How we test casino payment methods in Australia

What we measure, and what we refuse to measure

A payment rail is only interesting in two moments: when you fund an account, and when you try to take money out. Everything else — bonus terms, game libraries, interface polish — is a different review. So each test records six things: the rail used, the amount in Australian dollars, the seconds from pressing confirm to a playable balance, the minute the withdrawal was requested, the minute the funds were usable in the destination account, and any identity check that interrupted either leg.

The withdrawal timing is the part most comparison sites get wrong. We time to funds being usable in the receiving account, not to the operator marking a request 'processed'. Those two timestamps can be a full business day apart on card rails, and quoting the flattering one makes a slow payout look instant. Where a rail cannot receive a payout at all — paysafecard and Neosurf vouchers, phone wallets and biller payments — we record the substitute rail the money actually returned on, because that is the wait you will experience.

We do not measure things we cannot observe. We do not publish a security score, because we cannot audit an operator's infrastructure. We do not rate 'fairness', because we cannot inspect game RNG certificates ourselves. And we do not estimate a figure when a test failed; a failed session is logged as a failure and shown in the results rather than quietly dropped.

• Deposit time: from confirming payment to a playable balance, in seconds.
• Payout time: from submitting a withdrawal to usable funds, in minutes.
• Verification: every ID, selfie or source-of-funds check that interrupted a leg.
• Cost: any fee charged by the operator, the rail or the bank, in AUD.
• Return path: which rail the money actually came back on.

The test bench: accounts, amounts and timing windows

We hold funded accounts at five operators. Every account was opened by us, verified with our own documents, and funded from our own bank accounts and cards. We do not accept free test balances from anyone: a balance we did not fund is a balance whose withdrawal we cannot honestly verify, because the operator knows it never carried real settlement risk.

Each rail receives eight transactions per operator, spread across deposit sizes from A$10 to A$500. Small deposits and large deposits behave differently — larger amounts trigger source-of-funds questions far more often — so an average built only on A$50 tests would be misleadingly fast. Amount bands are held constant across rails so the comparison is like for like.

Submission windows matter as much as amounts. We run morning, evening and overnight cycles in Australian Eastern time, because the dominant variable in a payout is not the rail, it is whether a human on the operator's payments desk is awake. Overnight requests routinely land an hour or more slower than the same request submitted at 10am on a Tuesday, and that pattern shows up on every rail we bench, from PayID to Apple Pay.

How a rail earns its score

Each rail carries a score out of ten, and the weighting is published rather than implied. Speed of the return leg is the heaviest factor because it is the thing readers complain about most and the thing operators control most. Two-way capability comes next: a rail that cannot pay you back forces you into a second, unverified rail at the worst possible moment.

Cost is weighted third. Most Australian bank rails are free at the point of use, so this factor mainly separates cards — where international transaction fees and cash-advance treatment can quietly add several percent — from everything else. Reliability, meaning the share of sessions that completed without a decline or a hung redirect, rounds out the score.

Nothing in the score is influenced by commercial relationships. Ranking changes are recorded in the changelog on the homepage with the date and the reason, so a shift in position can always be traced to a measurement rather than a deal.

How operators get on the list — and how they come off it

An operator only enters the bench once we hold a funded, verified account and have completed at least one full deposit and withdrawal cycle. Nothing is listed on the strength of an affiliate feed. If we cannot withdraw, the operator does not appear, regardless of what commercial terms are on offer.

Removal is faster than addition. A refused or indefinitely stalled withdrawal, a change of terms that retrospectively affects existing balances, or a support channel that stops answering are all grounds for immediate removal while we re-test. We publish the reason in the changelog rather than letting a brand quietly disappear.

Australian readers should understand the regulatory backdrop, which our fact-checker verifies against primary sources on every page. Under the Interactive Gambling Act 2001 it is unlawful to provide most online casino games to people in Australia, so the operators Australians encounter are licensed offshore. That is a material risk, and we state it on every page rather than burying it — it is also why our testing focuses so heavily on whether money actually comes back out.

• Funded, verified account required before listing.
• At least one complete deposit and withdrawal cycle before publication.
• Immediate removal on a refused or indefinitely stalled payout.
• Every listing re-tested at least once per publishing cycle.
• Offshore licensing risk disclosed on every page.

Funding, affiliate links and editorial independence

This site earns commission when a reader signs up through some links. That funding pays for the deposits we make and the staff who make them. It never changes what you pay, and it does not change a ranking: the people who negotiate commercial terms have no input into scores, and the test log is written before anyone looks at whether a brand has a programme at all.

The practical safeguard is that the raw data is published. Every transaction that produces an average is visible in the test log on the relevant payment page, with its date, amount, timings and verification outcome. If a ranking did not follow from the data, the discrepancy would be visible to any reader who reads the table — which is precisely the point of publishing it.

Corrections work the same way. If you can show a figure is wrong, we will re-test rather than argue. Send the page URL, the disputed figure and any evidence to the editorial team via our contact page; verified corrections are normally published within five business days and the change is dated in the changelog.

The limits of this methodology

A bench of 320 transactions across five operators is large enough to expose consistent patterns and far too small to be a statistical guarantee. Our averages describe what happened to us, in Australia, in a specific ten-week window. An operator can staff its payments desk differently next month, a bank can change its gambling-block policy overnight, and a card issuer can start declining a merchant category it previously waved through.

Rails change too. Payment services withdraw from the Australian market, wallets update their acceptable-use rules by jurisdiction, and cashier menus are quietly re-arranged. Where a rail we previously covered stops being available to Australians, we retire the page rather than leave stale advice in the index — the wallet, card and voucher rails still under active test are the PayID, Visa, Mastercard, PayPal, Apple Pay, Google Pay, paysafecard and Neosurf pages — which is why the catalogue on the payment methods hub lists more rails than we have full test logs for.

Finally, this is payments testing, not gambling advice. Nothing here suggests any of these products are a way to make money. Gambling is entertainment with a negative expected return, and the fastest withdrawal in our log is still a withdrawal of money that was more likely than not to shrink. Free, confidential help is available from Gambling Help Online on 1800 858 858 and self-exclusion through BetStop.

Frequently asked questions