A checklist for any casino or game review: where the numbers came from, whether anyone played, what one session proves and what it admits.
A review is only worth reading if you can tell where its claims came from. Most of the industry's reviews are built from the operator's own marketing material and a score out of ten, and the score is doing the work that evidence should be doing.
This page sets out what to look for. It names no competitors and it is not an argument that we do it better, it is a list you can apply to any review including this site's.
Every figure should say its source, and the source should be checkable.
Return to player, hit frequency and maximum win odds are published by game suppliers. A review quoting them should say so and should name the version it read, because suppliers ship games at more than one return setting and the operator chooses.
A figure with no stated source is a figure you cannot check, and there is no reason for it to be unsourced except that checking it was skipped.
Measured play and marketing copy read very differently, and most reviews are the second.
A review of a game should be able to say how many spins, at what stake, over what period, and what came back. Not as proof of what the game is, because a single session cannot establish that, but as a description of what an evening with it is like, which the published figures never give you.
If the review cannot tell you whether the reviewer ever opened the game, that is the answer.
The honest framing matters more than the data.
A session shows texture: how often something happens, how long the quiet runs are, what the interface asks you to do while you are losing. It cannot show the return to player, which is measured over millions of rounds, and it cannot establish that a game is hot or cold.
A review presenting one session as evidence of a game's character is overreaching. A review that measures a session and says plainly what it does not prove is doing the job.
Any claim about a licence should cite the regulator's own register or published material.
Affiliate sites, review aggregators and casino portals are not sources for licence facts. They copy each other, the errors propagate, and licence numbers in particular get attached to the wrong legal entity because the brand name was matched instead of the company.
The part most often missing entirely.
Bonus buys, extra spins and side bets all have prices, and sometimes published returns. A review should say what the feature costs as a multiple of stake and, where the supplier publishes it, what that mode returns compared with simply playing.
Where the supplier does not publish the return for a paid feature, a review should say that the figure is unavailable rather than quietly leaving the impression that it is fine.
Three, and admitting them is a signal of quality rather than weakness.
Figures it could not verify. Claims it could not source. Questions it could not answer, such as what happens at a specific operator rather than in the game itself.
A review with no gaps in it has either not looked hard or is not telling you.
| Ask | A good answer | A bad sign |
|---|---|---|
| Where is this number from | A named supplier, regulator or document, with a date | No source given |
| Did anyone play it | Stake, spins, period and result stated | Only marketing descriptions |
| What does one session prove | Explicitly limited to texture | Presented as proof of the game |
| Who says it is licensed | The regulator's own register | Another review site |
| What do the extras cost | Multiple of stake, and the return if published | Features described, prices not |
| What could it not confirm | Stated plainly | No gaps acknowledged |
Only if you can see what produced it. A score with published inputs is a summary. A score without them is an opinion wearing a number.
Because the alternative is implying knowledge it does not have. Unverified figures are the main way errors spread between sites, since each one treats the last as a source.
Yes, that is the point of publishing it. Apply the checklist here and anywhere else.
This page is a reading checklist and general guidance. It cites no regulator and names no review site. Last reviewed October 15, 2026.