Guide

Stock-signal red flags

The tells that a service cannot be trusted, whatever its banner says.

Every one of these is a version of the same problem: the claim cannot be checked. Spot two or three together and the win-rate number on the homepage stops mattering.

  • Only the trades that came good ever reach the page; the ones that went wrong are never named.
  • Entries are vague enough — “long around here” — to score almost any outcome as a win.
  • A huge win-rate number sits on the page with no signal count beside it.
  • There is no drawdown figure anywhere, on a strategy whose whole story is risk.
  • The record lives in a chat that scrolls away and cannot be audited after the close.
  • Investigate any actually disclosed broker affiliation and its economic terms. A link alone does not prove an undisclosed payment, and subscription revenue does not prove the absence of commercial incentives.
  • “Proprietary” is used to avoid explaining the method at all.
  • No named person or credential stands behind the calls.
  • Nothing is timestamped, so any call could have been posted after the move.

Public product and reference documents inspected 12 October 2026. No paid software, selected broker route or subscriber account was tested. Where an original archive, audit report or selected contract is absent, this edition keeps that boundary explicit rather than inferring a completed product review.

The flags, mapped to the tests

Why the flags cluster by service type

No measured service-type pattern is established by this guide. The original matrix is now a request map: source versions for message streams, contract/account records for copying, origin records for reposts and frozen settings for algorithms. A channel can provide useful evidence and an algorithm can lack it. Inspect the named source instead of predicting its integrity from the interface.

Equity evidence starts with the security, original instruction and executable quote. A confirmed hash receipt establishes only its covered bytes by the attestation time. For VR2, target and stop are not covered; its existence does not establish stock-product suitability or an account outcome.

The category map is a consumer request list, not measured pass/fail data. Original evidence can come from several formats; a missing field stays unresolved for the specific product rather than turning every member of a category into a failed service.

Use the matrix to choose the relevant evidence request, not predict misconduct or certify a winner. A source with visible price still needs appropriate instructions and account records for a performance claim. A missing independent timing bound leaves chronology unresolved unless other evidence resolves it. One documented field does not settle every other question, and an entire copying or chat category is not automatically unverifiable.

How to weight the flags

Weight a limitation against the claimed task. A complete-return purchase needs a population and economic record; a candidate scanner needs its original universe and settings. An absent field can justify withholding that claim without accusing the operator of fraud. Preserve the original flags as follow-up questions, identify the source-specific missing record and avoid adding up unsupported character judgements. Evidence can improve when an actual archive, version record or account ledger is supplied; record its date instead of pretending it existed at the first review.

The clean way to act on all of this is the positive checklist rather than the negative one: run the four steps in how to verify a record, and a service either survives them or does not. The flags above are simply the fast version — the patterns that tell you a service will fail step four before you bother running it.

Risk indicators need specific evidence

The original flags remain an investigation checklist, not proof that whole formats are dishonest. A documented missing field supports a narrow limitation; a commercial relationship needs its own source. An unverified service can be unsuitable for a performance purchase without the publication accusing it of fraud.

Fictional profile has an identified publisher and visible billing, but no original instruction inventory. Record two documented fields and one unresolved population question. Do not convert that into two stars, a passed financial audit or a criminal conclusion. Ask for the missing inventory and preserve the specific limitation in the current comparison.