The scoring model on this site is not a black box. It is six weighted criteria: licence, payouts, bonus fairness, game range, support, and responsible-gambling tools. Every quarter we look at what our own review notes say about how well those weights predict player experience. When the weights are out of step with what actually matters, we change them and record the change here so the reasoning stays public.
This quarter the headline is payouts. We moved payout weight up two points, re-tiered the old Curacao licences, and re-ranked three brands. The detail, and the arithmetic behind it, is below.
How we recalibrate, and why quarterly
The model is only as good as the evidence behind the weights. Each cycle we pull three things together: the tips and complaints readers send us, the withdrawal timings and bonus outcomes we log during hands-on testing, and any regulatory changes in the markets we cover. We then ask a narrow question for each criterion. Does the weight it carries still match how much it separates a good operator from an average one?
If a criterion no longer sorts operators, its weight is too high. If a criterion sorts them sharply but sits low in the model, its weight is too low. The quarterly pace is deliberate. Move weights every week and the rankings churn for no reason. Move them once a year and the model drifts out of date. Three months is long enough to gather real timing data and short enough to stay current.
Weight change: payouts 20% to 22%, game range 15% to 13%.
The single most common complaint in the tips we get sent, and by a wide margin, is slow payouts. That is not new. What is new is that the gap between the top payout operators and the median has widened, not closed, since we last recalibrated in January. Median UK e-wallet withdrawals now clear in about eight hours; top-quartile operators clear in under two. The bottom quartile is still stuck at 48-plus. When the top and bottom of the market are that far apart on a criterion, the weight it carries in a score has to reflect that.
Game range gave up two points to fund the increase. This is not because catalogue size does not matter, it does, but because the marginal value of the ten-thousandth game in a catalogue is close to zero, and most operators cluster in the same two-thousand-plus range. It stopped separating good from average. A player notices a slow payout every single time they cash out. Almost nobody notices the difference between 4,000 and 6,000 games.
A worked example: what two points actually does
Weights only matter if they move scores, so here is the arithmetic on a realistic profile. Take an operator that scored 8.0 out of 10 overall in January. On payouts it rated 9 out of 10, with a fast and consistent e-wallet lane. On game range it rated 6, a mid-size catalogue.
Under the old weights, payouts contributed 9 x 0.20 = 1.80 and game range contributed 6 x 0.15 = 0.90, for 2.70 between the two criteria. Under the new weights, payouts contribute 9 x 0.22 = 1.98 and game range 6 x 0.13 = 0.78, for 2.76. That operator gains 0.06 overall for doing nothing, purely because the model now rewards the thing it is already good at.
Now flip the profile. An operator with a strong catalogue but slow cashouts, say 9 on game range and 5 on payouts, went from (5 x 0.20) + (9 x 0.15) = 2.35 to (5 x 0.22) + (9 x 0.13) = 2.27, a loss of 0.08. Neither operator changed anything about how it runs. The re-weighting alone re-sorted them by roughly a seventh of a point, which is enough to swap two closely matched brands in a ranked list.
Curaçao tier reset: legacy 1668/JAZ, 8048/JAZ and 5536/JAZ downweighted.
Curacao's regulator, the Gaming Control Board, has finished the first year of its new direct-licensing regime. New OGL/2024/ numbers are being issued directly to operators, not stacked under a master licence-holder. That is a meaningful improvement in accountability. The regulator now has a direct line to each operator instead of going through an intermediary who has no incentive to enforce anything.
We are treating the two tiers differently. Operators with a fresh OGL/2024/ or later licence get the standard Curacao licence score of 6.5. Operators still running under the legacy master-sub structure, 1668/JAZ, 8048/JAZ, 5536/JAZ and their peers, now score 5.8 for the licence criterion. The gap will widen further as those legacy licences approach expiry. We would rather nudge operators toward the more accountable structure than pretend the two are equivalent.
Three brands re-ranked.
Two moves up, one down. The two upward moves both came from payout audits. One UK-licensed operator confirmed a fresh e-wallet lane averaging 45 minutes across five paydowns, and one Swedish brand cleared its Swish withdrawals in under 20 minutes on three consecutive attempts. Both moved up one score band. The downward move came from a Curacao operator whose licence turned out to be a legacy sub-licence about to expire, not a fresh OGL number as their footer implied. They moved down one band.
We do not name the brands in change logs, only in the reviews themselves, because a change log entry is a snapshot and the ranked pages are the current picture. If you want to see where a specific operator sits today, check its review page. That one is always current.
What did not change.
Licence still carries 25% of the score, and always will. Bonus fairness stayed at 20%. We looked at moving it up to reflect how often bonuses are the reason someone deposits, but the ranking of operators does not change much when we do, so we left it. Support stayed at 10%. Responsible gambling tools stayed at 10% with the same caveat we noted last quarter: this weight will go up materially once we have manually audited the RG toolset of every operator on the site. That audit is in progress.
What this means if you read our rankings.
If a casino you follow moved a place or two this quarter and nothing about it seems to have changed, this is probably why. The operator did not get better or worse. The model started weighting payout speed more heavily, and the operator either benefited or paid for that depending on how fast it actually pays. Read the payout line on any review as the number that now carries the most swing after licence.
The practical takeaway is simple. A high overall score sitting on a weak payout line means less than it did in January, and a slightly lower overall score built on a fast, verified payout lane means more. If you value getting your money quickly, weight the payout figure yourself even more heavily than we do.
Next recalibration.
October, if nothing dramatic happens before then. Big regulatory changes, the ANJ decree, the UKGC financial-risk-check thresholds, Dutch deposit-limit rules, will move weights immediately, not on the quarterly cadence. We publish the change here when it happens.