How Jackpot Sounds Rates Online Casinos: Our Review Methodology

• Jackpot Sounds Casino Rating Methodology — Version 1
• Effective: June 2026
• Primary jurisdiction: Michigan (MGCB), West Virginia (West Virginia Lottery Commission)

A review without a documented method is an opinion. A review with a documented method is a standard.

This methodology defines how we evaluate licensed online casinos and how we produce both our Expert Score and our casino rankings.

Our reviews are built on two types of evidence:

• Structured data extracted directly from official casino documents, including Terms & Conditions, bonus terms, payment policies, and licensing disclosures.
• Editorial testing of the player experience, including banking, game selection, mobile usability, customer support, responsible gambling tools, and overall platform quality.

Every operator is assessed using the same framework and benchmarked against other licensed operators in the same jurisdiction. We do not compare casinos against a global standard. We compare them against the market they actually compete in.

The methodology is versioned and publicly maintained. When we add new evaluation criteria, modify scoring rules, or expand our data collection process, the version number changes and the update is recorded in the changelog.

How we make money, and what it doesn’t change

Jackpot Sounds earns affiliate commissions when readers sign up with some of the operators we cover. That commercial relationship does not buy a better score, a higher ranking, or a place in any “Best For” list. We score and rank every licensed operator in a market against the same framework whether or not we earn from it, we include operators regardless of any partnership, and we never reorder results by revenue. The scoring rules in this document are the only thing that moves a rating. Full details of our commercial relationships are set out in our Affiliate Disclosure.

Who we review: licensing is the gate, not a score

Before any scoring begins, an operator must clear a single eligibility gate, which we call Pillar 0. It must hold a valid licence from the regulator of the jurisdiction under review, and that licence must be verified against the regulator’s own public register — in Michigan, the Michigan Gaming Control Board (MGCB); in every other market, that market’s licensing authority.

Pillar 0 is pass or fail, and it is deliberately not part of the Expert Score. Because every operator we review has already cleared it, scoring “holds a licence” would hand the same points to everyone and do nothing but compress the scale — it would tell a reader nothing about how two licensed operators actually differ. A local licence is the price of admission to our reviews, not a point of differentiation between operators. What an operator does with that licence — its ownership transparency, sanctions record, time in market, and data-protection posture — is still assessed and scored, under the Trust & Safety pillar.

We never review, rank, or recommend offshore casinos, unlicensed operators, or brands not authorised to offer online gambling in the market covered by the review. Inclusion is not permanent: if an operator loses its licence, has it suspended, withdraws from the market, or becomes subject to significant regulatory enforcement, we re-evaluate and update the relevant review and ranking pages. Where it is useful for transparency, we may keep an informational review page that clearly states the operator is no longer licensed or no longer eligible for our rankings.

What this methodology covers

The framework consists of two data-driven audit pillars and six editorial evaluation pillars.

The data-driven pillars focus on information that can be directly extracted and verified from official casino documents:

• Terms & Conditions Quality Audit
• Bonus Intelligence Audit

The editorial pillars focus on the actual player experience:

• Trust & Safety & Operator Background
• Banking & Withdrawals
• Games & Software Providers
• User Experience & Mobile
• Customer Support & Responsible Gambling
• Product Development & Platform Freshness

Alongside our Expert Score, we are developing an Audience Score based on player feedback, complaint patterns, withdrawal experiences, and dispute outcomes. The Audience Score is currently in development and is not yet incorporated into operator rankings.

When published, the Audience Score will always be reported separately from the Expert Score. The two scores will never be blended into a single rating.

Two scores, never blended

Current status: The Expert Score is live and used throughout our reviews and rankings. The Audience Score framework is currently being developed and validated before public release.

Every operator we review carries two separate headline numbers. We never average one into the other, because they answer different questions.

• The Audience Score is what real players report. It is built from verified player feedback through regulated channels — recurring complaint patterns, withdrawal experience, dispute outcomes, long-term satisfaction. It is the players' verdict, not ours.
• The Expert Score is our own assessment, built across the evaluation pillars below. Its backbone is objective, source-quoted data extracted directly from each operator’s official documents and benchmarked against every licensed operator in the same jurisdiction.

A casino can have a high Expert Score and a mediocre Audience Score — clean terms on paper, friction in practice — and that gap is itself one of the most useful things a review can show. We surface it; we don’t hide it inside an average.

How we know what we know

We go to the source. Every data point is extracted directly from the operator’s own official pages. We never use affiliate summaries, bonus-comparison sites, or third-party databases.

We record what the document says, not what we think it means. Every extracted field carries a verbatim quote from the source. The quote is the evidence.

Every field has one of four states. This is how we stay honest about gaps:

• Disclosed — the operator states it; we record the value and the quote.
• Confirmed-absent (N/A) — we checked and the term genuinely does not exist (for example, an offer with no promo code). This is not a mark against the operator.
• Deferred — the document we are reading explicitly assigns this term to another document (for example, the master T&C states that a bonus’s wagering rules live in the offer’s own terms). We record the information from that source instead. Deferred is not a transparency flag — the operator is being clear about where the rule lives.
• Unclear — the term should exist but is not located anywhere we can find, or the page carries stale or contradictory text. This is a transparency flag.

We never infer, assume, or fill a gap. In the current extracts, Deferred cases are encoded as Unclear with an explanatory note (for example, “deferred to individual Offer Terms”); we read that note and score them as Deferred, not as a flag. “Confirmed-absent”, “Deferred”, and “Unclear” are different findings and are scored differently.

We benchmark against the field. A 15× wagering requirement means nothing in isolation. Against a field of MGCB-licensed operators it means a lot. Every measurable data point is compared against all licensed operators in that jurisdiction — field minimum, field median, field maximum, and where this operator sits.

We date everything, and we keep the raw capture. We record the extraction date on every field and the source document’s own “Last Updated” date where present. For each extraction we archive the raw page capture together with the fetch date and method, so any finding can be re-checked against exactly what we saw.

How the Expert Score is built

The evidence spine vs. editorial judgment

Two pillars — T&C Quality Audit and Bonus Intelligence — are structured extraction against a fixed parameter set, with a verbatim source quote required for every field. These are reproducible: a second analyst reading the same documents should record the same data. This is our “lab data.”

The remaining pillars involve editorial judgment — navigating the site, testing support, assessing the game library. We score these on a standardized 1–5 rubric per pillar so operators stay comparable, and we label them clearly as editorial assessment rather than extracted data.

The benchmark method (how a number becomes a score)

For every measurable parameter, we report the operator’s value against the licensed field, in the same format throughout the site:

Wagering requirement (deposit match): 15×

Compared to all MGCB-licensed operators, the field ranges from 1× to 15×, with a median of approximately 7×. This operator sits at the strict end of the field.

A single number is never shown without the field around it. Extracted data becomes a pillar sub-score by its position within the competitive field, not against an arbitrary absolute scale.

Here is the exact rule that turns a field position into a band. It is identical for every benchmarked parameter on the site.

Step 1 — Normalize. We score the disclosed value in its normalized form — a number with its unit, or a defined category — never the raw quoted text.

Step 2 — Orient. Every parameter carries a fixed direction in our parameter dictionary: whether a higher or a lower value is better for the player. Lower wagering is better; a shorter withdrawal time is better; a higher maximum cashout is better.

Step 3 — Build the field. The benchmark distribution for a parameter contains only operators whose value is Disclosed in that jurisdiction. A Confirmed-absent (N/A) value does not enter the distribution and is scored by the parameter’s own rule. An Unclear value does not enter the distribution either: the operator is scored at Band 1 on that parameter and carries a Transparency flag, because concealing a term a player needs is itself a failure, not a neutral gap. A Deferred value is resolved from the second document first, then scored normally.

Step 4 — Place a numeric parameter into a band. We locate the value against the field’s quintile thresholds, oriented by Step 2: best 20% of the field → Band 5 · 60th–80th percentile → Band 4 · the middle of the field, 40th–60th → Band 3 · 20th–40th percentile → Band 2 · worst 20% → Band 1. Banding is by value thresholds, so operators with identical values always share a band. If every operator holds the same value — the parameter differentiates no one — they all receive Band 3, so a flat field can neither lift nor sink a score.

Step 5 — Place a categorical parameter into a band. Non-numeric parameters — for example the self-exclusion mechanism or the bonus balance-return method — are scored against a short published rubric that maps each defined option to a band from 1 to 5.

Step 6 — Handle small fields. When fewer than ten operators disclose a parameter, quintiles are unstable, so we fall back to coarser anchors — field-best → 5, better than the median → 4, at the median → 3, worse than the median → 2, field-worst → 1 — and mark the affected sub-scores as provisional until the field is large enough. Several of our markets are small at launch, and this keeps early scores honest.

Worked example. The 15× wagering requirement above is the maximum of a field that runs from 1× to 15× with a median near 7×, and lower is better — so it falls in the worst quintile and scores Band 1 / Poor. By construction, an operator sitting at the field median would score Band 3 / Adequate.

These relative bands are the base layer. On top of them sits a second layer of absolute guardrails — the red lines defined below — which cap a score when a value or a behaviour crosses a hard player-protection or legal threshold, regardless of where the field sits.

Core vs. periphery

Not every field is directly comparable, and that is by design. Some parameters appear consistently across almost all operators, while others exist only in specific jurisdictions, products, or operator policies.

Fields that are used in rankings, benchmarks, field medians, or “Best For” recommendations are treated as core fields and must be recorded in a consistent format across every operator. Fields that appear only occasionally are recorded and reported, but are not used for direct benchmarking.

• T&C core (present in ≥85% of operators): withdrawal processing time, deposit limits, voiding-of-winnings conditions, KYC trigger, dormancy threshold, dormancy outcome, self-exclusion mechanism, account-termination grounds, license jurisdiction, license authority, operator legal entity, complaint procedure, regulatory escalation path.
• Bonus core (present in 100% of welcome offers): offer type & value, minimum deposit, promo code, opt-in, credit method, offer expiry, wagering completion window, wagering requirement, eligible games, excluded games, game contribution rates, promotion period, and maximum bet. Maximum bet is core: it is almost always stated, and where an offer omits it we record the field as Unclear and raise a Transparency flag, because a missing bet limit is a term the player needs before depositing. Maximum cashout (the cap on winnings) is handled separately. Most deposit-match offers carry no cap and simply do not mention one, so a missing cap is read as “no win cap” — a neutral-to-positive signal, not a hidden term. We score maximum cashout against an operator only when a cap is actually disclosed, where a lower cap is worse; we treat its absence as Unclear and flag it only for offer types where a cap is the norm, such as no-deposit and free-spin bonuses.
• Periphery (floats; recorded and shown, not benchmarked): chargeback fees, reverse-withdrawal clauses, parent-company identity, master-T&C bonus mechanics, the verbatim text of abuse definitions, named excluded titles, registered addresses, and any operator-specific clause.

Core fields carry a normalized companion. The verbatim value and quote stay as the evidence; a normalized value (a number with its unit, or an enum) sits beside them so the field can be ranged and ranked. Without it, a value like “approximately three business days” is perfect as a quote and useless as a median.

Product types

Welcome offers are benchmarked only against the same product type. A poker room releases its bonus against rake and tournament fees, not as a deposit-wagering multiplier; ranging a poker “multiplier” against a slots bonus compares unlike things. Each offer is tagged casino or poker, and the field benchmark is computed within type.

Transparency flags

Alongside parameter values, we track a countable, comparable signal that feeds the T&C and Bonus pillars: the number of Transparency flags against an operator. A flag is raised for any parameter recorded as Unclear, any stale or contradictory published text (for example, a live promotion page still showing “valid through December 31, 2022”), an undated T&C document, or placeholder contact details left in a published T&C. A Deferred field — where the document clearly points to another document we also read — does not raise a flag. More flags means lower transparency, and the count is comparable across the field.

Scoring bands

Each pillar sub-score maps to a band: 5 / Excellent · 4 / Good · 3 / Adequate · 2 / Weak · 1 / Poor. Bands are defined per pillar and published, so a rating is never a black box.

What each band means, relative to the field: Band 5 — among the best disclosed terms in the market · Band 4 — better than the market median · Band 3 — at or near the median · Band 2 — worse than the median · Band 1 — among the worst disclosed terms, or the term is hidden (Unclear).

A pillar’s sub-score is the weighted average of its core-parameter bands. The Expert Score is the weighted sum of the eight pillar sub-scores, using the published weights. Periphery fields are shown on the operator’s page but do not move the sub-score.

One set of bands, many rankings. The same parameter bands feed every view of an operator. The headline Expert Score applies the default pillar weights; each “Best For” ranking re-weights the identical bands toward what that audience cares about — low-wagering hunters up-weight the wagering and completion-window bands, high rollers up-weight maximum cashout, withdrawal limits and loyalty, and so on. Bands are computed once; only the weights change. That is why one operator can rank differently for different players with no inconsistency in the underlying data.

Red lines (absolute guardrails)

The benchmark bands rank an operator against its market. Red lines do the opposite: they are absolute limits that cap a score regardless of how weak the rest of the field is, so that “better than awful competitors” can never be reported as “good.”

A red line acts in one of two ways. A hard cap limits the operator’s overall Expert Score to 2.0 out of 5 and attaches a visible Caution note that states the reason. A pillar cap limits the relevant pillar’s sub-score to Band 2, so a predatory area cannot be averaged away by strengths elsewhere. Red lines and their thresholds are v1 values and will be tuned as we gather data across markets; every change is recorded in the changelog.

Red lines come from two places: the terms an operator writes, and the way an operator behaves.

A. Terms, read from the documents we already extract. A term that is hidden rather than disclosed is handled by the normal rule — Band 1 plus a Transparency flag — not here.

• Hard cap. A clause that lets the operator reverse a withdrawal the player has already requested.
• Hard cap. Discretionary confiscation of winnings — voiding a balance for vaguely defined “irregular play” or “abuse” with no objective definition.
• Hard cap. Inactivity or dormancy fees that can erase a real-money balance.
• Bonus → Band 2. A disclosed wagering requirement above 50× on the normalized basis.
• Bonus → Band 2. A disclosed maximum cashout lower than the player’s own qualifying deposit. The absence of a cap is not a red line; it is read as “no cap.”
• Banking → Band 2. Withdrawal terms that trap a balance — a payout limit so low that cashing out a typical win would take more than three months, or a fee charged on every withdrawal.

B. Conduct, from regulators and from verified player evidence. These come from outside the documents.

• Hard cap. The operator lets a player gamble from a place where it is not licensed — for example a player physically outside the licensed state who can still wager. This is a legal failure, not a quality one.
• Hard cap. The operator fails to honour an account-closure or self-exclusion request, especially from a player who has disclosed a gambling problem.
• Hard cap. Active regulatory enforcement or a fine against the operator. A full licence loss removes the operator entirely under Pillar 0.
• Banking → Band 2. A verified pattern of using document checks (KYC) to delay or block legitimate withdrawals.

Conduct red lines require evidence, not a single anonymous comment. We act on regulator publications and on verified player reports — a documented, reproducible pattern through channels we can check — and we record the basis for each red line on the operator’s page. A lone, unverifiable review is a prompt to investigate, never a red line on its own.

Pillar weights (v1)

The Expert Score is a weighted sum across the scored pillars. Weights are published openly and revisited each methodology version.

Player Feedback & Market Reputation is reported separately as the Audience Score, never blended into this table. VIP & Loyalty is evaluated within the editorial layer and reported, but carries no standalone weight in v1.

Pillar B: T&C Quality Audit

When a player creates an account, they agree to the operator’s master Terms & Conditions. Almost no player reads them. We do.

The promo page tells you how to get the bonus. The master T&C tells you what happens when things go wrong — when a bonus is voided, an account is closed, a withdrawal is held. We fetch the master T&C directly from the official URL, record the retrieval date and the document’s own stated “Last Updated” date, archive the raw capture, and extract the clauses that most affect players, each with a verbatim quote and one of the four field states.

What we find that players miss

Where the rules actually live. Operators structure their documents differently, and that itself is a finding. E.g.

• BetRivers states a universal default in its master T&C: Bonus Money expires 30 days after issuance. 
• Bet365's master T&C states no universal bonus validity, no master max-bet, and no master exclusion list — each offer’s rules live on that offer’s own page. 

Neither is wrong, but a player reading only the master T&C learns very different amounts from each, which is exactly why our two pipelines are read together: where the T&C defers a bonus rule, the Bonus Intelligence pipeline picks it up.

Dated vs undated documents. Without a date a player cannot know which version governs their account.  E.g.

• Bet365: effective 04/17/2026. 
• BetRivers: last updated June 10, 2025 — both dated. 

By contrast, operators that publish no date, or placeholder contacts (Play Gun Lake MI: “Phone Number TBD”), raise Transparency flags.

Withdrawal holds. BetMGM: withdrawals may not be accepted within 10 days of a credit-card deposit — disclosed only in the master T&C. Neither BetRivers nor Bet365 states any card-deposit hold; both are recorded as Unclear rather than assumed absent.

Dormant account outcomes. BetRivers: 3-year rolling dormancy; funds may be forfeited where permitted by law. Bet365: 3 consecutive years; zero-balance accounts closed fee-free, positive balances the operator tries to return, otherwise unclaimed funds remitted per state law. The threshold clusters at 3 years across Michigan; the outcome detail differs.

Self-exclusion and what happens to your balance. BetRivers spells it out: 1 year / 5 years / Lifetime, and the cash balance (less unconverted bonus money) is mailed as a check. Bet365's master T&C confirms a 72-hour minimum suspension and lists self-exclusion as a tool, but defers the periods and balance handling to its responsible-gaming page — captured as Deferred.

The escalation path — and the fine print on it. Both route disputes to the MGCB. E.g.

• BetRivers gives a Player Support email, phone number, a 3-business-day response commitment, and the MGCB complaint URL. 
• Bet365 routes through a Contact Us link then management, then the MGCB portal, and adds a binding-arbitration clause (AAA rules) with a one-year limitation period — a player-rights term that surfaces only if you read the master T&C.

The parameter groups (aligned to the T&C extracts)

Document identity — Publication / “Last Updated” date · legal entity and registered/principal address (e.g. Hillside (Michigan) LLC, Delaware; Rush Street Interactive MI, LLC) · license, regulator, and license number (s) where stated (Bet365: Internet Gaming License IGSL008630-26-001) · parent company / corporate group, recorded as Unclear where the ultimate parent is not named (as with Bet365).

Withdrawals — Stated processing window (BetRivers: ~3 business days; Bet365: deferred to the Withdrawals page) · withdrawal minimum · withdrawal maximum / daily cap (Bet365: no daily maximum, per-method limits apply) · payment-method or credit-card hold · deposit limits, player-set via responsible-gaming tools vs operator-imposed.

Bonus mechanics in the master T&C — Default bonus validity (BetRivers: 30 days; Bet365: deferred) · universal max bet during wagering (Bet365: deferred) · games excluded from wagering per master T&C · definition of bonus abuse / prohibited activity. These may differ from, supplement, or be deferred entirely to the promo page — the reason both pipelines are read together.

Account risk — Conditions under which winnings are voided or withheld (malfunction, error, prohibited activity) · grounds for account termination or suspension (BetMGM: 10 named grounds; Bet365: full Prohibited Activity list plus regulatory closure) · dormant-account threshold and outcome.

Player protection — KYC trigger point (s) · required verification documents (ID, selfie, proof of address, proof of payment method) · self-exclusion mechanism, available periods, and balance handling on exclusion.

Dispute resolution — Internal complaint procedure with working contact details and stated response time · whether any contact fields remain placeholders · regulatory escalation path (in Michigan, the MGCB, with portal and contact where stated) · binding-arbitration clause and limitation period where present.

Pillar C: Bonus Intelligence

When an operator publishes a welcome offer, we go to the official promotions page and extract every term that determines whether the bonus is actually playable. We do not summarize the marketing copy — we document the mechanics.

The headline — “100% up to $1,000!” — is the least useful piece of information about a bonus. The useful information is everything else: how the offer is credited, how long you have, how many times you must wager it, on which games, and at what contribution rate.

We extract thirteen parameters plus the offer’s dating metadata for every welcome offer (the eleven below, plus maximum bet and maximum cashout noted at the end of this section). Each requires a verbatim quote, and each carries one of the four states (Disclosed / Confirmed-absent / Deferred / Unclear). Multi-component offers — a no-deposit bonus plus a deposit match, for example — are documented component by component, never collapsed into a single summary, because each component typically carries its own credit method, its own clocks, its own wagering requirement, and its own game restrictions.

The thirteen parameters

Promotion period (start/end dates) is logged as dating metadata on every offer. Where a page shows stale or contradictory period text on a live offer, the field is recorded as Unclear and raises a Transparency flag.

The editorial pillars

Evaluated through direct assessment — site navigation, support testing, registry checks, player-feedback analysis — and scored on a standardized 1–5 rubric, clearly labeled as editorial.

• Pillar A — Trust & Safety & Operator Background. Licence validity is settled before scoring as the Pillar 0 eligibility gate; this pillar assesses what the licence does not tell you. Ownership transparency, corporate group and ultimate-parent disclosure, time in market, record of regulatory sanctions, and SSL and stated data-protection measures — the factors that separate two equally licensed operators. Independent RNG/fairness certifications are recorded as informational only — in a locally-licensed market the local license is the trust gatekeeper, so a third-party badge does not raise an operator’s trust score on its own.
• Pillar D — Banking & Withdrawals. Methods including local options, stated limits, fees, stated timelines and verification requirements. Stated withdrawal times are cross-referenced against the T&C pipeline; direct tests are noted and dated.
• Pillar E — Games Library & Software Providers. Slot depth, live casino, table games, video poker, jackpots; provider portfolio quality; what kind of player the operator serves best (basis for “Best For”).
• Pillar F — User Experience & Mobile. Navigation, search/filter, registration flow, accessibility of bonus terms and RG tools; mobile performance, app availability and quality, feature parity, store ratings.
• Pillar G — Customer Support & Responsible Gambling. Channel availability and hours; response speed and resolution quality, tested with a standardized question set; deposit/loss/session limits, reality checks, time-outs, self-exclusion, external support resources. The standardized question set — covering different stages of the player journey, including pre-registration, active play, withdrawal, and account closure — is currently in development; until it is published, support scores are marked provisional.
• Pillar H — Product Development & Platform Freshness. How actively the operator adds games and providers, refreshes promotions, and improves the platform. Assessed at review time and updated each cycle.

Reported separately as the Audience Score — Player Feedback & Market Reputation. Verified feedback from regulated channels; recurring complaint patterns; long-term satisfaction trends. A signal of volume and pattern, not individual reviews.

Jurisdiction & the benchmark field

Our benchmark is only as meaningful as the field it is measured against, so the field is always every operator holding a local license in the jurisdiction under review — no more, no fewer.

• Michigan (live now): all online casinos licensed by the Michigan Gaming Control Board (MGCB). License verification runs against the MGCB’s public record.
• Ready to deploy: West Virginia — benchmarked against all operators licensed by the West Virginia Lottery and evaluated using the same methodology framework.

We do not mix jurisdictions in a single benchmark, and we do not score operators against an absolute global scale. The field decides.

Some core fields are determined by regulation rather than by individual operators. In Michigan, for example, certain requirements — such as dormant-account rules and the escalation path through the Michigan Gaming Control Board (MGCB) — are largely consistent across licensed operators.

We still record and verify these fields because they confirm regulatory compliance. However, they rarely differentiate one operator from another. Meaningful differences are typically found in operator-controlled areas such as withdrawal policies, payment-method restrictions, fee structures, ownership transparency, customer-service procedures, and bonus terms. Our weighting reflects this distinction.

Keeping reviews current

Reviews are not static. Each carries an extraction date. We monitor licensing changes, bonus-offer updates, T&C revisions, payment-policy changes, regulatory actions, and shifts in player feedback. Material changes trigger a re-evaluation. Expired offers are flagged until the current offer is extracted and verified.