Published vs personalised rakeback: what can you calculate?

Published rates cover different amounts of the calculation

The calculation base is identifiable for 11 of the 15 programmes. Gamdom and Roobet describe personalised models. Winna and Rainbet’s Instant programme describe wagering activity without establishing a usable effective denominator and numeric rate.

These five categories are mutually exclusive and add to 15. The eight in the first row are not eight offers available on identical terms. Duel’s published slot example supports the nominal algebra while leaving ordinary account entitlement incomplete. Housebets’ 50% requires the selected reward mode. Yeet’s ordinary 5% and invitation-only 10% have different access conditions.

Payment information is a separate finding. Four programmes describe a split, five state a balance destination without establishing complete allocation, and six leave complete allocation unresolved. SpinCrypto is the only row with a specified earning cap: USD50 equivalent per bet. The other 14 have no established earning-cap rule in the material examined; that does not establish unlimited earnings.

The 15-programme disclosure matrix

Rows are alphabetical. “Described” means the named part is available in the programme information. “Not established” keeps a missing input open. Some conditions can be known even where the complete field remains unresolved.

No programmes match both choices. Change either filter to see another part of the sample.

Why the eight nominal models still leave questions

The rate must belong to the account and programme

500 Casino describes accrual from Level 0 and publishes a general 5% to 20% share of house edge. It does not expressly map 5% to Level 0 or 20% to an exact tier. The starting percentage cannot be filled in from the bottom of that range. Qzino publishes a net-loss base and rank ceilings, but no numeric rate for Newcomer levels 0 to 2. A missing entry rate remains missing even though higher-rank figures are available.

MetaWin identifies theoretical revenue and a tier multiplier for its Instant Bonus, covering the preceding 30 minutes. The public material does not provide a complete numeric multiplier table. Its below-2%-edge rule concerns Rank Up Points; that rule alone cannot supply the Instant Bonus’s game-contribution weights.

The same percentage needs correctly weighted eligible play

An edge-based formula needs each eligible stake and the relevant game configuration’s house edge. A single assumed RTP for every game cannot reproduce a mixed session. Yeet expressly allows different game or promotion weights, while most other rows establish product scope without a complete weight table.

Duel’s 50% explanation is about slot house edge, so it cannot be extended to its Originals, live games or zero-edge products. Stake’s 3.5% casino formula is separate from its poker-rake and sportsbook calculations. Rainbet’s Calendar Rewards model cannot be copied into its Instant rakeback row.

A cap has a unit and a place in the calculation

SpinCrypto caps each real-money bet’s Instant accrual at USD50 equivalent. Calculate the eligible per-bet amount, apply that bet’s cap, then add the accruals. Capping a whole session once would be a different rule. Its total accrued claim is explicitly uncapped, while general daily withdrawal limits still apply. Native-coin amounts use the operator’s internal exchange rate; a market price would substitute a different input.

The adjustment field also covers deductions and interactions, including any carry rule. Qzino expressly deducts winnings from losses in the accrual period, but a complete adjustment schedule is not established. An unmentioned carry or deduction cannot be assumed to be zero.

Earned, credited and withdrawable are separate amounts

Roobet’s 50/50 balance-and-Vault split allocates an already determined reward. It does not reveal the rate that earned that reward. Rollbit likewise sends half of a base-rakeback claim to balance and half to a three-day calendar. A known nominal result therefore does not mean the entire amount arrives immediately.

Balance wording also leaves different questions. Stake and Shuffle describe cash without a separately established 0x reward-wagering rule. Housebets explicitly specifies 0x reward wagering, yet separate 1x deposit playthrough can restrict withdrawal. Its selected 50% mode halves XP and stops new Hubie cashback, weekly/monthly bonuses, VIP reloads and discretionary Hubie rewards; previously earned rewards retain their schedules.

Gamdom and Roobet describe personalised factors without a complete public algorithm. That limits independent reproduction of the amount. It does not establish unfair treatment or a lower payment. For an existing account whose reward varies, the next task is understanding why a personalised rakeback amount changes.

What this sample can establish

One selected programme per crypto brand

The sample contains one selected main casino or slot programme for each of the catalogue’s 15 crypto brands. “Main” identifies the selected record, not a ranking. Qzino’s selected record is cashback; its supplementary rakeback is excluded. Extra weekly rewards, sports-only records, limited boosts and other programme contexts are not added as independent casinos.

This is a purposeful catalogue sample. It is not a random survey, a European market average or a worldwide disclosure rate. The wider source catalogue has 66 programme and product contexts; a separate casino/slot comparison includes 49. Neither number is the denominator for the findings here.

Fields are checked separately

The classification checks the programme’s stated base, a numeric rate tied to its described scope, contributions, adjustments and carry, cap scope, allocation, eligibility, minimums, rounding and final payment. A conditional nominal formula is counted when the base and a usable scoped rate can be identified, including algebra reconstructed from an explicit example. It is not counted as complete payment reproduction.

Known, partial and unknown are kept distinct field by field. An explicit personalised model remains a personalised model; it is not treated as a concealed fixed rate. Conflicting terms require a visible qualification rather than an invented resolution. No row in this nominal-model classification needs a separate conflicting category, which does not mean every term of every programme agrees.

Descriptions of terms, with limits

The findings come from programme descriptions and terms, with the original material retained for checking. They are not deposit, bet, claim or withdrawal tests. They describe the information available for these selected records, including its gaps, rather than the fairness or speed of payments. Our comparison methodology explains the treatment of sources and unresolved conditions.

For the arithmetic itself, continue to calculating rakeback from the right base. That guide explains how to use known inputs; it cannot provide an unpublished account multiplier.

Read the conditions behind three different models

These examples follow from the matrix: a published casino rate, a personalised Instant reward and a personalised reward with a documented split. They are selected to explain the differences, without declaring a highest-paying programme.

Find the other studies in the casino rewards research library.

Questions about published reward formulas