How we rate casinos
Every score on Casino.net is four separate inputs at once:
arithmetic from our own operator records, a named reviewer’s verdict, votes from readers, and ratings collected from public platforms. This page says exactly how much each one counts, and what happens when one of them is missing.
4
inputs
26
algo criteria
282
casinos rated
19
markets
10/10
the score
Every casino review on Casino.net opens with a single score out of 10. It is not one measurement and it is not one person’s opinion. It is a weighted blend of four inputs, and the point of this page is that you can see which is which.
One input is deliberately mechanical. The Casino.net Algo reads 26 criteria out of our own operator database, applies fixed weights, and ranks each casino against the others we rate in the same market. Anyone can check it against the table printed on the review.
The other three are deliberately human. A named reviewer plays the casino, deposits and cashes out whenever we can, checks the support channels, and says what the data cannot. Where we cannot open an account, we rely on players we stay in touch with and player feedback online. Readers vote. Public platforms are counted too, because our verdict is not the only one worth reading.
Four inputs, one score
These are the live weights. Reviews with reader votes use the 30/30/30/10 split. Reviews with no votes use a separately defined 40/40/20 fallback, so an empty vote pool is never counted as zero.
Written as arithmetic: with votes, (algo × 3 + expert × 3 + votes × 3 + platforms × 1) ÷ 10; without votes, (algo × 4 + expert × 4 + platforms × 2) ÷ 10. Every input is converted to tenths before it is added, so the published score is exact rather than a rounded float.
The Casino.net Algo
The Algo is the part of the score with no opinion in it. Twenty six criteria across seven categories, each with a fixed ceiling, each read from a record in our operator database rather than from a view about the brand.
Nothing is stored
The table is recomputed every time a review loads, so a corrected record shows up in the score immediately.
Local pools only
A casino is ranked against the casinos we rate in the same market, so its score is a position among local competitors rather than a global verdict.
Independent ceilings
Safety and responsible gambling make up a quarter of the Algo, our largest category, so a failure there cuts the score sharply.
Inside the Casino.net Algo
All 26 criteria with how each one is measured, the points split category by category, where scores actually land across the index, what changes from market to market, and every version of the algo on the record.
The expert assessment
A named reviewer signs off every review. The assessment is what the arithmetic cannot reach: whether registration works, whether support answers, whether a withdrawal arrives when the terms say it will, and whether the site is pleasant to use. These are checked on our own account whenever we can and against player reports when we cannot. A person can only open an account at a casino once, so later re-checks lean on the published terms and recent player feedback.
What is examined
Registration and verification, the deposit and withdrawal path, the terms behind the bonus, support responsiveness, game discovery, and site performance on mobile and desktop.
Who signs it
A named author and, where the subject calls for it, a named specialist. Both appear on the review with their credentials and their other work.
Kept separate on purpose
The reviewer cannot change a criterion, a weight or an Algo score. Judgement enters the total only through this one input, at the weight shown above.
Reader votes
Any visitor can rate a casino on its review. The number is exactly what readers left, with no editorial adjustment applied to it.
One rating per reader
A repeat rating from the same reader replaces the earlier one instead of stacking, so a single enthusiast cannot lift a casino.
Absent, not zero
A review nobody has rated yet carries no vote weight at all. An empty vote pool is never counted as a bad score.
What it is worth
Votes are unverified by design, which is why they sit alongside the other inputs rather than above them.
Third party ratings
The smallest share of the score comes from outside Casino.net entirely: public ratings of the operator, and two independent measures of the site itself.
| Source | What it rates | Scale |
|---|---|---|
| App Store | The casino’s iOS app, where one exists | out of 5 |
| Google Play | The Android app, read from the same storefront listing | out of 5 |
| Trustpilot | The public score for the operator | out of 5 |
| Reviews.io | A second independent review platform | out of 5 |
| Lighthouse | Google’s own performance audit of the casino site | out of 100 |
| Ahrefs | Domain strength (Ahrefs DR): how strongly other websites link to the operator. A strong backlink profile takes years to build, so it signals an established brand. | out of 100 |
Where a market has no web rating for an operator, the missing rating is taken from another market in a fixed order: the page’s own market first, then the UK, Germany, Canada, Spain, Italy, Ireland, Malta, Sweden and the other markets in our list, and finally any market that holds a value. App Store and Google Play ratings are never borrowed from another market. We would rather show a real rating from elsewhere than an empty column. The web score is only shown when at least two channels return a valid rating, and an Ahrefs Domain Rating below 10 counts as not listed. These figures are read from our own database rather than fetched as you load the page, so they reflect the last time we collected them.
What we never do
Four commitments that hold across every market and every review, and each one is checkable against the page in front of you.
No paid placement
A commercial relationship cannot buy a score, a criterion, a weight or a place in the main ranking. The only exception is Our Monthly Picks, a toggle you switch on yourself, which can reflect commercial relationships. Our arrangements are set out on the commercial disclosure page.
No score without data
Where we hold no record for a criterion it scores nothing and the review says so. An empty field is never treated as a pass.
No advice to gamble
A high score means a casino performs well against the things we measure. It is never a recommendation that gambling is a good idea.
No silent edits
Changes to the criteria or the weighting are published in the algo changelog, including the ones that pushed published scores down.
Who writes the reviews
Every review carries a named author, and specialist subjects carry a named expert alongside them. Their credentials, their market and their other reviews are listed on their own profile.
The author
Writes the review for one market in that market’s language, and signs the expert assessment that feeds the score.
The specialists
Slots, live and table games, legal, and the technical side each have a named expert whose opinion appears on the review under their own byline.
The method owner
Katya Vasileva owns the scoring method itself and signs off every version of the Algo in its changelog.
A figure about your casino wrong? Write to editorial@casino.net with the record and the market. A corrected record recomputes the score on the next page load, and a correction that moves a published score appears in the algo changelog.
When a score changes, and why
A score is not a verdict frozen on the day of publication. Four separate things move it, and each one leaves a trace.
01
A record changes
A licence, a bonus, a withdrawal time or a game count is corrected in our operator database. The Algo is recomputed the next time the review loads, so the score follows immediately.
02
A criterion changes
We add, remove or rewire a criterion. Every affected score moves at once, and the change is written into the algo changelog with the date and the direction.
03
A reviewer revisits
At each scheduled review update the expert assessment is redone against the live site, which can move the score in either direction.
04
Votes arrive
Votes accumulate continuously. A brand new review carries no vote weight at all until the first ratings arrive.