Data hub: every figure on this site, with its source and date

Every number used anywhere on this site appears below with its value, unit, primary source, document identifier, publication date and freshness status. Nothing is averaged and nothing is estimated. Where two credible sources disagree, both appear as separate rows. Where we looked for a source and could not find one, the row says so. The table is the site's working method made visible: if a claim on any page cannot be traced from here, it should not be on that page.

Published 27 August 2026 · Updated 28 August 2026 · AI Trading Book Editorial

Regulation

Retail losses and consumer harm

Performance, accuracy and adoption

Market size estimates

Eight estimates spanning roughly 22 times. Published in full because the spread is definitional rather than statistical — the reasoning is on the market size page.

Platform pricing

All captured from vendor pages in August 2026. Every row is verify before use — tariffs, tier limits and promotional pricing change often, and a stale price is worse than no price.

Documented incidents

Where the data does not exist

Twelve gaps recur across the site. Each was searched for and each is published as a gap rather than filled.

How this table is maintained

A figure enters only with a named source, a retrievable identifier where the source type has one, and a date. Where two credible sources disagree, both get rows — retail loss rates and market-size estimates are the two places this happens most. Nothing is averaged, harmonised or rounded to a convenient value.

Rows marked verify before use are re-checked before being cited anywhere new; vendor pricing is the main category. Rows marked historical describe a fixed past period and will not change. When a source publishes a new version, the row updates and the change is logged on the corrections page rather than applied silently. The full standard is set out in the editorial policy.

If you find an error, or a primary source for one of the twelve gaps, the corrections page is the route. A document number or URL is more useful than a description.

Frequently asked questions

What is the data hub?

A single table containing every figure used anywhere on this site, with its value, unit, primary source, document identifier where one exists, publication date and a freshness status. It exists so that any claim on any page can be checked in one place.

What does the freshness status mean?

Current means the source is the latest published version we could locate. Historical means the figure describes a specific past period and will not be updated. Verify before use means the value changes frequently — vendor pricing in particular — and should be re-checked at the source before being relied on.

How are figures selected for inclusion?

A figure appears only if it has a named source, a retrievable identifier such as a report or media release number, and a date. Where two credible sources disagree, both are listed with their own rows rather than averaged or reconciled.

Why are some cells marked data not found?

Because we looked for a primary source and could not locate one. It means the figure could not be verified, not that it does not exist. Publishing the gap is more useful than filling it with an estimate borrowed from a neighbouring jurisdiction.

Can I reuse these figures?

Yes, and preferably by citing the original source listed in the row rather than this page. Each entry names the regulator publication, paper or vendor page it came from precisely so that it can be traced past us.

Which figures conflict with each other?

Retail loss rates range from 68% to 82% across different measures, and market-size estimates span roughly 22 times. In both cases the conflict is definitional — different populations, periods and boundaries — so both are published in full rather than resolved.

How current is the platform pricing?

Captured from vendor pages in August 2026 and marked verify before use. Platform tariffs, tier limits and promotional pricing change frequently, so any pricing figure should be re-checked at the vendor's own page before it is relied on.

How often is this page updated?

When a source publishes a new version, when a figure we marked verify before use is re-checked, or when a correction is made. Changes are logged on the corrections page rather than applied silently.

What is the largest gap in this data?

New Zealand. Of fourteen comparable regulatory data points tracked across three markets, we located New Zealand primary sources for three, against thirteen for the United Kingdom and twelve for Australia.

Do any of these figures measure whether AI trading is profitable?

Only indirectly. The Eurekahedge AI Hedge Fund Index comparison and the peer-reviewed review of 27 machine-learning experiments are the closest available evidence. No regulator publishes an assessment of AI trading tool performance.

How do I report an error?

Through the corrections page, ideally with the document number or URL of the source that contradicts what is published here. Corrections are logged with the date and what changed, rather than edited in silently.

Informational research only. Nothing on this page is personal financial, legal, tax or investment advice. Regulatory positions, vendor pricing and published statistics change; verify at the named source before relying on any figure.