Fraud detection built into the platform

Each player is checked against a set of rules. Every rule that fires raises the player's risk score, and the score places them in a band — green, yellow or red — shown next to the player wherever staff work with them.

Scores are recalculated at the moments that matter — registration, bonus claims and withdrawal requests — and refreshed for every player on a regular schedule, which is when links between accounts are recomputed. Staff can also recompute a score on demand from the dossier.

Every fired signal is shown with the details that triggered it. A support agent explaining a held withdrawal needs to be able to say why, and an operator reviewing a case needs to see what the system saw, not just a number.

The three ways a casino is usually exploited: several accounts run by one person, a referral programme paid for players who do not really exist, and bonus money taken out without playing it through.

No single signal proves fraud: housemates share an IP, a travelling player changes country, a cautious one uses a throwaway address. The score weighs signals against each other, so an honest player who trips one check is not treated like an account that trips several at once.

From a review queue filtered by band, decision, score and search, and a dossier for each player: the signals that fired with their details, the other accounts on the same IP, the other accounts on the same device, and the player's own devices.

The overview shows how players are distributed across the bands, how many are on hold and which signals fire most often. A signal that suddenly fires for many players at once — a wave of registrations from one source, say — shows up there before it shows up in withdrawals.

The dossier's linked accounts are usually what settles a case. One suspicious account is a judgement call; several accounts on one device that all claimed the same welcome bonus is not.

Two decisions. HOLD blocks the player's withdrawals, bonus claims and promo code activations. CLEAR marks the player as reviewed, so they pass checks even in the red band. Both ask for a fresh two-factor code and can carry a note.

A hold stops money leaving and stops more promotional money arriving, without banning the account — the player can still log in, and the case can be resolved either way. If the review ends badly, a ban is a separate action.

CLEAR is what stops a legitimate high-value player from being flagged again every time they travel or change phones. Every decision is written to the audit log with who made it.

It can. How the red band is enforced is agreed with each operator, because the right setting depends on the brand's traffic. Manual holds work from day one.

The reason for agreeing it rather than switching everything on is the cost of a false positive. A blocked withdrawal from a genuine depositor costs more — in support, in chargebacks, in reputation — than most individual frauds, so automatic enforcement is tuned against real decisions on the brand's own players rather than against someone else's.

A player on hold cannot create a withdrawal request at all. For everyone else, every withdrawal is approved by a person, and the player's risk profile is one click away from the approval screen.

Withdrawals sit in escrow until a decision, so checking a dossier does not put anything at risk — the money has already left the player's balance and has not yet left the casino. Declining returns it to the player; declining and holding keeps the account from trying again. The flow is described on payments.

No. The fraud system scores behaviour on the platform — accounts, devices, referrals, bonuses and payouts. Identity and age verification is done with a KYC provider, and card and transaction screening by payment providers.

The two complement each other. KYC tells you who a player is; behavioural scoring tells you what they are doing with the casino's promotions and referral programme, which is where most casino-specific abuse happens.

How the rest of the platform is secured — hosting, isolation between brands, staff permissions and the audit log — is on security.