On-chain tiers T0 to T4: the ladder we use
Five rungs from unmeasured to fully on-chain, what evidence moves an operator between them, and why the tier caps what transparency can score.
Most comparisons of crypto gambling sites use a binary: on-chain or not. That label collapses at first contact with the products, because a site can settle deposits on a public chain, publish seed pairs from a private database, run a bankroll contract, and serve third-party slots from a data centre, all at once. We use a five-rung ladder instead. This page explains each rung, the evidence that moves an operator between them, and what the tier does and does not affect in the score.
Why a ladder and not a label
The useful question is not “is this on-chain” but “how much of this can I check without asking anyone”. That is a quantity, and it varies along a consistent axis: the further up the ladder, the more of the product is reconstructible from public data.
The tier is also a description, not a grade. It does not feed the total directly. It constrains what one criterion can be worth, and it tells a reader what kind of evidence exists for everything else.
T0 — Unmeasured
We have not placed the operator on the ladder. There is no reading yet of how much of the stack, if any, touches a public chain.
T0 is a statement about us, not about the operator. It appears for a site we have added to the index but not yet examined closely enough to defend a rung, or one whose product was not reachable from where we were looking. It is not a criticism and it is not a low score; it is an empty cell that we would rather show than fill in.
T1 — Fiat-first
Crypto is a payment method and nothing else. The operator is a conventional house — usually with a fiat cashier, fiat accounting and a fiat licence — that accepts coins at the till, often through a payment processor that converts on receipt.
The signal is that the crypto layer is outsourced: deposit addresses rotate through a third-party processor, the balance is denominated in a currency rather than an asset, and nothing in the documentation discusses chains. There is nothing on-chain to verify because the operator never intended there to be.
T2 — Crypto casino
Deposits and withdrawals happen on-chain; the games are a black box.
This is the rung most recognisable operators occupy, and it covers a very wide range of quality. You deposit to an address, the operator credits an internal balance, and you play games served from its own infrastructure. Provably-fair seed pairs may well be published — see Provably fair, explained — but they are the operator’s own records. Accounts, game logic and the bankroll all live on private servers, and no individual bet is verifiable from chain data.
As of September 2026, Stake, Shuffle, Rollbit, Gamdom, Roobet and Yeet are among the rows we record at T2. Being at T2 is not a mark against an operator’s payment record; it is a statement about what an outsider can confirm.
T3 — Hybrid
Crypto rails plus a genuine on-chain artefact, while the games themselves still run mostly off-chain.
The artefact has to be real and it has to be readable. A public treasury or bankroll contract qualifies. In-house originals anchored on-chain qualify. A token with an on-chain revenue mechanism qualifies in part. Marketing that describes the site as “built on” a chain does not.
EarnBet, Dexsport and CoinPoker sit here in our current reading, as of September 2026 and from public documentation rather than a deposit test. T3 is the rung where a reader has to do the most work, because the on-chain portion varies enormously between operators — the guide on reading a treasury is the practical follow-up.
T4 — Fully on-chain
Game logic, randomness and payout all execute in contracts on a public chain, played from a self-custodied wallet. Anyone can replay the outcome from chain data.
The test we apply is the replay test: given nothing but the chain, can a stranger reconstruct the bet, the random draw and the settlement? If a private server has to be asked about any of the three, the operator is not at T4.
Solpump, Solpot, JustBet and Overtime are the rows we currently record at T4, again as our reading of public architecture rather than a verified finding. T4 operators are typically narrower products with smaller catalogues and, often, much shorter operating histories. That trade-off is real and it shows up in their survival sub-scores.
What the tier does to the score
One rule, and it only ever reduces:
A custodial operator at T2 or below has its transparency sub-score capped at 12 out of 25.
The reasoning is architectural. If the funds are held by the operator and none of the game logic is on a public chain, there is a ceiling on how much an outsider can verify, no matter how much documentation exists. Publishing more does not break the ceiling; changing the architecture does.
The tier does not touch payout, survival, freedom or value. A T4 operator with no operating history scores badly on survival, and should. A T2 operator with years of visible withdrawals scores well on payout, and should. The full rule set, including the two other caps, is on the methodology page.
The question mark
A tier written as T3? or T4? means the rung is claimed rather than confirmed by us. There are two ways that happens: the operator asserts the tier and we have not been able to verify it, or the operator asserts the rung above and we have recorded the lower one until we can check.
We record claims as claims, with the date we read them. We do not promote a row because the marketing is confident.
Movement is expected
Tiers describe the current stack, and stacks change. An operator that ships a bankroll contract moves up; one that quietly retires an on-chain component, migrates originals to private servers, or lets an address go stale moves down. We date every change and keep the previous reading visible rather than overwriting history.
Downward movement is worth watching for another reason: it is often an early symptom of the pattern described in Why crypto casinos die. Infrastructure gets retired when it stops being paid for.
You can filter the whole index by tier on the trust score ranking, or put two operators from different rungs side by side on the compare page.
18+ only. Gambling involves real financial risk and can be addictive. Nothing here is financial, legal or gambling advice.
Questions
Is a higher tier a better casino?
No. The tier describes the architecture, not the quality. A T4 operator can be young, thin and badly run; a T2 operator can have years of paid withdrawals behind it. The tier tells you how much you can check without permission, which is a different axis from how well the site works.
Why does the tier cap the transparency score?
Because at T2 and below with custodial funds, no bet can be reconstructed from chain data. There is a ceiling on how much an outsider can verify, whatever documentation is published, so transparency cannot pass 12 of 25.
What does the question mark after a tier mean?
That the tier is claimed rather than confirmed by us. Either the operator asserts it and we have not verified it, or it asserts the rung above and we have recorded the lower one until we can check.
Can an operator move down the ladder?
Yes. Tiers describe the current stack. An operator that retires an on-chain component, migrates games to private servers or takes a bankroll contract offline moves down, and we date the change.
How do you decide a tier without depositing?
From public evidence: documentation, contract addresses, explorer data and the interface itself. Where that evidence is thin we record the lower rung and mark it. Every tier on the site today is a desk reading, because no operator in the index has been deposit-tested.
How the scores behind these guides are built
Five weighted criteria, three published caps and a tier ladder. Every total is a seed score you can recompute from the public registry.