The tests, in detail
Three original evidence questions have dedicated pages; the full framework remains on the method page. Their presence does not establish that the entire field was tested or that a category automatically fails. The expanded forex layer investigates economic and operational records separately.
Related retained references: method page
What the five tests are really sorting
The original service-type matrix remains as an investigation map, not a factual verdict about every member of a category. Ask which original record, source account, instruction population or learning programme is represented. A polished site cannot answer a missing record question, but neither can a blanket category fail. The table below separates the appropriate requests and limits for each format.
| Record | Represented finding | Boundary |
|---|---|---|
| Messaging-app channel | Request original versions and expected inventory | Do not infer deleted losses from format |
| Copy-trading / PAMM route | Request source/follower or managed-account records | Different operating contracts must remain separate |
| Social-media caller | Request original identity, chronology and relationships | No category-wide incentive finding |
| Signal aggregator | Request origin mapping and amendments | Forwarded record is not a new original |
| Algorithmic service | Request frozen method and actual evaluation population | A backtest is not a subscriber account |
| Organiser-referenced publisher | Keep contest, instruction and account objects separate | Founder evidence does not audit the later stream |
The bottom row is the only category that answers both questions, which is the structural case the guide makes for the desk we back — not that it is louder, but that it belongs to the one type a stranger can confirm. The three tests below each take one column apart in full.
Why these two questions decide most cases
Of the five tests, two carry most of the weight. An outside witness is the one that cannot be retrofitted: either a party other than the seller saw the forex trades on a live account, or no one did, and a marketing team cannot manufacture an organiser's leaderboard after the fact. A real denominator is the one that cannot be faked without an outright lie: a win rate is only evidence when the full count of calls, losers included, sits beside it. A service that answers both has handed a reader a record they can interrogate. The remaining three — locked at issue, a measured grade and clean incentives — are real and they matter, but they tend to confirm a verdict the first two have already reached rather than overturn it. That is why a service can run a slick site, an active community and a confident pitch and still land in a row of crosses.
The three tests with their own page
Witnessed outside the room
Why the strongest forex evidence is a record someone other than the seller has already confirmed.
Locked before settlement
Why a public timestamp is the test a fast forex call cannot edit its way around.
Grade definitions and calibration questions
Related retained references: Grades that are measured
Investigate the actual frozen label definition, calibration boundary and forward evidence; none is supplied simply by quoting a letter.
The remaining two tests — pricing on a public page and clean, non-rebate incentives — are covered on the method page, because they are quicker to confirm and rarely the deciding factor.
Read categories as evidence requests
A chat needs originals and versions; a copier needs source/follower mappings; an aggregator needs origin trails; an algorithmic product needs its actual method version and evidence period. No category automatically loses its losers or supplies an honest denominator. The retained matrix now names missing requests rather than assigning unsupported yes/no passes.
The outside witness must match the asserted record. A founder competition result can be useful background while a current instruction population is still unverified. A time-bound packet can protect some fields while others remain outside it. Preserve those scopes rather than awarding the featured desk all five questions by implication.
One fictional mixed-evidence packet
Imagine a seller provides a founder contest row, one completed earlier-bound instruction packet and no full expected inventory. The first object supports its contest statement; the second supports its covered timing. Complete population, grade calibration, selected renewal and relationship claims remain unresolved. There is no arithmetic five-pass result.
This example belongs to no real completed review. It demonstrates how a consumer can preserve useful evidence while asking for what is still missing. Missing material should stay unknown rather than being converted into an assumed loss or an invented all-clear score.