How we rate sweepstakes casinos — our testing method

Our method

How we test and score

We fund every account with our own money, play the sweeps balance down, file a real redemption and time it. Three times, on three separate weeks.

We pay for our own accounts

Nobody on this site gets a free account, a comped balance or a test environment. We sign up the way a reader does, buy a coin package the way a reader would, and go through the same identity check. Where a brand offered us press access, we declined it, because an account that a brand knows is being watched is not an account worth writing about.

The scoring weights

Payout behaviour is 40% of the score — median time from filing to funds, how close the slowest run was to the median, and whether the amount that arrived matched the amount requested. Terms are 30%: the sweeps figure relative to the headline, the play-through multiple, the redemption floor and whether anything unpleasant is buried in the rules. Games and software are 20%. Support is the last 10%, tested by sending a real question with a real problem in it.

Three redemptions, not one

A single fast payout proves very little. Brands run promotions, queues empty on quiet weeks, and a first-time redemption sometimes gets hand-processed. So we file three times, spaced across three weeks, and publish the median rather than the best. If the slowest run was far off the median we say so in the review, because inconsistency is its own kind of problem.

What we deliberately ignore

Welcome package size in gold coins. Total game count. Anything a brand describes as exclusive without saying what it is exclusive to. Loyalty tiers you would need months to reach. None of these correlate with getting paid, and all of them are easy for a brand to inflate.

Commercial relationships

Some links on this site pay us if you sign up. The rate is agreed before testing starts and never changes afterwards, and it has no bearing on the score or the position. Several brands in the lower half of our ranking have no commercial relationship with us at all, and at least two brands that pay well sit below brands that pay us nothing.

When a score changes

We re-test the full list twice a year and spot-check the top five quarterly. A score moves when the numbers move: a change to the play-through multiple, a change to the redemption floor, a payout time that drifts, or a support desk that stops answering. When a score moves we say why at the top of the review rather than quietly editing the figure.

What a score is not

A rating here is a summary of measured behaviour on a funded account over a fixed period. It is not a prediction, and it is not a recommendation to spend money. Treat every figure on this site as a starting point for your own reading of the brand’s rules.

See what the method produced

Eighteen brands in order, with the numbers each score was built from.