Most casino reviews are a rewrite of another casino review. This page describes what we do instead, in enough detail that you can catch us failing to do it.
A score is one number and you have no way to audit it. A method you can audit: if we say a figure came from the Curaçao register, you can open the register. If we say the terms analysis flagged a confiscation clause, you can read the analysis. The point of publishing this page is to make the review falsifiable.
This method is documentary. It is built on registers, regulators, terms documents and public dispute records, and it does not include real-money play. That is a deliberate limit, not an oversight, and Stage 8 says exactly what it costs us.
Before anything else: who is the counterparty. The licence badge names a company; the company goes into a corporate register. For RollXO that returned Stable Tech N.V., company number 160129, registered in Curaçao - and immediately contradicted the operator name printed across most of the Australian search results.
An operator name that fails at this stage invalidates everything downstream. If a review cannot tell you who takes your money, its withdrawal timings are decoration.
The licence number goes into the regulator's register, not into a search engine. We record the issuing body, the issue date, the stated term, and what else the same number certifies. That last field is the one nobody checks: at RollXO the same licence number serves a sister brand in the same group, which explains far more about the operator than its own marketing does.
We also record the regime. Curaçao's LOK ordinance took effect on 24 December 2024 and converted existing licences to provisional ones automatically. A reviewer who does not know this reads an expired term as a scandal.
Wagering multiple, what it multiplies, max-bet rules during bonus play, game contribution, max cashout on bonus winnings, dormancy fees, and every clause permitting the operator to withhold or confiscate. Where the terms are behind a logged-in cashier and not indexed, we say so and lean on a published clause-by-clause analysis rather than pretending to have read the source.
Confiscation clauses get quoted in substance, never paraphrased into something softer. "Low-risk play may result in confiscated winnings" is not the same sentence as "the casino discourages low-risk play".
Methods in and out, minimums, fees, processing windows, and the withdrawal ceilings - daily, weekly, monthly - with the currency they are denominated in. Currency matters: a euro-denominated cap on an AUD-native account moves with the exchange rate, so the ceiling you screenshot in January is not the ceiling that applies in July.
Every figure is taken from at least two independently maintained trackers. Where they disagree, both numbers are printed. The RollXO withdrawal cap is the live example: one tracker records €2,500 weekly and €12,000 monthly, the other $3,000 weekly and $15,000 monthly, and the honest output is the disagreement rather than an average.
Public dispute platforms publish full case text. We read cases, not counts. A resolved dispute worth $46 that ran past a month tells you how the operator behaves under pressure; a rejected-complaint ratio tells you how often players give up. Cases involving Australian players are pulled out separately, because AU-specific friction is the whole point of an AU review.
Case categories are reported in full - resolved, rejected, open, unresolved - because a partial ledger is how a review flatters an operator without lying.
Player ratings get read, not counted. At RollXO, reading them surfaced reviews under this brand's listing that describe other casinos in the same group by name. A 10-out-of-10 average built from text like that is not evidence of quality; it is evidence of something else, and it belongs in the review.
We read what the market says about the operator, and we treat contradictions as findings rather than noise. Five mutually exclusive welcome packages for one brand is a fact about the market that a player needs before they trust any of them - including us.
A competitor's review is never a source for a specification. It is only ever evidence about the competitor.
It cannot tell you how long your withdrawal will actually take. It cannot confirm whether a payment method appears in the Australian cashier. It cannot tell you what the wagering line says on your own bonus panel. Those need an account, a deposit and a clock, and until that work is done this review does not claim them.
Where a question is only closable by a live test, the review says the question is open and names the artefact that would close it. It does not fill the gap with a number borrowed from somewhere plausible. A review that admits its limit is worth more than one that invents past it.
Every figure is re-checked against its live source on the day of publication, and the date is printed on the page. Arithmetic is recomputed rather than carried over: if a review says a bonus requires $750,000 of turnover, someone multiplied it that morning.
Anything that cannot be traced to a named source class is cut. Not hedged - cut.
Licence status, withdrawal caps, bonus structure and complaint counts are re-checked on a fixed cycle and after any regulatory change. When a figure moves, the affected section is rewritten and the fact-check date at the top of the review changes. When nothing moves, the date does not change either - a "last updated" stamp that advances without any content changing is a lie about freshness, and the editorial policy commits us against it.
