How a review here gets built
Slot figures decay by transcription. A studio states something once, an aggregator paraphrases it, a content farm rounds it, and eighteen months later forty pages agree on a number that has never existed anywhere except in each other. The clearest symptom is a maximum win. Pragmatic Play does not publish one for Starlight Princess 1000, yet you can find a precise figure for it in a dozen places, all of them downstream of one another and none of them downstream of the studio.
The counter to that is not diligence, it is labelling. If a reader can see which sentences came from the manufacturer, which came from an hour at the reels, and which are simply our opinion, then the ones that deserve scepticism can be treated with it. That distinction runs through every review here and it is worth setting out explicitly.
Three kinds of statement, and they are not equal
What the studio says
Return percentage, volatility rating, grid layout, ways to win, stake range, buy-feature status and, where it exists, a stated ceiling. These come from the manufacturer's own catalogue page for the exact title, and the review carries the link and the date beneath the specification block so you can put the two side by side. Where another site contradicts the manufacturer about its own product, that is not a dispute worth reporting - the manufacturer is describing something it built.
One thing this category cannot promise you: which configuration your casino licensed. Several Pragmatic titles ship in more than one return build and the licensee chooses. The catalogue figure is a reference, not a guarantee about a particular lobby, and every review here says so rather than implying the number follows you around.
What one session showed
The logged run on each review is a sample of one, written while it happened rather than reconstructed into a tidier shape afterwards. It is evidence about rhythm and nothing else: how far apart the features sat, what the near misses looked like, where the round's value actually came from. It is not evidence about the return figure, it does not forecast yours, and nobody including us can repeat it to the same result. A review presenting a session as proof that a game is generous has misunderstood its own material.
What we think
The verdict, the pros and cons and the mark out of ten. These are judgments and they are printed as judgments. You are entitled to disagree with all three while accepting every figure above them, and the pros and cons are written so you can: if the drawbacks listed do not bother you, the score is close to irrelevant to your decision.
Naming the exact build, before anything else
This discipline matters more on the games covered here than it would on most sites. The Starlight Princess name now sits on at least four separate machines - the earlier release, the 1000 build, a pachinko-style version called Pachi, and a Super Scatter entry - each with its own catalogue symbol and its own behaviour. A page that treats them as one product hands you a specification for something you are not playing.
So a review here starts by pinning the release, and every page names the other builds only as other builds. The same rule applies elsewhere in the lobby: the Big Bass family has several members and only the original is reviewed on this site, and the buffalo game most Canadians picture is an Aristocrat cabinet rather than the Pragmatic title we cover. When you want to check what you have loaded, the game client prints its own name on the rules screen. That line is authoritative in a way the lobby thumbnail is not.
What we count while a multiplier game is running
A totals-only session log hides the entire mechanic on the machines this site exists to explain. Watching the balance tells you whether an hour went well; it tells you nothing about why. So the running notes record particular events instead:
- Multipliers that arrived on a dead drop. A 100x symbol on a grid with no match pays nothing at all. Recording those is the only honest way to describe what a multiplier ladder is worth in practice.
- Multipliers that arrived on a live chain. The ones that converted, and what they combined into once summed.
- Doors that nearly opened. On a four-scatter trigger, three scatters is frequent and worthless. How often it happened is part of the machine's character and never appears on a spec sheet.
- Which single event carried the run. On every very high volatility title tested here, the answer was one feature round, which is exactly what the volatility rating is trying to warn you about.
- The stake and the spin count. A session note missing either of those is an anecdote wearing a lab coat.
An empty field is a finding
Where a manufacturer publishes nothing, the cell stays blank and carries a note saying why. We do not average competitors, borrow from a sibling title, or reason backwards from a screenshot somebody posted. Two of the five games here have no stated ceiling, and that absence is genuinely informative: it tells you the number circulating online has no source, and it tells you something about how the studio wants the product understood.
If Pragmatic Play publishes a ceiling for Starlight Princess 1000 tomorrow, it appears here with the date it appeared. Until then the blank stays, and readers who write in to correct it get the explanation rather than a silent edit.
Where each figure on this site came from
Five games, five manufacturer pages, and a different amount of disclosure on each. The table lists what we were able to take and what the studio keeps to itself.
| Game | Manufacturer | Return stated | Ceiling stated | Left blank here |
|---|---|---|---|---|
| Starlight Princess 1000 | Pragmatic Play | 96.50% | No | Maximum win |
| Big Bass Bonanza | Pragmatic Play | 96.71% | Yes - 2,100x stake | Nothing |
| Sugar Rush 1000 | Pragmatic Play | 96.00% | Yes - 25,000x stake | Nothing |
| Buffalo King Megaways | Pragmatic Play | 96.52% | No | Maximum win |
| Elvis Frog TRUEWAYS | BGaming | 96.7% | Yes - 5,000x stake | Nothing |
Each review links the manufacturer page it was read from, with the date of the reading, directly beneath its specification block.
How a demo gets on the page
The playable preview above each review is the manufacturer's own build, loaded from the manufacturer's own public endpoint, with a Canadian dollar counter where the studio supports one. Before a machine appears here we confirm three things: that the endpoint belongs to the studio, that it loads in a frame without leaning on somebody else's licence, and that the client reports its own name and return figure so a reader can verify the build against what we published.
What happens inside that frame is not ours to influence and should not be. Outcomes resolve on the studio's servers, the frame is cross-origin, and Ontario's own standards require a free-play version to run the same odds as the paid one. Deciding when a preview ends is a product choice. Altering what it returns would be a regulatory offence, and the answer to that is simply no.
The mark out of ten, and why it stays a judgment
A score here compares a machine with its own class rather than with every slot ever released. A very high volatility title is not marked down for having long silences - that is the category. It is marked down when the silence buys nothing: a ceiling nobody will see, a feature that under-delivers when it finally opens, or a mechanic that turns out to be an old engine in new artwork.
We could dress that up as a weighted formula and publish the weights. It would be theatre. The inputs are qualitative, the weighting would be chosen to produce scores we already believed, and a number with four decimal places behind it is not more honest than a number with one. So it is stated as an opinion, and the pros and cons underneath it are written to let you overrule it.
Money: who pays for this, and what it buys
Some outbound links to operators pay us commission when a reader registers. That funds the site and pretending otherwise would undo the point of this page.
Here is the boundary. Commission reaches the placement of a link and stops there. It does not reach the specification block, the volatility rating, the blank where a ceiling would go, the mark out of ten, or the paragraph on several pages explaining that an offshore operator answers to no Canadian regulator. Those are settled before anyone knows which offer a reader will click. An operator asking us to soften any of them would be told no, and would be named on this page for asking.
What we will not print
A method is defined as much by its refusals as by its procedure. These are written down so they can be held against us.
- A ceiling the manufacturer has not stated, regardless of how many competitors quote one.
- A spin count or a credit ceiling from a promotion we have not verified against live terms - an invented figure is a false claim, and in Ontario an advertising matter as well as a credibility one.
- A demo borrowed from another affiliate's tokenised embed. Previews here come from the studio's own endpoint or the page says there is none.
- A description of a build we have not played. Pachi and Super Scatter are named here only to keep them separate from the machine we did test.
- Any suggestion that a betting pattern moves an outcome. Results resolve on a certified generator and no staking system touches it.
- The grey-market paragraph relegated to small print at the bottom of a page.
How to audit this site in five minutes
None of the above is worth much if you have to take it on trust, so here is the short version of checking us:
- Open any review, follow the manufacturer link under the specification block, and compare the rows line by line.
- Look for the blanks, then confirm the same absence on the manufacturer's page.
- Load the preview at the stake we logged and see whether the rhythm resembles what we wrote.
- Read the drawbacks before the score and decide whether the two agree.
- Check whether a return figure or a mark ever moves in a direction that would suit an advertiser. They do not.
Anything that fails those checks is a correction we want. Send it with the manufacturer link attached and it is usually live the same week - the process, and what we do when the manufacturer turns out to agree with us instead, is on the contact page. Background on who is writing this is on who we are, and the numbers themselves are unpicked on RTP explained and slot volatility.