The casino register method
How is the casino register built?
24 operators are on the casino register, and 18 have a licence confirmed in the issuing authority's own register. 24 carry a published payout score, from a formula whose every weight is published on the methodology page. 112 contradictions between marketing and binding terms are documented, 38 at high severity. None has a measured withdrawal test yet.
Documents last checked Reviewed by CryptoGuardz Register Desk, editorial desk
This is the one register on the site that publishes a number. The formula that produces it, every weight in it and the authority weighting that drives most of it are published at how we rate operators, which is generated at build time from the constants in the scoring function itself.
What goes on the register
An operator is included where it accepts cryptocurrency deposits from retail players, runs a casino product rather than only a sportsbook, and publishes terms of service in English that could be retrieved and read in full. Inclusion is not an endorsement and exclusion is not a warning: an operator absent from the register has usually not been read yet.
There are 24 entries. The register is not a top list and is not ordered by the score. It is ordered by the register table’s own column headings, which a reader can sort by whichever fact they came for.
What is measured, and how
A licence counts as confirmed only where the number was found in the issuing authority’s own public register, on a date recorded on the entry. 18 of 24 are confirmed that way. An operator’s footer is not evidence of its own licence, and a register we could not query produces asserted and unconfirmed, which is a different finding from invalid and is never rendered as one.
A contradiction is recorded only where both sides can be cited: the marketing claim with its URL and date, and the binding term or register row that contradicts it with its own. 112 are documented across the register, 38 of them at high severity. Severity is about the consequence to a player who relied on the claim, not about how badly the operator behaved.
The score is a weighted sum of five components: licence verification at 30, measured payout speed at 25, on-chain wallet health at 20, claim consistency at 15 and an independent safety rating at 10. A component with no data is excluded from the denominator rather than counted as zero, and where fewer than 30 points of weight have data no score is published at all. Every one of those figures is read out of the scoring function, here and on the formula page.
What is not measured
Two components currently have no data anywhere on the register: measured payout speed, worth 25 of the 100 points of weight and on-chain wallet health, worth 20. No withdrawal test has been run on any of the 24 operators, and no payout wallet has been identified on any of them. Both are stated on every page that carries a score, and both are the reason every entry on the register is low confidence. That is a statement about our coverage and not about the operators.
The score does not measure whether an operator will pay you. It measures how much of what an operator says about itself can be confirmed from somewhere other than the operator. Those are different things, and only the second one is checkable from a desk.
Nothing on this register measures game fairness, and nothing here is derived from player sentiment, review-site aggregate ratings or complaint counts on their own. Fairness is a separate register with its own method at the provably fair method.
Figures we borrow, and from whom
Two numbers on an operator entry were measured by somebody else. Both are named, linked and dated where they appear, and neither touches the score.
The first is Casino Guru’s safety index, which is the one input to the score that is not our own work. It is worth 10 of the 100 points of weight and it is included because it publishes the criteria behind its own number, which is the test it had to pass to be used at all.
The second is Tanzanite’s casino deposit volume, and it is worth nothing at all in the score. 9 of the 24 operators on this register are covered by it: BC.GAME, BetFury, Cloudbet, Flush, Gamdom, Roobet, Shuffle, Sportsbet.io and Stake. The remaining 15 are not tracked by Tanzanite and carry no figure, which their entries say rather than leaving the space blank.
Deposits into casino deposit addresses Tanzanite has identified, across 10 chains (ARB, AVAX, BASE, BSC, ETH, POLY, SOL, TON, TRX, XRP) and 16 tokens (ARB, AVAX, BNB, DAI, ETH, LINK, MATIC, PEPE, SHIB, SOL, TON, TRUMP, TRX, USDC, USDT, XRP). It does not count fiat deposits, deposits in coins or on chains outside that list, deposits to addresses Tanzanite has not identified, internal transfers, bonus credit, and every bet placed with money already on the site. It is money arriving, not money wagered, and it is not a measure of how much an operator holds or pays out.
We publish it because we cannot produce it. No payout wallet has been identified for any operator here, so the one question a reader most wants answered on chain, whether money leaves as fast as it arrives, is one this register cannot answer. Deposit volume answers a smaller question honestly: how much money is actually moving, and in which direction, at a site whose own marketing will only ever say “millions of players”. Alongside the deposit count it also gives the average deposit, which separates a site taking a few very large deposits from one taking a great many small ones. That figure is ours, computed from their two.
The terms of the borrowing matter, so they are stated. The figures were read once, on 2026-09-22, from Tanzanite’s published deposit volume page, for the 9 sites of theirs that overlap this register. They are not mirrored on a schedule and the dataset behind them is not republished here. Ten three-day buckets. The final bucket was still open on the day we retrieved this, so the total is a running one, not a closed month. Tanzanite's own terms state that its on-chain and deposit analytics "may be delayed, incomplete, mislabeled, or affected by data-source limitations" (their terms), and that caveat travels with the number.
How an absent value is written
| Token | What it means | On this register |
|---|---|---|
| Not measured | The measurement is in scope and we have not made it yet. The gap is ours. | Used on every payout and wallet field. 0 of 24 operators have a withdrawal test and 0 have an identified wallet. |
| Not applicable | The concept does not fit this kind of entity at all. Reserved for a type mismatch, never for a count of zero. | Used where a field cannot apply to the entity, such as a bonus term on an operator that offers no bonus. |
| Not disclosed | The company publishes nothing on the point. The gap is theirs and it is a finding about them. | Used where an operator publishes no figure, such as a processing time its terms decline to commit to. |
| No data | We looked and either nothing exists or nothing would load. The gap is in the record. | Used where an authority publishes no queryable register, so the licence could not be checked either way. |
Corrections
If an entry about an operator on the casino register is wrong, send the primary source that corrects it to the address on the contact page. A correction needs the document, not a request: a link to the clause, the register row or the page that shows the recorded value is not what it was on the date we read it.
Corrections are published with their date at corrections, and the original value stays visible rather than being edited away. A change to what an entry records about a company is logged at the change log on the day it is made. A register whose history can be rewritten quietly is not a register.
Commercial disclosure
CryptoGuardz holds commercial affiliate relationships with 3 of the 24 operators on the casino register, named in full at the disclosure page and named again on every page that carries one of them. Being paid through them is the intended model rather than a description of the current build: no outbound link here carries affiliate tracking yet. The scoring function cannot see a commercial relationship, because a partner flag is not one of its inputs, and a data file carrying a hand-set score fails the build. No commercial relationship is an input to anything computed on this site, and the build rejects a data file that tries to set a derived value by hand.
The data
The register is published as a CSV with its columns and its known gaps at the data page. The score is not a column in it. It is derived from the columns by the published formula, so a reader who disagrees with the weighting can recompute the whole table from the inputs.