How we score operators, what counts as a verified incident, and what we explicitly refuse to do for affiliate money.
How we decide what goes on the Fairness Ledger, how we score operators, and what we explicitly refuse to do for affiliate money.
Every operator we review is scored 0–5 against the same five tests. Pass = ✓, fail = ✗, mixed = ½. Public ranking is the sum.
Does the operator publish a SHA-256 of the server seed BEFORE accepting bets? The hash must be visible client-side, not buried in a docs page.
When the server seed rotates, the OLD seed must be revealed. We re-compute its hash and verify it matches what was published.
The player must be able to set their own client seed. Operator-only RNG = no provably fair.
The outcome formula must avoid modulo bias (use rejection sampling or a domain that divides cleanly). We test this with the modulo bias detector and a chi-square sample.
The exact outcome-derivation algorithm must be documented. "Trust us" is not provably fair.
An incident moves from alleged to verified only on one of:
Reddit threads and Twitter screenshots are alleged-tier — useful as starting points, never enough on their own.
We do accept affiliate revenue, but ONLY from operators currently scoring 5/5 on the five criteria. If an operator's score drops, the affiliate link comes off the same day. Current list is on the operators page.
Send leads (with sources) to [email protected]. We publish credit unless you ask for anonymity. Corrections to existing entries go to [email protected].
Most “operator review” sites are affiliate marketing. The editorial position is downstream of the affiliate payout schedule: pay more, score higher. This produces leaderboards that are functionally reverse-correlated with player welfare — the operators most willing to skim from players have the biggest affiliate budgets.
Our methodology fixes that by being public and falsifiable. Every operator we list is scored against the same five technical criteria. If an operator fails a criterion, the score drops. We don’t take affiliate money from operators that fail any criterion — so there is no incentive to soften the audit.
Of the five criteria, the most-violated in the wild is criterion 2 (seed reveal on rotation). It’s also the cheapest for an operator to fake: publish a static hash, never rotate, claim “PF” on the landing page. Players almost never check, so the operator gets the marketing benefit without the engineering cost.
Our audit reverses this: we routinely sample published hashes from operators we cover, wait for rotation, attempt SHA-256 verification of the revealed seed, and publish the result. Operators that don’t rotate get flagged. Operators whose reveals don’t match the published hash get added to the Fairness Ledger.
Tips arrive at [email protected]. Every submission is independently reproduced before publishing:
Every operator we cover is welcome to dispute any entry. If they can show technical evidence that contradicts ours, we update the entry and credit them. We have never softened an entry in exchange for promotional consideration; we have repeatedly softened entries in exchange for hard evidence. The two are not the same.