Editorial policy
How QuantMedia research is produced, funded, corrected and disclosed. The purpose of this page is to let a reader judge the work without having to take anything on trust.
Independence
QuantMedia is independently operated by Cemil Ertürk and is not sponsored by, affiliated with, or paid by any fund, broker, exchange, data vendor or issuer. No paid placements, sponsored posts, affiliate links or gifted subscriptions appear anywhere on this site. No third party reviews or approves research before publication.
If that changes, it will be disclosed on this page and on the affected page before publication, not afterwards.
How this site is funded
QuantMedia currently generates no revenue. There are no sponsorships, no paid placements, no affiliate links and no paid subscriptions, and nothing on this site has ever been written for payment.
An application to Google AdSense has been submitted and is pending. If it is approved, advertising may appear on this site in future, and this page will be updated to say so plainly. No other form of monetisation is planned at present.
Whatever funding arrives later, the rule is fixed in advance: advertisers get no input into research topics, conclusions or publication timing, they are never shown content ahead of readers, and no revenue arrangement will depend on reaching any particular conclusion. Any change to this will be disclosed here before it takes effect, not after.
No investment advice, and no positions disclosed as recommendations
Everything here is published for informational and educational purposes. It is not investment advice, not a recommendation, and not an offer or solicitation to buy or sell any security. Cemil Ertürk is not a licensed investment adviser or broker-dealer in any jurisdiction. The signals published by the daily scan are the mechanical output of a documented rule set, not a view on what any reader should do.
Data sources
| Data | Upstream source | Nature |
|---|---|---|
| US equity OHLCV, signals, indices | Yahoo Finance, via the yfinance library | End-of-day bars |
| Market news | Finnhub | Third-party headlines, attributed to their publisher |
| Foreign exchange | open.er-api.com | Reference rates |
| Crypto | CoinGecko | Reference prices |
QuantMedia does not own or license a professional market-data feed. All upstream data is free-tier and carries the accuracy limitations of that tier, including occasional gaps, late revisions and unadjusted corporate actions. Anything computed by QuantMedia from that data — the confluence scan, the Signal Breadth and Sector Confluence indices — is original to this site, and is labelled as such.
News headlines are third-party content and are always attributed to the publisher that produced them. QuantMedia does not rewrite, re-report or claim authorship of them.
Live versus snapshot
Nothing on this site is labelled live unless the pipeline actually refreshed it. The machine-readable state is published at /data/status.json and each source reports independently, so one failing feed cannot silently stale the others. When a source has not refreshed, the site says the feed is delayed or paused and shows the timestamp of the data it is displaying. A failed update never removes data that was already published — the last known good value is preserved and disclosed as such.
Methodology versioning
The production signal methodology carries an explicit version and effective date, published in /data/signal_config.json and stamped into every scan output. The current engine is version 2.0, effective 14 April 2026.
Research pages carry their own research version, publication date and last material revision date. The version is incremented only when the substance changes — a new limitation, a corrected calculation, a changed conclusion. Typography and layout fixes do not bump it, because a version number that moves for trivial reasons tells the reader nothing.
Corrections policy
Mistakes are corrected in place. The correction is recorded rather than silently absorbed, so that the history of a claim stays legible.
- Material errors — a wrong number, formula, or conclusion. The page is corrected, the last-material-revision date is updated, and the correction is described on the page itself.
- Non-material fixes — typography, broken links, clarified wording. Fixed without ceremony and without a version bump.
- Withdrawal — if a piece of research cannot be supported, it is marked as withdrawn with the reason, rather than deleted. Removing the evidence of a mistake is worse than the mistake.
Two corrections have been published so far, both found by running the code rather than reading it: an arithmetically wrong worked example in the Probabilistic Sharpe Ratio paper, and a VPIN degenerate case that returned zero toxicity for a perfectly one-sided tape. Both are disclosed on the affected pages.
To report an error, email contact@quantmedia.io.
Reproducibility standards
Where an implementation exists it is published with example data, expected output and tests, and the reproducibility index states honestly which papers have code and which do not. Papers without an implementation are labelled research-only rather than left to imply otherwise.
Example datasets shipped with the packages are synthetic and generated from fixed seeds. This is stated in the README, the module docstring and the console output. Synthetic data demonstrates that code runs and is deterministic; it does not validate a strategy, and no result computed on it should be read as evidence about live markets.
What has not been validated
The published signals have no forward-tested track record. The methodology is documented and the outputs are published daily, but no out-of-sample performance record has accumulated yet, and none is claimed. Backfilling one would be fabrication. Any performance figure appearing on this site is explicitly labelled as in-sample, illustrative, or computed on synthetic data.
Use of AI tools
AI coding and writing assistants are used materially in building this site: in drafting and editing prose, in writing and refactoring code, and in auditing pages for errors and inconsistencies. This is disclosed because it is true and because readers are entitled to know how content they are asked to trust was produced.
What that does not change: every factual claim, formula, reference and
numeric result is the author's responsibility. Formulas are checked against
their primary sources, numeric examples are computed with
numpy/scipy rather than asserted, and the code is
covered by tests that run in continuous integration. Where an AI-assisted draft
produced something wrong — as with the PSR worked example — it was
caught by executing the calculation, and the correction was published.
No page on this site is auto-generated and published without review, and no AI system is presented as an author.
Privacy and analytics
QuantMedia uses Google Analytics 4 for aggregate traffic measurement. Google AdSense code is present because an application is pending; should it be approved, advertising cookies would apply as described in the privacy policy. No accounts, logins or personal profiles are collected by this site.