Accounts and funding
Every casino in the ledger is tested on a real player account created for the purpose, funded with the minimum needed for a normal session in each coin under test. We play like ordinary customers — a spread of slots and quick games, no bonus on ledger accounts unless the entry is explicitly about bonus terms. Where a KYC check was triggered, the entry says so; in this season, none was.
Accounts are not anonymous to the casinos we test; that is not the point of the exercise. The point is that our withdrawals are ordinary in size ($40–$400), ordinary in frequency (one to three per casino per week), and ordinary in every visible way, because a fee structure that only misbehaves on ordinary players is the only kind worth catching.
Baselines and the reference exit
Before any casino withdrawal is logged, we price the lane itself: three transfers between wallets we control, at ordinary priority, in each of the five coins, in the same week as the casino tests. The median of those three is the network baseline — BTC ≈ $1.80, ETH ≈ $1.20, USDT-TRC20 ≈ $1.00, LTC ≈ $0.03, SOL < $0.01 this season — and every operator “network fee” is compared against it.
The reference exit is $200 equivalent, requested in every withdrawal lane the cashier offers. Operator fee is the pre-confirmation quote's deviation from the baseline plus any named operator charge. A quote matching the baseline to the cent reads as $0. The receipt — quote, charge, arrival — is screenshotted with the date visible.
What the fee score weighs
The score is a cost-weighted average across lanes, not a vibe. Operator fees on the reference exit (60%): what the casino adds, lane by lane, in dollars. Floors (20%): the lowest minimum across supported lanes, scored logarithmically, because the difference between $10 and $20 matters more than between $70 and $80. Structure (20%): meters, unpublished limits, coin-denominated floors that float, and any pattern of padded network quotes — capped so a clean receipt sheet with ugly terms cannot outrank a slightly pricier cashier that is honest everywhere.
Speed is deliberately not scored. A cheap-but-slow cashier and a fast-but-padded one are different products, and mixing the axes produces meaningless averages; where a casino's pace was notable in our logs, the review says so in words, and readers who care about time should consult a stopwatch site. FeeScout counts cents.
Independence and limits of the method
The ledger is editorially independent and affiliate-funded — placements noted in the footer never move a number, a class or a rank; the formula is published above precisely so that claim is checkable. Figures are a photograph of Ledger 26-D: fee tables, minimums and limits are retuned by operators without notice, and our receipts are evidence of behaviour in a window, not a contract for eternity.
Known limits: we test five major lanes, not the long coin tail; our sizes stop at $400, so caps above that range are read from published tables rather than lived; and dormancy clauses, by nature, need seasons of idleness to observe. Where a claim in a review could age, it carries the season tag. Re-walks happen when terms change materially or a reader receipt contradicts ours — and contradictions, when verified, get printed.