How we test and score operators

Four kinds of evidence, kept separate

You'll see one of these labels next to facts across the site. We never blur them.

• Tested by Bettreview Something we did ourselves, recorded with a method, a date and the conditions.
• Verified official information A fact we checked against a current primary source, such as a regulator's register or the operator's own terms, with the date and scope.
• Operator claim Something the operator says that we haven't confirmed independently. A promised payout time is a claim until we test it.
• Reader report Something a reader told us privately. Unconfirmed unless further evidence supports it, and never averaged into a score.

What a real test looks like

A test answers one question. For example: how long did one bank-transfer withdrawal take, from submitting the request to the money arriving, on a verified account?

Before the test

We define the question, the start and end points, and what would count as a pass or fail. The tester must be eligible to use the platform and follow its terms. If money is involved, we set a small approved budget. We don't gamble just to create content.

During the test

We log each step with timestamps in West Africa Time, plus the device, browser or app, the account's verification state and the payment route. Receipts and screenshots are saved privately. We note anything that could affect the result, such as a weekend, a bank delay, a manual review or a bonus on the account.

After the test

A second person checks the evidence and how we've described it. We redact personal information, then publish the observation, the conditions, the date and the limitations. We keep the operator's advertised time separate from the time we observed.

What we won't do

We stop a test if it would need prohibited conduct, too much financial exposure or unnecessary personal data. We don't bypass identity checks or account restrictions, and we don't pretend to have an emergency to test support.

The tests we publish

• Payments and withdrawals: request, acknowledgement, processing and receipt measured separately, with the method, verification status, fees, limits and sample size.
• Support: the same ordinary question asked of each operator, recording the channel, the first response and whether it actually answered the question.
• Mobile and account controls: navigation, form errors, loading on slower connections, accessibility, and where limits and self-exclusion sit. Finding a control isn't the same as it working, so we test both where we can.
• Bonus terms: the original terms, eligibility, wagering, maximum cash-out, exclusions, expiry and withdrawal effects, with the version and date we read.

One test is one observation. It doesn't promise that your withdrawal, or your support chat, will go the same way.

How scores are earned

A review shows an overall score only when every category below has enough evidence, the scoring method has been approved by the editor, and there's no unresolved eligibility problem or open safety concern. Until then the score is withheld, and the review says why.

Each approved subcriterion is scored from 0 to 5:

• 0: Documented failure or absence.
• 1: Serious shortcomings.
• 2: Substantial limitations.
• 3: Meets the defined baseline.
• 4: Performs well with minor limitations.
• 5: Strong documented performance against the published criteria.

Category points = average subcriterion score / 5 x the category weight. The overall score = total points / 10, rounded to one decimal place. So a score is out of 10.

These are editorial assessments. A high score doesn't guarantee future performance, and it doesn't mean gambling is safe.

What we never do with scores

• We don't treat an untested category as zero, and we don't quietly drop it and rescale the rest.
• We don't include reader reports or app-store ratings in our score.
• We don't change weights to favour a result. A change means a new method version, and affected reviews are re-evaluated before scores are compared.
• Commercial relationships never change a score, a verdict or the order of a list.

Method status

Current method: version 1. The weights above come from our editorial guideline. The detailed subcriteria are still being finalised and approved, so no review on this site shows a numerical score yet.

Dates you'll see, and what they mean

• Tested on: when our own test happened.
• Information checked on: when we last checked facts against sources.
• Method version: which scoring rules applied.

Editing a page isn't a retest, and we don't change a date to make something look fresh. As internal targets, we aim to recheck official promotional terms about once a month (31 days) and core facts about every three months (92 days), and to repeat tests after a material change. These are targets, not a promise that every operator is monitored continuously. When information is older than that, the page shows Needs recheck and any score is paused.

Licences

We check licences against the relevant authority, not just a logo in a site footer. In Nigeria, the Supreme Court ruled in November 2024 that lotteries and gaming are mainly for states to regulate, so a licence from one state doesn't automatically cover the whole country. We state the authority and the territorial scope we could confirm, and we say when we couldn't confirm it.

Found a mistake? Tell us through the corrections page.