Why Our Methodology Is Public
Most betting review sites won't show you how they score — because many simply republish affiliate marketing material. We publish our methodology because the whole value of a review site is trust. If you can't verify how a rating was produced, the rating is worthless. This page documents exactly what we test, how we score, and what we refuse to accept from operators.
The Testing Pipeline: Real Money, Every Time
Every bookmaker in our rankings goes through the same pipeline:
- 1. Registration & KYC — we sign up with real details and document the verification process and how long it takes
- 2. Deposit test — a real deposit ($15–$50) via the site's most popular method, timed
- 3. Bonus claim — we claim the welcome offer and audit every term against industry norms
- 4. Betting test — pre-match and live bets across football and cricket, testing bet acceptance speed
- 5. Odds audit — 50 markets compared against 6 rival bookmakers to measure margins
- 6. Withdrawal test — a real withdrawal request, timed with a stopwatch to the minute
- 7. Support test — three live-chat contacts with real questions, response times logged
The 5 Scoring Categories
Each bookmaker is scored 1–10 in five categories, then weighted into a final 5-star rating:
| Category | Weight | What we measure |
|---|---|---|
| Odds & Markets | 25% | Margin on 50 markets, market depth, niche sports |
| Bonuses | 20% | Bonus size, wagering fairness, max bet rules, expiry |
| Payments | 20% | Deposit speed, measured withdrawal time, fees, method variety |
| Live Betting | 20% | Bet acceptance speed, streaming, cash out, live market depth |
| Support & Trust | 15% | Licences, complaint history, support quality, responsible gambling tools |
Trust Scoring: Licensing and Beyond
Trust isn't vibes — it's checkable facts. UKGC and MGA licences score highest because they give players real dispute resolution. Curacao licences are legal but offer fewer protections, so those sites must compensate with payout history and transparent behaviour to score well. We also check operator complaint patterns, payout disputes and sanctions history before publishing any review.
Bonus Scoring: Size Is Not Value
A 200% bonus with 60x wagering on deposit+bonus scores worse than a 100% bonus with 25x on bonus only. We calculate the effective value of every offer — expected cost to clear, max bet rules, game weighting and expiry — and score that, not the marketing headline. Our bonus guide teaches you to do the same maths.
How Affiliate Relationships Work (And Don't)
PagBets.org earns affiliate commissions when readers register through our links — this is how the site stays free. Three rules keep that from corrupting our content: scores are produced by test data before any commercial conversation; reviews are never shown to operators before publication; and commercial teams cannot change a rating. Our full disclosure is on the Affiliate Disclosure page.
Update Cadence
Markets move fast. Every review carries its last-updated date, and we re-test each bookmaker at least quarterly. Bonus changes, licence changes or withdrawal complaints trigger an immediate re-test. If a site's score drops, the published review drops with it — no exceptions.
Why Our Scores Move Between Quarters
Scores change four times a year, and the change is a feature rather than an inconsistency.
An operator's payout speed, support quality and margin do not stay still, and a score that reflects a test from nine months ago is a historical record rather than a current assessment. What moves a score is evidence: a withdrawal that takes four days where it previously took one, a support agent who cannot answer a payment question, a price that has drifted measurably wider on a market we track.
Every re-test follows the same protocol so the comparison is fair. Deposit, bet, withdraw, then dispute. The dispute step is the one most reviews omit and the one that separates an operator that has good policies from one that honours them. Our the withdrawals guide covers what we time.
What We Deliberately Do Not Score
Three things we do not score, because scoring them would make the scores worse at being useful.
We do not score how many sports a bookmaker covers, because breadth is a product choice rather than a quality measure and a player who only wants football does not benefit from a bookmaker covering eight more sports. We do not score the number of casino games, for the same reason. And we do not score how generous a bonus looks, because that would reward the operators whose promotions are worth the least. Our the free bets guide covers how we value a promotion instead.
How the Odds Audit Actually Works
The odds audit is the part of our methodology most likely to be challenged, so it is worth describing precisely. For each bookmaker we take 50 markets across football, cricket, basketball, tennis and esports, sample them inside a fixed window before kickoff, convert each price to implied probability, divide each by the bookmaker's own overround on that market, and take the mean across the 50. The result is a percentage margin comparable between operators. Our margin guide explains the underlying arithmetic.
Sampling inside a fixed window matters more than most readers expect, because prices move. If one bookmaker's prices are sampled an hour before kickoff and another's at the line, the comparison measures timing rather than quality. We sample within a fifteen-minute window and record the timestamp for every market, which is why our odds figures can be reproduced by anyone who checks at the same time. Margin figures on this site carry that caveat by design.
The 50-market sample is a compromise, and it is biased in a way we accept. Sampling evenly across sports avoids a bookmaker's cricket strength being hidden by a football-heavy set, but it also means a bookmaker genuinely excellent at tennis and cricket markets can score lower than one that is uniformly adequate. Where we identify a specific strength the headline number misses, we say so in the review text rather than adjusting the score. Our value betting mathematics guide covers why a sharp average matters more than a good average.
One number cannot describe a sportsbook's pricing, which is why we treat the margin figure as a summary rather than a verdict. Two operators both at 4.5 percent on average can differ by a full point on any single match, and a bettor placing one Premier League bet a week should compare that specific market rather than rely on an annual figure. Our ranked betting sites page publishes the per-sport breakdown where we have measured it.
Esports markets are included deliberately even though most affiliate comparisons exclude them, because the spread on esports between operators is wider than on football. Our esports betting guide covers the market structure; the methodological point is that excluding a market where pricing varies widely makes a margin average look better than the bettor's experience of it.
From Five Scores to One Number
Five categories weighted into a single 5-star score is a convenience with a cost, and the cost is that one number hides the profile behind it. We publish the breakdown on every review because a 4.5 built on strong odds and weak support is a different proposition from a 4.5 built on the reverse. Our bookmaker selection guide walks through matching a profile to a type of player, which is the step the single score cannot do for you.
The weighting reflects what we observe in testing rather than what sounds reasonable. Odds at 25 percent reflects that margin compounds across volume and cannot be recovered later. Bonuses at 20 percent is lower than most affiliate sites assign, because a bonus is a one-time transfer while odds apply to every bet. Payments at 20 percent reflects that an account you cannot withdraw from is not a bookmaker at all, whatever its price. Support and trust at 15 percent is the lowest weight, and we think that is defensible.
The obvious criticism is that trust should carry more than 15 percent, and there is a reasonable argument for it. Our answer is that trust operates largely as a filter rather than a score: an operator without a verifiable licence, or with a pattern of unpaid withdrawals, is excluded outright rather than scored down. Once that filter has run, the residual difference between a UKGC licence and a Curacao one is about dispute handling quality, which is worth less than a point of margin to most players.
Rounding is the other place the single score misleads. Two operators at 8.9 and 8.8 in a category are not meaningfully different, and we have published identical star ratings for operators with quite different profiles. Treat the star rating as a sorting tool and the breakdown as the actual information. Where a score sits near a boundary, 3.9 against 4.0, the difference falls inside our own measurement error and should not drive a decision.
The category scores are also subject to a floor and a ceiling. An operator with a genuinely excellent product cannot score above 9.5 in any category because we reserve the top band for sites that are exceptional on that dimension, and no operator in testing has earned it. Our casino app rankings apply the same scale, which is why almost nothing scores a perfect ten.
Where This Methodology Falls Short
Three honest limitations. First, we test as a single customer with a modest amount of money, so we cannot see how a player turning over six figures is treated — which is exactly the segment that encounters limits and delayed withdrawals. Second, we are not resident in most of the markets we cover, so our view of regulatory conditions is documentary rather than lived. Third, our test window is a fortnight, which is too short to see seasonal degradation.
The single-customer limitation is the most serious, because limit behaviour is where players are most commonly failed and precisely what we are least able to observe. A bookmaker that pays everyone under 1,000 dollars promptly and begins delaying at 5,000 dollars will pass every test we run. Our payment methods guide covers the withdrawal mechanics we can test; it cannot cover this. Treat the payments score as a floor rather than a ceiling.
The documentary limitation affects the legal sections of our market guides. We read the legislation and the regulator's register, which tells you what is permitted on paper rather than what is enforced. Where the two diverge — and in several markets we cover they do — the practical answer rests on enforcement practice, which we report from published sources and player accounts rather than measure directly. Our fake betting site guide is still the first routine we would run.
The two-week window is a deliberate trade. A longer test would cost more per review and reduce the number of operators we can cover, and we would rather assess more operators badly than fewer operators well. The mitigation is the quarterly re-test: a score that survives four consecutive test windows is more trustworthy than one from a single fortnight. Our responsible betting tools guide is the other half of our published standard, because a rating that ignored player protection would not be worth much.
The largest gap is the one nobody in this industry publishes: how long a review takes. Ours is roughly 12 to 18 hours per bookmaker across registration, funding, betting, withdrawal and three support contacts, which is why we cover fewer operators than some competitors and re-test fewer of them. Anyone claiming a comprehensive ranking of hundreds of bookmakers updated monthly should ask what that claim costs, and the answer is usually that it is not testing.
✓ What We Like
- Tells you how to check a licence rather than asking you to trust one
- Covers the payment problem that decides whether the product is usable for you
- Explains the tax structure, which is the part that sets the price
- Uses named regulators, laws and figures rather than general description
✕ What Could Improve
- Several rules are being tightened, so a figure written today may not hold
- Applies to one jurisdiction, so it does not transfer to your market
- A limit or ceiling constrains how much you can actually stake


