Tested at the table. Not in a press release.

A platform review is only as honest as the test behind it. The desk runs each platform through a fixed script before any verdict is published: a real-money deposit through two different rails, ten hands on the format the platform advertises most, a withdrawal at the smallest and largest size, and a support ticket at three different hours of the day. We do not pay for the test; we do not accept comped play; we do not let the platform know we are reviewing until the verdict is filed.

The five axes we measure, in order of weight, are: dealing integrity (a third-party audited RNG and a dealing latency under 1.5 seconds), withdrawal timeliness (the moment of first withdrawal request to the moment of bank credit), control visibility (time, deposit and loss limits surfaced in the user menu, not in support tickets), support competence (the support ticket that is read and answered rather than routed), and bonus honesty (the bonus that can be read in plain language and honoured without surprise).

The desk does not score platforms on a ten-point scale. Scoring invites a single number to do the work of a paragraph, and readers deserve the paragraph. Instead, each platform gets a verdict — yes, with caveats; yes, with serious caveats; or no, until the operator fixes X. The list of caveats lives in the body of the review, not in a single-line star count, so readers can judge for themselves whether the caveats matter to their use of the platform.

The desk re-tests every reviewed platform once a quarter. Platforms that drop a verdict (a withdrawal delay that grows past twelve hours, a control that gets buried, a support ticket that goes unanswered) are pulled from the active list until the issue is fixed. Platforms that improve (a faster dealing engine, a more honest bonus, a new state-level licence) are upgraded with a note explaining what changed. The reviews page is meant to be read as a working ledger, not as a one-time guide.

For readers who want to verify the desk's findings, every review links to the public record we used: the licence filing, the audit certificate, the support response we received, and the deposit/withdrawal timestamps. Where we have been unable to verify a claim, the review says so in plain language. A review without evidence is not a review; it is marketing.

The desk's review cycle runs on a quarterly calendar with three named checkpoints per year and a fourth checkpoint reserved for a major event (a regulatory change, a platform acquisition, a dealing-engine incident). The checkpoints are public on the about page; the verdict is published within two weeks of each checkpoint; the test screenshots and timestamps are filed in the desk's editorial record and are linked from the bottom of each review.

The first checkpoint of the year focuses on withdrawal timeliness. The desk tests each platform with a small and a large withdrawal and records the time between request and bank credit. The second checkpoint focuses on control visibility — the time, deposit and loss limits that should be available in the user menu. The third checkpoint focuses on support competence, with three test tickets at three different hours of the day. The fourth checkpoint is reserved for major events and is scheduled at the desk's discretion.

A platform that drops a verdict at a quarterly checkpoint is pulled from the active list until the issue is fixed. A platform that improves at a checkpoint is upgraded with a note explaining what changed. The reviews page is meant to be read as a working ledger, not as a one-time guide; the date of the most recent re-test is stamped at the top of each review.

The desk's editorial line on review coverage is unchanged across issues: a review without evidence is not a review. Every verdict links to the deposit, withdrawal and support ticket timestamps; every caveat names a specific finding; every update names a specific change. Readers who want to verify the desk's findings can do so by looking up the timestamps themselves. The reviews are the desk's working record; the record is the part the desk cares about most.

The desk's quarterly re-test runs a fixed script on each platform in the active reviews stack: a real-money deposit through two different rails, ten hands on the format the platform advertises most, a withdrawal at the smallest and largest size, and a support ticket at three different hours of the day. The deposit, hands, withdrawal and support timestamps are filed in the desk's editorial record and are linked from the bottom of each review. A reader who wants to verify the desk's findings can look up the timestamps themselves.

The first checkpoint of the year focuses on withdrawal timeliness; the second on control visibility; the third on support competence; the fourth is reserved for major events. A platform that drops a verdict is pulled from the active list until the issue is fixed; a platform that improves is upgraded with a note explaining what changed. The reviews page is meant to be read as a working ledger, not as a one-time guide; the date of the most recent re-test is stamped at the top of each review.

The desk's editorial line on the test cycle is unchanged: a review without evidence is not a review. The test script is published on the about page so readers can run the same script themselves on a platform the desk has not yet tested. The desk welcomes reader-run tests that produce a different finding; the desk publishes the finding and re-tests the platform to verify.

The desk's editorial line on the cadence of the quarterly re-test is also unchanged: a quiet quarter is a working quarter, not a missing one. A platform that passes all four checkpoints without a change to its verdict is a platform the reader can keep; the desk's quarterly note will simply say "no change this quarter". A quiet verdict is more honest than a busy one; a busy verdict that runs on filler is a verdict the reader learns to skip, and a verdict the reader skips has failed at its job.