computes the actual equations in your browser rather than replaying an animation of a result: turn a dial and the numbers are recalculated. Constants are taken from published sources and named on the instrument. Where a constant is fitted rather than published, the bench says which one, and the fitted value is usually the finding itself.

The numbers on an instrument's screen, and the numbers quoted in the copy around it, are re-derived from the instrument's own data file by a verification script kept beside it and run before it ships. That rule exists for one reason: the copy cannot quietly drift from the arithmetic. Change the data or the maths and the check fails first.

How a bench prices a claim

Every claim put on a bench is labelled with what kind of statement it is: a figure reported from its source, a measurement anyone can check, a modelled result carrying assumptions, or a reading. Modelling assumptions are printed on the instrument's own face rather than buried in a footnote. A reading is a way of taking the evidence, not a fact, and it is marked as one. These are plain status words, not the evidence tiers below.

When a bench tests a book, the book's own numbers are quoted with their chapter or page, so you can go back and check them. When the arithmetic is done, the competing readings are set out at equal size. The instrument prices; it does not vote. Which reading holds is yours to decide.

How a document is selected

The site works from the public record. A document earns a briefing when two things are true. First, it is a primary record: a government, military, contractor or agency file, a cleared book, a hearing transcript, or a long-form interview. Second, it is either new to the public record, or it has not yet been read carefully in plain language. Preference goes to documents that can be linked back to an original a reader can open. The site does not try to cover every file. It tries to cover the ones a reader would want explained properly.

How a claim is checked

Every quote is taken verbatim from the document or transcript it is attributed to. Dates, names, document numbers and figures are read from the primary source itself, not from press coverage of it. Where a document is ambiguous, the briefing says so, rather than resolving the ambiguity for effect. Where press coverage and the document disagree, the document is followed and the discrepancy is noted. Each briefing separates, in plain terms, what a document establishes from what it only suggests. Nothing is presented as proven that the source does not prove.

The evidence tier

Every briefing and note carries an evidence tier that rates how direct and verifiable its underlying source is. The tier rates the source, not the verdict: a strong tier means the source is solid, not that anything anomalous occurred, and not that the site endorses a conclusion. The tiers are set out in full on the Evidence Standards page.

The reading room

Culture comes in two forms, and which one you are reading is stated on the piece. A signed review covers work I actually care for; the opinion in it is mine, and I say what I liked and what I did not. A neutral source note covers the contested books and interviews behind the UAP record: what the book says, where it says it, sourced, with no verdict attached. Both are read against what the source actually says rather than what the headlines made of it, and where an instrument has priced the numbers in one, that instrument is linked alongside.

What this site will not publish

The site will not present unsourced rumour as fact. It will not name private individuals who are not already public figures in the record. It will not claim that a document or an instrument proves more than it does. It will not chase speed at the cost of accuracy: a thing is published when it is right, not when it is first. When the site gets something wrong, the error and the fix are recorded on the Corrections page, rather than quietly edited away.

This site is a single-author project. It has no special access and claims none. Its only method is careful reading, arithmetic shown in the open, plainly explained, with every original's location named so you can read it yourself.