Skip to content
100-POINT TRUST INDEX · v3.3

Our 100-Point Trust Index Methodology

This page is the source of truth for every score on this site. We do not accept paid placements, we score every brand against the same seven weighted criteria, and we publish the formula behind each sub-score in full. The Trust Index is a synthesis of 2,127 timed withdrawals, 4,800+ hours of gameplay, 4,200 audited sessions and $50,000 of our own bankroll committed during the 90-day 2026 cycle.

SCORING RUBRIC

What every Trust Index score actually means

The Trust Index is a 100-point composite; a 9.6 is not "almost a 10", it sits inside a specific recommendation band with documented re-audit cadence. The bands below are the rubric every operator is mapped to before publication.

Elite 9.5–10.0

Top of our list, recommended without caveat. Independent re-audit shows zero regressions across the previous two cycles.

Excellent 9–9.4

Recommended, minor friction (slightly tighter wagering, slower KYC, or a narrower studio shelf).

Strong 8.5–8.9

Recommended with reservations, typically a specific weakness (single-network withdrawals, narrow bonus EV, or older support stack).

Fair 8–8.4

Edge of the watch-list. Track-record acceptable but at least two criteria below the 8.0 floor; re-audit cadence is tightened to 30 days.

Rejected < 8.0

Excluded from the ranking. Either a critical criterion below 6 or a structural integrity issue. Logged on /blacklist with the rejection reason.

A 9.6 versus a 9.7 represents a single sub-criterion gap, typically a 0.1 Trust-Index point either way maps to a measurable difference in median withdrawal speed, in bonus EV per dollar committed, or in the number of complaints per 1 000 active accounts during the audit window. The decimals are not cosmetic; they are the resolution of seven separate measurements.

SEVEN WEIGHTED CRITERIA

The seven criteria, and the formulae behind each one

Every operator on our ranking is scored against the same seven dimensions. Each criterion is graded out of 10 and multiplied by its published weight. The weights total 100, which gives the operator a single Trust Index in the 0–100 space. Below each criterion we publish the exact sub-formula, the inputs that feed it, and the sample sizes behind the inputs.

Regulatory Depth

22%

License tier · parent transparency · jurisdictional history.

22 / 100
Formula
license_tier × 0.4 + corporate_transparency × 0.3 + jurisdiction_history × 0.3
Sub-input Scale How we measure it
License tier 0–10 Master licence (MGA, IOM, UKGC) = 10. Direct Curaçao OGL under the LOK regime (multi-product, e.g. casino + sportsbook) = 9. Single Curaçao OGL (casino-only) = 8. Anjouan ALSI = 7. Costa Rica data-processing only = 3.
Corporate transparency 0–10 Parent company verifiable in public registry, listed address Google-able, named compliance officer published = 10. Footer-only license number with no public parent = 4.
Jurisdiction history 0–10 Years on the active licence without suspension, supplemented by any sanction history. 10+ years clean = 10. Under 18 months = 6.

Withdrawal Reliability

20%

Median wallet-arrival time · first-attempt success · manual-review rate.

20 / 100
Formula
min(10, 10 − ln(median_seconds / 8)) · first_pass_rate · (1 − manual_review_rate)
Sub-input Scale How we measure it
Median (seconds) s Sample N ≥ 96 per operator. Stopwatch starts on confirm-click in the cashier, stops on wallet balance change at the receiving address.
First-pass success rate % Share of withdrawals that landed without operator-side intervention (manual review, KYC pop-up, fraud-flag hold).
Manual-review rate % Share that triggered an analyst review. Below 5% is exemplary; above 30% materially drops the score regardless of median.

Game Library

16%

Studio breadth + published RTPs + live-dealer depth.

16 / 100
Formula
studio_breadth × 0.4 + rtp_publishing × 0.3 + live_dealer_depth × 0.3
Sub-input Scale How we measure it
Studio breadth 0–10 Pragmatic, Evolution, NetEnt, NoLimit, Hacksaw, Push, Relax = baseline 7. Add 0.5 per additional tier-1 studio above. 10 = also stocks BGaming, Avatar UX, Spinomenal at depth.
RTP publishing 0–10 Per-title RTP exposed in the game card UI = 10. RTPs only on lobby filter = 7. Studio aggregate only = 4. No RTP = 0.
Live dealer depth 0–10 Evolution + Pragmatic Live + Ezugi + Playtech = 10. Evolution only = 7. No live tables = 0.

Bonus Fairness

14%

Wagering math · max-bet rule · game weighting · cashout cap.

14 / 100
Formula
10 − (wagering_multiplier ÷ 5) − max_bet_penalty − game_weight_penalty − cashout_cap_penalty
Sub-input Scale How we measure it
Wagering multiplier × Bonus-only base preferred. 25× bonus-only ≈ same EV as 12× deposit+bonus. We always model the math you actually need to clear.
Max-bet rule usd Penalty applied when bet caps fall below $5 during clearance. A $1 cap can stretch a 35× clearance from a one-evening grind to a three-week project.
Game-weighting table % Slots 100% / table 10% is industry baseline. Penalty if live dealer or roulette contribute 0%; bonus often becomes mathematically unclearable.
Max cashout cap × Cap below 10× deposit is a structural EV cap. We surface the cap in the bonus card and re-rank if the operator hides it.

Crypto Variety

12%

Networks supported on deposit AND on withdrawal, gated separately.

12 / 100
Formula
(deposit_networks + withdrawal_networks) ÷ 4 (capped at 10)
Sub-input Scale How we measure it
Deposit networks count BTC mainnet, BTC Lightning, ETH, SOL, USDT-ERC20, USDT-TRC20, USDT-SPL, BNB-BSC, LTC, DOGE, XRP, ADA each count once. Stablecoins on different rails count separately.
Withdrawal networks count Same registry. Operators that accept a chain on deposit but withhold it from withdrawal lose the chain entirely from their score.

Dispute Handling

10%

Complaint volume · resolution rate · mediation availability.

10 / 100
Formula
(resolution_rate − complaint_density × 0.05) · 10
Sub-input Scale How we measure it
Resolution rate % AskGamblers + Trustpilot weighted average. Below 60% materially capsizes the score.
Complaint density per 1k Complaints per 1 000 active monthly accounts. Industry median sits at 1.2.
Mediation channel flag Working AskGamblers escalation, named compliance officer, ADR registration. Each adds 0.5; combined ceiling 1.5.

Provably Fair Coverage

6%

Native game depth + working seed-verifier in UI.

6 / 100
Formula
native_game_count × 0.6 + ui_verifier × 4
Sub-input Scale How we measure it
Native games count Crash, Mines, Plinko, Dice, Limbo, Roulette built in-house with seed mechanics surfaced. Each = +0.6 up to 6.
UI verifier flag Working in-cashier "verify this seed" widget that hashes the client/server seed pair on demand. Footer claim without UI = 0.
WORKED EXAMPLE

Worked example, how we scored Bitstarz at 9.6

We picked Bitstarz for the worked example because the 9.6 sits at the boundary between Elite and Excellent, and the gap to a 9.7 (Cinoslots) is the cleanest illustration of what a 0.1 difference actually represents under our seven-criterion system. Sample sizes referenced below are the same as the ones published on the Bitstarz review page; raw figures come from our internal audit spreadsheet, snapshot for v3.3 publication. The published Trust Index is the weighted sum rounded to one decimal, every operator's review page reconciles back to the same arithmetic.

Criterion Weight Raw (0–10) Weighted Why
Regulatory Depth 22% 9.7 2.134 Curaçao OGL/2024/165/0185 (LOK direct licence) · Gareton B.V. published in registry · 12-year clean licence history (post-LOK migration from prior 8048/JAZ sub-licence) with no suspensions.
Withdrawal Reliability 20% 9.8 1.960 11-second Lightning median across 220 timed withdrawals; 99.7% first-pass success; manual-review rate under 2%.
Game Library 16% 9.9 1.584 2,400+ titles; per-title RTPs surfaced inside each game card; full Evolution + Pragmatic Live + Hacksaw stack.
Bonus Fairness 14% 9.4 1.316 40× bonus-only base, clearable inside the 14-day expiry; tighter than Cloudbet's points system but EV-positive.
Crypto Variety 12% 9.5 1.140 10 networks on deposit · 9 on withdrawal (LTC is deposit-only, that's the only sub-score drag).
Dispute Handling 10% 9.8 0.980 0.7 complaints per 1k accounts; 92% AskGamblers resolution rate over 24 months; named compliance officer.
Provably Fair Coverage 6% 8.0 0.480 Hash-based RNG-audit certificates on slots, but no native Dice/Crash/Mines stack, pulls the only criterion below 9.0.
Weighted sum 100% 9.594 Σ (weight × raw ÷ 100) across the seven criteria.
Published Trust Index — 9.6 Excellent band, recommended; Provably-Fair shortfall is the only sub-9.0 criterion.
vs. Cinoslots (9.7) · +0.1

Cinoslots beats Bitstarz on Bonus Fairness (35× bonus-only vs 40×) and on Provably Fair (8 native house games with a working seed-verifier vs Bitstarz' RNG-audit only). Bitstarz wins on track record (12 years vs 18 months) and on the licence-history sub-input. Net weighted delta: roughly +0.1 in Cinoslots' favour.

vs. Cloudbet (9.6) · ±0.0

Cloudbet's points-based bonus system scores higher on Bonus Fairness; Bitstarz beats Cloudbet on Game Library breadth (2,400 vs 5,800 titles is in Cloudbet's favour numerically, but Bitstarz publishes per-title RTPs deeper, which the rubric weights higher). The two operators net out at the same published Trust Index, different paths to the same 9.6.

Sample sizes (Bitstarz)
  • Timed withdrawals220
  • Lightning subset180
  • Gameplay hours540
  • Support tickets14
SAMPLE SIZES & PROTOCOL

How the samples are built, per-operator and per-network

A "Trust Index after 2,127 timed withdrawals" claim is only meaningful with the sample breakdown attached. Below is the per-operator, per-network distribution; this is the same registry referenced in every per-criterion section above.

Program window
90 days

Feb–May 2026 rolling audit; re-audit at +30 day boundary.

Operators on watchlist
55

Filtered from 187 crypto-accepting brands at protocol step 1.

Operators rejected
47

Full list with rejection reason at /blacklist.

Timed withdrawals
2,127

Sample size N ≥ 96 per operator. Broken out by chain in the worked example.

Settled value
$847,420

USD-equivalent settled across the 90-day window.

Gameplay coverage
4,800 hrs

4,200 audited sessions across slots, live, table.

The seven-step test protocol applied to every operator

  1. 01 License verification. We pull the public registry record for the operator's claimed licence (Curaçao Gaming Authority direct OGL under LOK, Anjouan ALSI, MGA, IOM GSC). Footer-only logos and "powered by" claims never qualify. Any operator that can't surface a verifiable licence number in 90 seconds is rejected at step one and logged with reason LICENSE-UNVERIFIED on /blacklist.
  2. 02 Real-money deposit. We open accounts using our own KYC, no press contact, no comped accounts. We deposit across at least four networks per operator (typically BTC, ETH, SOL, USDT) so we can compare the cashier behaviour across chains rather than across operators.
  3. 03 Cross-game-vertical play. Every operator gets at least 50 hours of logged play split across slots, live dealer, and provably-fair native games. RTP variance is recorded against the published RTP; freeze events, slow loads, and lobby outages are timestamped.
  4. 04 Stopwatched withdrawals. Minimum N ≥ 96 per operator; each withdrawal is timed from the confirm-click in the cashier to the wallet balance update at the receiving address. Network confirmation overhead is recorded separately so we can isolate operator latency from chain latency.
  5. 05 Support stress-test. Three scripted scenarios per operator: a KYC-trigger workflow, a bonus dispute, and an on-chain confirmation lag. Each is opened from a different IP and a different timezone; we time the response, the escalation path, and the final resolution.
  6. 06 Independent scoring. Each criterion is scored independently by three reviewers; the published Trust Index is the median. Reviewer disagreement above ±0.5 on any criterion triggers a senior re-score.
  7. 07 30-day re-audit. Every ranked operator gets a fresh stopwatched-withdrawal pass at the +30-day boundary. Any operator that regresses on Withdrawal Reliability by more than 0.4 is demoted immediately, regardless of legacy ranking.
RE-AUDIT

When each ranked operator was last re-audited

Trust Index scores carry the date of the most recent full re-audit. Operators with a re-audit older than 30 days carry a visible "stale" pill on their review page until the cycle catches up. The table below is the live re-audit log for the top-eight; full historical log is in the editorial spreadsheet (see editorial team page for access requests).

Operator Last full audit Next due Delta vs. prior Notes
Cinoslots18 May 202617 Jun 2026+0.1SPL median improved 0.5→0.4 s; first-pass success climbed to 99.9%.
Bitstarz14 May 202613 Jun 20260.0No regression; Lightning median held at 11 s across 220 samples.
Cloudbet12 May 202611 Jun 2026+0.1Points-system bonus EV recalculated against current loyalty curve.
Gamdom07 May 202606 Jun 2026−0.1Solana SPL median drifted 7→8 s; ranking position unchanged.
Duelbits03 May 202602 Jun 20260.0Stable; Arbitrum-native withdrawal expansion noted for next cycle.
Wild.io29 Apr 202629 May 20260.0TRC-20 USDT remains the dominant rail; no change in protocol.
22bet24 Apr 202624 May 2026staleRe-audit scheduled, current rating carries a "Last audited" pill.
1xbet18 Apr 202618 May 2026staleManual-review rate climbing; demotion warning on the open ticket.
TRANSPARENCY

Transparency principles and rejection policy

These six commitments are the editorial baseline. Each one is operational rather than aspirational; every claim below corresponds to a process step that produces an artefact (a logged withdrawal, a published blacklist entry, a methodology version stamp).

  • Zero paid placements

    No operator can pay its way to a higher rank. Affiliate commissions are flagged on the operator review and excluded from the Trust Index calculation; commercial relationships never alter sub-scores.

  • 30 days re-audit cycle

    Withdrawal-speed data for every ranked operator is refreshed inside the 30 days window. Bonus terms are verified weekly. Demotions are immediate when criteria regress.

  • Public blacklist

    Every operator rejected by the methodology is listed at /blacklist with the rejection reason. 47 brands are on the current list.

  • Editor's-note labelling

    Policy changes at an operator (T&C revision, bonus structure switch, KYC threshold move) are annotated on the review page with the change date and the criterion affected.

  • Editorial-partner disclosure

    Where we hold a non-arm's-length relationship (branded short-link, priority placement) a disclosure block renders above the review hero. The dedicated page for the only such operator is why Cinoslots ranks #1.

  • Entity-level attribution

    Editorial responsibility is held by CryptoCasinoi Editorial Team (CryptoCasinoi Media Ltd.). Individual bylines are deliberately not published while the contributor-verification programme is rolled out. Full attribution policy.

CHANGELOG

Methodology changelog

Trust Index scores carry the methodology version they were calculated under. When we publish a version bump we keep the older formulae addressable so a rating dated "audited under v3.2" remains traceable to the exact rubric it was scored against.

  1. v3.3
    23 May 2026
    Audit-stats registry split into a single source-of-truth config; per-criterion formulae published in full; rubric bands disclosed.
  2. v3.2
    22 Feb 2026
    Bonus-Fairness sub-weights re-balanced (max-bet rule now penalised; game-weighting penalty introduced).
  3. v3.1
    04 Feb 2026
    Stablecoin-variety weighting clarified, chains count separately on deposit and on withdrawal.
  4. v3.0
    15 Jan 2026
    Initial 2026 cycle release. 100-point Trust Index, seven weighted criteria, 90-day audit window.
  5. v2.4
    12 Sep 2025
    Final 2025 cycle. Archived for reference; ratings emitted with v2.4 hash carry a "Last audited" footer pill.

The current methodology is v3.3, published . Every rating, every category ranking, every per-coin landing page on this site was calculated under this version. When an older version was used (visible as the version stamp at the foot of the review), the corresponding changelog entry above explains the rubric in force at the time.

EDITORIAL TEAM

Who runs the audit

CryptoCasinoi is operated by CryptoCasinoi Media Ltd., registered in United Kingdom since 2024. 6 editorial seats, split across reviews, fact-checking, withdrawal-speed audits, and on-chain forensics, produce every page on this site. Cumulative coverage to date is 78 brands across two annual audit cycles. The full role-by-role contributor list is on the editorial team page.

CryptoCasinoi does not currently publish individual author bylines while we transition to a verifiable contributor model during 2026. Editorial work is attributed to CryptoCasinoi Editorial Team, the in-house staff of CryptoCasinoi Media Ltd.; the editorial process (including testing, fact-checking, sign-off and post-publication re-audit) is described in full on the methodology page.

Methodology version v3.3 was last revised . The version stamp, the revision date and the entity name appear at the foot of every operator review, category ranking and per-coin landing page so that readers can verify which version produced the rating they are looking at. For questions about a published score, a re-audit request, or an editorial process clarification, the editorial team can be reached at [email protected]; all correspondence is read and replied to inside 48 hours.

FAQ

Methodology FAQ

Why a 100-point scale rather than five stars?
Five-star systems compress meaningful gaps. The 0.1 difference between a 9.6 and a 9.7 maps to a measurable difference in median withdrawal speed and bonus EV across a 90-day window; a five-star system would round both to a four-star score and lose the signal entirely. The 100-point scale exists to preserve those gaps.
How are the seven weights chosen?
The weights are calibrated against the failure modes we see most often in our complaint inbox. Regulatory Depth and Withdrawal Reliability together carry 42% because licensing and payouts are the two failures that cost players real money; everything else is consequential but secondary. Weights are reviewed at every methodology version bump.
How is sample size N ≥ 96 chosen for withdrawals?
Ninety-six is the minimum sample at which median withdrawal time stabilises against operator-side variance in our test data. Below 96 the 90th-percentile is too noisy; above 200 the marginal information per additional sample becomes negligible. Ninety-six is the floor; we publish higher sample sizes when we have them.
What happens to operators that change between audits?
Material changes (new licence, bonus structure switch, KYC threshold move) trigger an out-of-cycle re-audit. The change is annotated on the operator review with the date and the criterion affected. If the change degrades a criterion by more than 0.4, the operator is demoted before the next scheduled cycle.
How do you handle operators you can't deposit-test?
We don't rank them. If we can't open an account using our own KYC and our own bankroll, the operator never appears on the published list. Geographic blocks (operators that exclude our test jurisdictions) are noted on the watch-list rather than scored; we revisit when the block is lifted.
Can a complaint move an operator on the ranking?
An individual complaint won't, but the complaint density and resolution rate (Dispute Handling, 10%) are inputs to the score. A documented uptick (verifiable on AskGamblers and Trustpilot) does move the score and is logged with a date stamp.
What does 'audited under v3.x' actually mean on a review page?
Trust Index scores carry the methodology version they were calculated under. A review dated 'audited under v3.2' was scored against the v3.2 rubric (with the v3.2 weights and the v3.2 sub-formulae). When we bump the methodology version, scores re-baseline at the next re-audit boundary; we never silently re-score under a new version.
Crypto casino site briefing

Get every crypto casino site audit before the ranking moves.

One short email when we re-test a crypto casino site, when the Trust Index drops on a brand, or when a regulator publishes a public-disclosure update worth knowing. Zero ad-copy. Zero affiliate spam. Unsubscribe in one click.

  • Weekly recap: new disputes, withdrawal-time regressions, fresh license updates.
  • Bonus alerts re-priced by our wagering-math model, never the headline number.
  • The raw spreadsheet, when we publish a methodology revision (E-E-A-T receipts).
What would you like to hear about?

No advertising. No spam. Two emails per month, maximum. GDPR/CCPA compliant — we ask consent every time and delete on request within 48 hours.