How ranking works here, and what it refuses to do

What counts as a field

A field earns a column when a reader can repeat the check without asking anyone for permission. Three qualify at present.

The first is how many of the three rankings name a given brand, which is a count anyone can redo from the published pages. The second is the published return to player of 97%, which every ranking consulted prints at the same figure. The third is arithmetic performed on the second: subtract 97 from 100 and the house edge is 3%, a figure that comes from a sum rather than from a source, and is labelled that way wherever it appears.

Everything else is either a claim or a blank. Round length between 10 and 30 seconds, a smallest stake of 0.10 in dollars, the presence of two simultaneous bets and an automatic cash-out: these are claims, taken from public material, held at that level and phrased at that level. The maximum multiplier, the largest possible win and the upper stake limit are blanks, because no ranking read for the operator comparison prints any of the three.

What happens to a disagreement

A disagreement between sources is kept, printed, and left standing. Disagreement is material rather than a defect, and resolving it by picking the more convenient version would destroy the only thing that made the comparison worth reading.

The clearest example is our own. The return figure is recorded in our file as agreed by all three rankings, while what was actually captured from those pages is brand names and nothing else, so the agreement on 97% is inherited from the reading rather than reproducible from what was stored. It is printed anyway, because it is the only return figure anyone publishes, and it is printed as a figure nobody has confirmed at source. A second gap sits underneath every probability on this site. No distribution of the crash point is published, so the hit rates on the page working through cash-out targets come from the ordinary construction for a game of this shape rather than from a measurement. Neither gap was discovered by a reader writing in. Both were found by redoing our own sums before publication, which is the only reason they can be reported here rather than corrected quietly. A correction made in silence looks identical, from the outside, to a mistake that survived.

Both gaps are small, both are printed, and a site that asks readers to check other people’s sums has to survive its own.

The paid placement, named

Vave is the only paid placement on this site. It is also the only brand named by all three rankings, and those two sentences are printed together everywhere the finding appears, because apart they do damage: the agreement would read as verification of an operator whose position here is bought.

Payment buys a position and buys no points. Nothing in a table changes because of it, no column is weighted for it, and the count that puts Vave at the top of the home page finding was taken from pages with no commercial relationship to this one. The finding measures how much checking three rankings did. An operator standing in the middle of that overlap is a consequence of the measurement, and reading it as a verdict would invert what the whole exercise was for.

A version of this site would put all of that in a footer, where it performs worse for the reader and better for us, which is how footers usually work.

Numbers that came from a sum

Arithmetic on published figures is encouraged here and marked as arithmetic. The 3% house edge is one instance. The count of 21 duplicate entries across 79 recommendations is another, and so is the observation that 38 of the 58 brands carry a single mention apiece.

Sums are the strongest material available, because a reader can rerun one in a minute and catch us if it fails. Where a figure came from a sum, the sum is shown. Where it came from somebody else’s page, the page is described by what it is rather than dressed up as a measurement, and the working behind the return figure is set out at length.

What stays missing

Two things are absent from this site by design and will stay absent, and a third has changed. The game itself was opened and played in fun mode at Vave on 7 September 2026, on a play balance, which is why the round pictures on these pages are the game rather than a drawing of it, and why the arithmetic printed under them can be checked against the frame it came from. One lobby was opened and nineteen were not, so the operator table carries exactly one confirmed catalogue entry, and the nineteen blanks beside it record what nobody looked at rather than what anybody failed. The frames show the game window and nothing around it, so the venue is named on our word and not shown in the picture. No deposit has been made and no withdrawal has been requested, so a stated payout time is still described as stated rather than measured, and it will keep that word until somebody here has actually waited for one. No-deposit offers, which are excluded from the work entirely. And any instruction aimed at beating a game or an operator, since the pages about predictors and modified applications exist to explain why the mechanism forbids it.

What remains missing is shorter and worth naming. The catalogue column in the operator table is empty in every row: confirming that a venue runs the Spribe title under its own name means opening that venue’s game lobby and reading the supplier label, which is the next piece of work rather than a refusal. Bonus terms are the same case. The wagering multiple and the rate a crash round contributes are operator-specific, they were not part of the read that filled the other columns, and until they are the arithmetic on the bonus page runs on the operator’s own published pair of numbers rather than on a name.