Ethics and publication policy

1. What this publication does

Fully Priced publishes the backtest results of strategies I have built. For now these are strategies designed for futures prop-firm challenges. Strategies for trading your own capital will follow. Results may be published free or to paying subscribers, under the standard in §10.

The site also keeps one earlier test of a popular trading method (ICT on MNQ). If I ever publish an audit of someone else's trading code, sections 2 to 4 apply to it.

What I write about is a strategy, a rule or a result, never a person. That includes the author of any code and anyone who teaches or trades a method discussed here.

The public interest served is specific and narrow: published performance claims about trading code influence what other people build, fund and deploy. Where such a claim is unsound, the people best placed to be harmed by it are the people least equipped to test it. Testing is cheap. The absence of anyone doing it is the problem.

2. What is audited, and what is not

Sections 2 to 4 govern code audits, where a specific person's published work is the subject. A market-claim test examines a method as it is commonly taught, not any individual's implementation of it, so there is no single author to notify. It is governed by the pre-registration rules in the methodologyand by sections 5 to 12 here.

Eligible. A repository is eligible only if it makes a specific performance claim in a durable, quotable artefact — a README, a paper, release notes, documentation. The claim must be quotable verbatim and locatable at a commit.

Never audited.

• Anyone's first project. If the repository is a learning exercise and the author has no other substantive public work, it is out of scope permanently, whatever it claims.
• Work by anyone who appears to be a minor.
• Private, unpublished or leaked code. Only what the author chose to make public, under a licence permitting analysis.
• Repositories that make no performance claim. Publishing code is not a claim. Sharing an experiment is not a claim. Absent an explicit claim there is nothing to audit and no public interest in auditing it.
• Anything where I hold a conflict. See §7.
• Any repository whose author is in evident personal distress, where an audit would compound it. A judgement call, erring toward not publishing.

Pseudonymous authors are audited on the same terms as named ones. I never attempt to identify a pseudonymous author, and never publish identifying information about any author beyond what they have published themselves.

3. Notification and right of reply

Before publication, without exception:

1. The author is contacted at their published address with the full draft audit, the commit tested, the environment specification, the run logs, and every figure relied on.
2. Where a fix is feasible, the remediation pull request is openedat the same time, so the author has the correction in hand while the response window runs.
3. The author has 14 days to respond. The deadline is stated in the notification.
4. On request the deadline is extended once by a further 14 days, and further where the author is actively working on a fix.

If the author responds, the response is published in full and unedited alongside the audit, at equal prominence. Not summarised, not excerpted. If it is long it gets its own page, linked from the top of the audit rather than the bottom.

If the author corrects the claim before publication, the audit is published with the verdict withdrawn, describing what was claimed, what was found, and what changed. A corrected claim is a success of the process and is reported as one.

If the author disputes a finding on technical grounds, the dispute is tested before publication. If they are right, the finding is removed and the correction noted. Being shown wrong before publishing is the cheapest possible outcome, and it is actively invited.

If no response is received, the audit is published stating the notification date, the deadline, and that nothing was received by it. A response is published whenever it arrives, however late, appended with its date.

4. Remediation

Every audit that identifies a fixable defect ships a pull request:

• Minimal and single-purpose. One defect, one PR. No refactoring, no style changes, no unsolicited improvements.
• Referenced to the specific checklist item and finding.
• Opened at notification, not at publication.
• Merged fixes are recorded on the audit with a dated note. They are the best outcome this publication can produce.

Where a defect is not fixable — a claim depending on unobtainable data, an approach unsound at its foundation — the audit says so and says why, and no pull request is opened.

5. Corrections

I will be wrong. The policy for when I am:

• Corrections are published on the audit itself, dated, describing what was wrong and what changed. Never a silent edit.
• Every audit carries a permanent correction log. Nothing is retroactively altered without a record.
• If a correction changes the verdict, that is stated in the correction, in the audit summary and in the audit list, and it is distributed on the same channels as the original with the same effort.
• Corrections are solicited. Every audit carries a contact address and an invitation to dispute any finding, with a commitment to test the dispute.
• A dispute made after publication is treated exactly as one made before it.

6. Distribution

Audits are published here first. Onward distribution — the repository's issue or pull request thread, the mailing list, and elsewhere — happens no earlier than publication, and never before the response window closes.

A repository's own issue and pull request threads are treated as a first-class publication surface. What is written there is written for readers, held to the same standards as the audit, and subject to this policy in full.

7. Conflicts of interest and sponsorship

Never audited: any current or prospective sponsor; any direct competitor of a current sponsor; any repository I have contributed to; any repository whose author has a commercial relationship with me, past or present.

Sponsorship is restricted to outcome-neutral infrastructure whose commercial interests are unaffected by any audit verdict: market data vendors, backtesting and research platforms, and compute providers.

Categorically excluded, without exception:

• Course sellers, educators and content businesses in this field
• Signal services, copy-trading services and managed accounts
• Proprietary trading firms and funding programmes
• Brokers and execution venues, including execution API providers.Accepting payment from an execution venue while publishing about trading claims is the precise arrangement that regulatory guidance on financial commentary treats as evidence that commentary has become financial product advice. It is excluded on that basis, and would be excluded on independence grounds regardless.

Terms of every sponsorship, repeated on every page one funds:

• No sponsor sees an audit before publication.
• No sponsor has any input into audit selection, methodology, findings or verdicts. No exceptions, and no informal ones.
• Sponsorship is disclosed on every page it funds, above the content.
• Every sponsor, current and former, is listed on a permanent public page with dates.
• No affiliate links to financial products, ever.
• The right to end a sponsorship at any time, for any reason, is retained and is stated in the agreement.

Personal positions. Where I hold a position in an instrument material to an audit, it is disclosed in that audit. Where the position is material enough to matter, the audit is not published.

Own trading. I trade futures on my own account, including through proprietary trading firm evaluation accounts. I am a paying customer of those firms. I am not an affiliate or partner of any of them, and I receive nothing from any of them in connection with this publication. Where an audited claim concerns an instrument or a strategy type I trade, the audit says so. My own trading returns are not published or marketed here, and strategies I currently trade live are not published here. A strategy I have stopped trading may be published after it is retired, labelled as such.

In-kind support. Free or discounted credits, software or compute from any vendor, model providers included, is treated as sponsorship under this section. It is disclosed and listed in the same way. Where an audited repository depends on a supporting vendor's product, or on a competitor's product, the audit states that.

There are currently no sponsors.

8. How findings are written

Alongside that:

• Suspicion, inference and demonstrated fact are labelled distinctly, in the text, every time. Where a defect is inferred rather than demonstrated, the audit says which.
• Findings attach to the artefact, not the author. An audit concerns a commit, not a career. Findings are never aggregated across a person's work into a characterisation of that person, and no lists of authors are kept.
• No audit of an author's other work follows from an adverse finding.One claim, one audit.
• Every audit states what it did not test, and what would overturn it.
• The verdict “unverifiable” is not a negative finding and is never presented as one.

9. Records

For every audit, retained permanently and produced on request in any dispute:

• The commit SHA, the quoted claim, and a capture of the source artefact
• The full environment specification, run logs and computed figures
• The notification, with timestamp, and all correspondence with the author
• The response window dates and any extensions granted
• Every version of the published audit
• The verification steps taken for each finding, and their outputs

10. Regulatory position

Fully Priced publishes the historical backtest results of my own trading strategies. It doesn't give financial product advice, general or personal. It doesn't recommend any financial product, and it isn't meant to influence anyone's decision about one.

Permanently:

• No signals: no live entries, exits or alerts, and no trade calls of any kind, free or paid
• No affiliate links to financial products or to any provider of them
• No claim, express or implied, that any strategy should or shouldn't be traded

Strategy results. Fully Priced may publish the historical results of trading strategies, free or to paying subscribers. Every result is published to one standard regardless of price:

• each result states its development status: whether the reported period was used to design, tune or select the strategy, was left untouched, is a mix, or is not documented
• costs are included and stated, and in-sample and out-of-sample results are shown separately where both exist
• drawdowns, losing periods and limitations are shown with the same prominence as gains
• strategies that lose are published to the same standard as those that win
• paid content is labelled as paid; payment buys access to research, never to a recommendation
• results are historical: they describe how a strategy behaved in the past, not how it will behave, and not whether anything is a good investment
• strategy code, full rules and parameters are not published

When a result discusses an instrument or a market, it describes what happened in the past. Nothing published here is a recommendation to any person, and nothing published here takes any person's circumstances into account.

11. Disputes

Any person may dispute any finding at any time, before or after publication, at [email protected]. Disputes on technical grounds are tested, and the outcome is published whichever way it falls.

Where a finding cannot be sustained it is withdrawn, and the withdrawal is published at the same prominence as the finding.

12. Amendment

This policy is versioned. Amendments are published with a dated changelog. An amendment never applies retrospectively to relax a commitment already made to an author under an earlier version.

Changelog

• 1.3 — 23 October 2026. §1: the publication now covers the backtest results of the author's own strategies, starting with prop-firm strategies; testing popular claims is no longer part of its scope, and the ICT test is kept as earlier work. §10: strategy research is now published as results only. Strategy code, full rules and parameters are not published. Version 1.2 required rules and the test specification to be fixed before testing; that is replaced by a requirement to state each result's development status, including whether the reported period was used to design, tune or select the strategy. Added: results are stated as historical, and drawdowns, losing periods and limitations are shown with the same prominence as gains. No signals and no affiliate links remain permanent. Commitments to authors of audited code are unchanged.
• 1.2 — 19 October 2026. §10: versions 1.0 and 1.1 said, "specifically and permanently", that there would be no strategy sales and no paid access to any purported edge. That commitment is withdrawn. It is replaced by the strategy-research standard in §10, which applies equally to free and paid content. No signals and no affiliate links remain permanent. §1: strategy research added to the scope. §7: strategies currently traded live by the author are not published. Commitments to authors of audited code are unchanged.
• 1.1 — 19 October 2026. §7: the reference to a named national regulator's guidance was replaced with a general reference to regulatory guidance on financial commentary. No commitment was added, removed or relaxed.
• 1.0 — 19 October 2026. First published version.