Editorial policy

This publication writes about money, which means a confident sentence that happens to be wrong can cost a reader something real. The process below exists to make that harder, and it is enforced by the publishing software rather than left to good intentions.

Sources we will use

In order of preference: financial regulators and official bodies; peer-reviewed and open-access academic work; and the official documentation of the platform or product being described. Platform documentation is treated as evidence of mechanism only — what a piece of software does and does not do. It is never evidence that a tool is profitable.

We do not cite a vendor's marketing as evidence for anything except what that vendor says about itself, and we say so when we do. We do not cite other blogs, content farms or ranked lists. A URL that looks relevant is not a source: if the page does not say it, it does not support it.

The research ledger

Before an article is written, each factual claim it needs is entered in a ledger with its source, the publisher, the date it was read and the quoted words that support it. Each claim carries one of four statuses:

• Verified — someone read the supporting sentence in the source document and it supports the claim as written.
• Qualified — the source supports something narrower than the claim originally said, so the claim was rewritten to match the source.
• Editorial — a framing or definition this publication adopts, resolved by precise wording rather than by evidence. It is presented as our standard, never as a research finding or a regulatory requirement.
• Unresolved — no source yet. An unresolved claim blocks the article, and is shown on the page while the article is unpublished.

A figure is only recorded if it was read in the source itself. A number that arrives through a search-result summary, a press release or another publication's account of a study does not qualify, however widely it is repeated — and if it cannot be traced, it does not appear on this site at any strength of hedging.

Before an article is published

Every article moves through four states, and the last step is a person:

• Draft → Research → Review → Published.
• An article cannot enter Review while any claim in it is unresolved.
• It cannot be Published without at least one verified source, no outstanding research notes, and a recorded approval naming a real editor. The software rejects the publication's own name in that field, so "approved by Trading Tech Review" is not an approval.
• Articles are never published in bulk.

What we will not publish

• An invented author, biography, credential or qualification.
• A statistic, study, quote, ranking, search volume or price that we have not read in a source — including ones that look harmless.
• Our own performance figures, or any claim that a tool beats the market.
• A guarantee, a promise of returns, or any suggestion that something is risk-free. We do not describe ourselves as having tested a tool unless we actually did.
• Personal financial advice, signals, or recommendations to buy or sell anything.
• Review, rating or FAQ structured data. Our pages describe only the organisation, the site, the breadcrumb trail and the article itself — we do not emit markup implying scores we never awarded.

Images

Diagrams are drawn for the article they illustrate and are generated here, not licensed from a stock library and not scraped. Every image carries alt text describing what is visible in it rather than what the article argues. We do not publish artwork containing a readable balance, price, profit figure or equity curve, even a decorative one: an invented chart is a performance claim, and it would be a performance claim we could not source.

Corrections

If something here is wrong, we want to fix it rather than quietly adjust it. Write to the editorial desk with a link to the article, the specific claim, and a source if you have one.

A factual error is corrected in the text and the correction is noted on the article. A claim that turns out to be unsupported is cut or narrowed to what its source actually says — we do not leave a weakened version of a wrong sentence standing. Where a change alters the substance of an article, the revision date on the page changes with it.

Commercial relationships

Trading Tech Review has a commercial relationship with TheNextLevel — this publication is produced as part of a marketing project for it. Where you see a link to them on this site, it is there because of that relationship, and it is labelled as promotional at the point you meet it. We say so plainly because readers are entitled to know who is behind what they read.

What the relationship does not do: it does not decide which subjects we cover, it does not change a claim that is backed by a source, and it does not buy a recommendation — we do not rank or score tools for anyone. We never repeat TheNextLevel's performance figures. Their links carry rel="sponsored nofollow noopener" and pass no ranking credit, and while this arrangement exists we do not describe this publication as independent.

Whatever the arrangement, it never decides which tools are covered, never changes a sourced claim, and never buys a favourable description. Promotional placements are labelled as such, appear on a minority of articles, and link out with rel="sponsored nofollow noopener" so they pass no ranking credit. We will not describe this publication as independent while any commercial relationship exists. The commercial disclosure page carries the current position in full.

Artificial intelligence in our own workflow

We write about AI tools, so it would be poor form to be vague about our own use of them. AI assistance is used in research and drafting. It is not permitted to be the source of a fact: every claim still has to trace to a document a person read, which is precisely what the ledger above is for. Generated text is not published unchecked, and the approval naming a real editor is the last gate before anything goes live.