Methodology v1.0 · last updated 2026-04-26
How we score operators
Every operator on Ash Reserve is scored against six weighted dimensions. The rubric is identical for every review, weights are public, the evidence is community-submitted and moderator-verified, and every score change is logged on the changelog.
Where the evidence comes from
Scores are derived from community-submitted evidence — withdrawal proofs, payout-time logs, fairness-verifier reproductions, license-register lookups, item-value spot checks, complaint clusters. Moderators verify each submission for primary-source attribution; the rubric does the scoring math. Anyone can contribute.
Calibration note
Reserve Scores for mystery-box operators are calibrated to this vertical. The entire vertical's licensing posture is weak relative to mainstream casinos: a 7/10 on License & Jurisdiction here is materially less reassuring than a 7/10 for a UKGC-licensed sportsbook. We don't grade against a hypothetical perfect operator that doesn't exist.
Provably-Fair Quality
Weight: 25%Whether a user can independently verify each roll wasn't manipulated, end-to-end, including the item-mapping leg.
Verification: Operator fairness page + manual verification on a real bet using the operator's verifier tool.
| Sub-criterion | Points |
|---|---|
Pre-commit hash visible before roll Hashed server seed shown and recordable before the player commits. | 2 |
Strong, standard primitive HMAC-SHA256 or SHA-512, not a custom or undocumented hash. | 1 |
Independent verifier provided Runnable code (browser or Node), not just prose. | 2 |
Seed rotation policy + retrievability Documented rotation with old seeds queryable post-rotation. | 2 |
Item-mapping reproducibility Per-box weight tables + reduction function published; user can re-derive the winning item. | 2 |
Third-party audit / open-source verifier iTechLabs, eCOGRA, GLI, or a public open-source verifier repo. | 1 |
| Total | 10 |
Odds & EV Transparency
Weight: 20%Whether a player can compute expected value before opening a box, and whether the published numbers actually add up.
Verification: Manual EV calculation on at least five cases per operator using disclosed odds and item values.
| Sub-criterion | Points |
|---|---|
Per-item probabilities published Numeric odds shown for each item, not vague rarity tiers. | 3 |
Probabilities verifiable via PF Disclosed odds match the PF integer ranges used by the mapper. | 2 |
EV calculable from odds + item values Together with item values, a player can compute house edge. | 3 |
Machine-readable weight tables Per-box weights exposed as JSON or equivalent, not just rendered HTML. | 2 |
| Total | 10 |
Withdrawal & Redemption
Weight: 15%Can players actually get their winnings off the platform — quickly, predictably, and without arbitrary holds?
Verification: Documented payout policy + Trustpilot + Reddit complaint patterns + community-sourced payout-time evidence.
| Sub-criterion | Points |
|---|---|
Cash-out option available Non-item redemption (crypto/fiat) — not item-only. | 2 |
Observed payout time Median payout time documented; sub-24h preferred. | 3 |
Withdrawal-ban incidents (penalty cap) Up to -3 deducted from this dimension if credible withdrawal-ban incidents are documented. | 0 |
Limits + fees disclosed Min/max withdrawals and fee schedule publicly stated. | 2 |
KYC trigger transparency When KYC is required and what it asks for is published in advance, not invoked at withdrawal. | 3 |
| Total | 10 |
Item Value Integrity
Weight: 15%Whether the item values shown on the site match real-world market prices, and how punitive the cash-out ratio is.
Verification: Spot-check ten items per operator against Steam Market (skins) or retail comps (physical goods).
| Sub-criterion | Points |
|---|---|
Displayed values match market Listed item values within 10% of current market price. | 3 |
Markup % on cash-out Cash-out ratio of redeemable item value vs displayed value (≥80% is solid). | 3 |
Bait-item ratio Share of items priced below 25% of box price; lower is better. | 2 |
Sell-back rate transparency Sell-back ratio published per item and consistent across the catalog. | 2 |
| Total | 10 |
License & Jurisdiction
Weight: 15%Calibrated to the mystery-box vertical: the entire vertical's licensing posture is weak relative to mainstream casinos. A 7/10 here is materially less reassuring than 7/10 for a UKGC sportsbook.
Verification: Direct regulator-register lookup + ToS parse + VPN spot-check for geo-blocking.
| Sub-criterion | Points |
|---|---|
Credible regulator named Curaçao GCB (post-LOK 2024) = 2, Anjouan = 1, Costa Rica or none = 0. | 2 |
License number verifiable on regulator's public register Linked, working register entry with matching legal entity. | 2 |
Restricted-territory list published in T&Cs Explicit list, not buried 'consult your local laws'. | 1 |
Geo-blocking enforced for US (and UK if no UKGC) Verifiable via VPN spot-check or terms. | 2 |
KYC + age verification required pre-withdrawal Documented and consistently enforced. | 2 |
Independent dispute resolution / ADR named Required under Curaçao LOK; bonus signal elsewhere. | 1 |
| Total | 10 |
Operator Conduct
Weight: 10%How the operator behaves with players over time — community sentiment, scandal transparency, data handling, responsible-gambling tools.
Verification: Trustpilot + Reddit + breach databases + ToS responsible-gambling section.
| Sub-criterion | Points |
|---|---|
Trustpilot signal Volume-weighted Trustpilot rating, with scoring calibrated to vertical norms. | 3 |
Reddit / forum complaint patterns Recurring themes (withdrawal locks, rigged-feeling pulls, account bans) penalized. | 2 |
Responsible-gambling tools Deposit limits, self-exclusion, cooldowns documented and accessible. | 2 |
Scandal/breach transparency If the operator has had a scandal (e.g. Drakemall, MysteryBrand-style), how it was handled. | 2 |
Data-handling history Past breaches disclosed; current data-protection posture documented. | 1 |
| Total | 10 |