Honest casino reviews: every verdict, with the clause that proves it.
RAW publishes 25 operator verdicts of the 32 casinos it tracks, and every one is built the same way: the withdrawal terms, fee policy, KYC gate and licence are read from the operator's own pages, the clause is cited, and the date it was read travels with the claim (newest read 2026-09-02, across 170 source documents). The overall score is capped at 9, because the inputs an honest 10 needs do not exist yet, and every score renders next to the axis table that produced it: a number nobody can audit is the thing this site exists to argue against. Where an operator's own record contradicts its marketing we publish both, and where we could not verify something we say so instead of filling the gap. 18 operators sit on a published refusals ledger with the reason attached.
A review here is not a number pulled from feel. It is the operator's own published terms, read on a date we name, with the parts that contradict the marketing left in, and the score printed beside them shows the formula and axis table that produced it.
Every answer in this topic
Common questions
How is an honest casino review different from a rated one?
A rated review compresses an operator into one number whose weights you cannot check. RAW publishes the number with its weights and every fact under it: the stated payout window, the fee class, the KYC gate in the operator's own words, any published no-KYC ceiling and the licence, each with a clause reference and the date it was read, plus RAW Score v0 rendered next to the axis table that produced it. 25 operators have a verdict page today.
What decides where an operator ranks?
Typed dataset fields, and nothing else. The ranking is computed from the records described above: stated payout window, fee class, KYC gate, published ceilings, licence, and any documented delay record. If a value is not in the dataset it cannot move a position or a score, which is also why RAW Score v0 is capped at 9: the inputs an honest 10 needs, a complaint ledger and self-measured payout times, do not exist yet.
Can a bad review be reconsidered?
Yes. Any operator may request a re-evaluation every 3 months, and the criteria are public so the operator can fix what the review found without asking us. A re-evaluation is a fresh read of the live documents against the same public criteria, not a negotiation.
What does RAW refuse to publish?
Anything it could not verify. 18 operators are on a published refusals ledger, each with the reason, and the site states plainly when an operator publishes no figure at all rather than inventing one. Publishing nothing is not the same as an operator having no limit, and the two are never merged here.
How this topic is verified
Every answer collected here is built from operator documents or checkable public records, carries the date it was read, and is asserted against the dataset when this site builds: if the words and the data disagree, the build fails rather than the page going out wrong. The full criteria are on the method page, and the operators we refused to rate, with the reason, are on the refusals ledger.
Operators: reviews here are not permanent. You can request a re-evaluation every 3 months, publish a right of reply on any page about you, and see exactly what the criteria are before you ask. Details on the operators page.