Scorecard

The criteria, in detail

These criteria retain the original research questions and dedicated routes. They do not establish that most services fail or that one marketing format is uniquely honest. Reference: method page

Sorting the field

What the scorecard is actually sorting

Options evidence is contract-specific. These are functional product assessments, not audited performance rankings. VR remains an unranked model-record example because an options subscription and its contract coverage have not been confirmed. The framework is a document request list, not measured category-wide pass/fail data. Inspect an identified service and label missing evidence as unavailable.

ObjectDeclared findingEvidence boundary
MessagingOriginal contract/leg versions and deliveryNo presumed deleted losses
CopyingSource/follower contract and quantitiesNo automatic options mapping
Social eventOrigin, chronology and stated inferenceNo assumed incentive model
AggregatorParent records and amendment linksRepetition is not separate validation
AutomationFrozen inputs, rule and account transitionsNo assumed live performance
Claimed reviewed deskNamed reviewer, actual scope and full inventoryAudit label alone insufficient

The formats below have different evidence objects. No single category is the only one outsiders can investigate. Original versions, account histories and frozen rule settings can support different bounded findings. Reference: pick

Why these two columns decide most cases

An options record needs the original contract terms and broker-side position evidence. A completed timestamp attests only to specified bytes by its block time. VR2 does not cover stop, target, option legs or a subscriber fill, so its presence cannot justify an options-service score.

Read each test

The three tests with their own page

Criterion

Sealed before the outcome

Why a public timestamp - with the grade hashed inside it - is the test a service cannot fake its way around.

Criterion

A defined-risk structure

What a defined-risk call looks like, and why a bounded worst case is the first thing an options buyer should demand.

Criterion

Grades that are measured

How an A-to-D conviction grade is calibrated to a model's own returns instead of a mood word.

Complete population and incentive records can require substantial work. They are not automatically easier to verify than chronology. Retain missing records and their actual source dates. Reference: method page

The map is a request inventory

An original alert needs contract and amendment records. An account-copying product needs source and follower routes. A rule platform needs frozen settings. A news tool needs source and event chronology. These are different request objects; no category receives a universal yes or no.

Fictional row: a news event has a documented originating timestamp but no option instructions. It is not a failed options-return record when it never claimed that job, and not a verified trading edge either. Match evidence to the actual selected function.

A completed format comparison

Fictional message source retains the first contract pair, amendment and actual receipt time but supplies no subscriber fill. The documented object is the instruction chronology. A separate fictional account source supplies fills but no parent publication record; execution is documented while original publication chronology remains unresolved. Neither format earns a universal yes in every column.

A reviewer can retain both findings with their source dates and population limits. Do not replace this evidence ledger with a tick because a service calls itself audited, or with a cross because it uses a messaging app. The underlying source objects, not the interface labels, determine the conclusion.