Skip to content

Cashback and rebates

Cashback does not usually return a percentage of everything wagered. It returns a percentage of a loss defined by the promotion. That definition can exclude games, markets, bonuses, fees, voids, or activity outside a settlement window. The word “cashback” also does not guarantee cash.

The calculation starts with the offer’s loss base. Only then should the rate, cap, and form of payment be applied.

Gross rebate = eligible loss x rebate rate
Credited rebate = lesser of gross rebate and rebate cap
Cash-equivalent rebate = credited rebate x conversion rate

If the rebate is withdrawable cash, conversion rate is 100%. If it is a restricted credit, conversion must be modeled. The headline amount is the maximum advertised rebate. Expected value is the probability-weighted change the rebate makes to the full activity. Cash-equivalent value is the modeled withdrawable value of the credit received. Actual result is the settled payment and the gambling outcome that produced it.

Net loss usually means qualifying stakes minus qualifying returns over a stated period, but definitions vary. A casino may net all included games together. A sportsbook may net settled stakes and returns, perhaps by sport or promotion. Deposits and withdrawals are not automatically gaming wins or losses.

Eligible loss is the portion that remains after the offer’s exclusions. It can differ from the account’s net result.

Hypothetical casino scenario: During one weekly window, a customer stakes $1,200 on eligible slots and receives $1,050 in returns. The customer also stakes $300 on excluded table games and receives $220.

Eligible slot loss = $1,200 - $1,050 = $150
Excluded table loss = $300 - $220 = $80
Account gaming loss = $150 + $80 = $230
Eligible loss for rebate = $150

At a 10% rebate rate, the gross rebate is $15, not $23. The table result remains economically real even though the promotion ignores it.

Wins can offset losses within the netting window. In a hypothetical example, if a customer loses $100 on Monday and wins $90 on Tuesday, a weekly net-loss rebate may use $10, while a daily program could calculate Monday and Tuesday separately. Neither interpretation is universal. Check the start time, end time, time zone, and whether unsettled activity rolls into the next window.

A cap limits the credited rebate. A tier can change the percentage after loss or activity thresholds. Some programs apply one rate to the entire eligible loss after a tier is reached. Others apply marginal rates only to the slice within each tier. These structures produce different results.

Hypothetical tier scenario: A weekly casino rebate pays 5% on the first $100 of eligible loss, 10% on the next $200, and 15% on eligible loss above $300. The total rebate is capped at $60. The tiers are marginal.

For a $500 eligible loss:

First tier rebate = $100 x 5% = $5
Second tier rebate = $200 x 10% = $20
Third tier rebate = $200 x 15% = $30
Gross rebate = $5 + $20 + $30 = $55
Credited rebate = $55

For a $700 eligible loss:

First tier rebate = $5
Second tier rebate = $20
Third tier rebate = $400 x 15% = $60
Gross rebate = $85
Credited rebate after cap = $60

The extra $200 of loss raises the rebate by only $5 after the cap. Losing more to reach a tier is not a saving strategy.

Opt-in can be required before the first qualifying wager, once per settlement window, or before a claim deadline. Examples of exclusions can include bonus-funded wagers, low-margin games, opposite-side wagers, cashed-out bets, voids, pushes, particular markets, or wagers below minimum odds. These are examples only. Use the exact written rules and save the accepted terms. The pre-deposit checklist gives a place to record them.

A cash rebate enters a withdrawable or ordinary cash balance after settlement, subject to normal account review. A bonus-credit rebate carries a separate playthrough or conversion condition. A site-credit rebate may need to be wagered once and may follow stake-returned or stake-not-returned settlement.

Hypothetical credit-form scenario: A customer earns a displayed $40 rebate. Compare three payment forms.

  • Form A is $40 withdrawable cash.
  • Form B is $40 casino bonus credit with an estimated 65% cash conversion after wagering.
  • Form C is a $40 sportsbook bonus bet with stake not returned and an estimated 58% cash conversion.
Form A cash-equivalent value = $40 x 100% = $40
Form B cash-equivalent value = $40 x 65% = $26
Form C cash-equivalent value = $40 x 58% = $23.20

All three can display “$40 cashback.” Their economic values are not equal. Form B also adds casino variance and completion risk. Form C can settle at zero. Wagering requirements and free spins and insured bets explain those conversion stages.

A loss rebate changes only eligible losing outcomes. It does not reduce winning outcomes, and it does not make session results cluster around the new average.

Hypothetical casino scenario: A customer has already planned a $200 slot session. The modeled outcomes are:

  • 25% probability of a $200 loss;
  • 35% probability of an $80 loss;
  • 25% probability of a $40 gain;
  • 15% probability of a $220 gain.

First calculate expected net result without cashback:

Baseline EV
= (0.25 x -$200) + (0.35 x -$80)
+ (0.25 x $40) + (0.15 x $220)
= -$50 - $28 + $10 + $33
= -$35

The casino offers 20% cash back on eligible net loss, capped at $25. Apply the cap to each loss branch:

$200 loss: raw rebate = $40, capped rebate = $25
Revised outcome = -$175
$80 loss: rebate = $16
Revised outcome = -$64
Winning outcomes: no rebate

Now recalculate:

EV with cashback
= (0.25 x -$175) + (0.35 x -$64)
+ (0.25 x $40) + (0.15 x $220)
= -$43.75 - $22.40 + $10 + $33
= -$23.15
Incremental cashback value = -$23.15 - (-$35) = $11.85

Cashback improves expected loss from $35 to $23.15. It does not make the session positive EV. The actual outcome can still be a $175 loss after rebate or a $220 gain. The distance between those outcomes remains $395, so meaningful variance remains.

The rebate rate is not the same as a reduction in house edge. The 20% applies only to net losing sessions and is capped. If it applied without a cap to every negative unit of loss, the transformation would be easier. Real settlement rules usually require outcome-by-outcome modeling.

Hypothetical casino scenario: Use the same outcome distribution, but the 20% rebate is bonus credit with a 60% expected cash conversion, still capped at $25 credited.

$200 loss:
Credited rebate = $25
Cash-equivalent rebate = $25 x 60% = $15
Revised expected outcome for branch = -$185
$80 loss:
Credited rebate = $16
Cash-equivalent rebate = $16 x 60% = $9.60
Revised expected outcome for branch = -$70.40

Full expected result:

EV with restricted-credit rebate
= (0.25 x -$185) + (0.35 x -$70.40)
+ (0.25 x $40) + (0.15 x $220)
= -$46.25 - $24.64 + $10 + $33
= -$27.89
Incremental value = -$27.89 - (-$35) = $7.11

The $11.85 face-value benefit from the cash version falls to $7.11 in cash-equivalent value. The actual result can be worse if the credit expires or reaches zero during conversion.

How cashback changes sportsbook expected loss

Section titled “How cashback changes sportsbook expected loss”

A sportsbook rebate can settle across a bet, a day, or a group of wagers. Model the same unit used by the terms. Do not sum per-bet rebates when the offer nets the full week.

Hypothetical sportsbook scenario: A customer plans two independent $100 cash wagers at decimal odds of 1.91. Each has an estimated 50% win probability. A weekly promotion returns 25% of eligible net loss as withdrawable cash, capped at $30.

Each winning wager produces $91 profit. The combined outcomes are:

Both win:
Probability = 25%
Net result = +$182
One wins and one loses:
Probability = 50%
Net result = -$9
Both lose:
Probability = 25%
Net result = -$200

Baseline EV:

Baseline EV
= (0.25 x $182) + (0.50 x -$9) + (0.25 x -$200)
= $45.50 - $4.50 - $50
= -$9

Apply weekly cashback:

Both win: no rebate, outcome = +$182
One wins: rebate = $9 x 25% = $2.25, outcome = -$6.75
Both lose: raw rebate = $50, capped rebate = $30, outcome = -$170

Revised EV:

EV with cashback
= (0.25 x $182) + (0.50 x -$6.75) + (0.25 x -$170)
= $45.50 - $3.375 - $42.50
= -$0.375

Rounded to cents, expected loss falls from $9 to about $0.38. Yet one quarter of modeled sequences still lose $170 after rebate. Lower expected loss does not mean low risk.

Now compare per-bet settlement.

Hypothetical sportsbook variation: The same 25% rebate, with a $30 cap per losing bet, settles each bet separately. In the one-win branch, the $100 losing bet earns $25 even though the pair’s net result is only -$9.

One-win revised outcome = -$9 + $25 = +$16
Both-lose revised outcome = -$200 + $25 + $25 = -$150
EV
= (0.25 x $182) + (0.50 x $16) + (0.25 x -$150)
= $45.50 + $8 - $37.50
= $16

The settlement unit changes EV from about -$0.38 to +$16. This is why “25% cashback” cannot be valued without the window and netting rule.

Rebates may remain pending until all included activity settles. Late stat corrections, canceled events, reversed deposits, or game adjustments can change the loss base. Check whether the operator can recalculate a credit and how a pending withdrawal affects eligibility. These are operational questions, not reasons to assume wrongdoing.

Record:

Window start and time zone
Window end
Eligible products
Gross eligible loss
Rate and tier method
Cap
Credit form
Expected credit date and expiry
Actual credit and cash received

The bonus calculator can model branches, while understanding expected value explains why the average is not a forecast. For casino offers, check game weighting and contributions. For sportsbook offers, verify minimum odds and excluded settlement types in the offer itself.

A rebate is not a loss limit. A platform limit or personal stop controls how much activity can occur; cashback only changes settlement after activity has happened. If the rebate is delayed, disputed under the terms, or issued as restricted credit, the original loss still affects available cash in the meantime.

Hypothetical liquidity scenario: A customer sets a $120 weekly loss stop and participates in a 10% cash rebate program. After an eligible $115 loss, the account shows an estimated $11.50 rebate that will settle next week. A new $20 wager could push the weekly cash loss to $135 before rebate.

The wager breaches the $120 stop and should not be placed. Subtracting the expected rebate in advance would report $103.50 and make room appear where none exists. The payment is not settled cash, and the stop was defined on gambling loss before promotional recovery.

Good records preserve both figures:

Gross gambling result = -$115
Pending rebate = $11.50
Net result after rebate, once received = -$103.50

This separation also prevents the operator’s credit date from rewriting the period in a personal loss ledger. Keep gambling result, promotional adjustment, and cash receipt date in distinct columns.

Responsible-gambling controls remain relevant even when a rebate improves expected value. The National Council on Problem Gambling’s voluntary Internet Responsible Gambling Standards include recommendations for account limits, clear promotion information, and access to help. They are not nationwide law. A reader can use stricter personal limits, and a promotion should never be used to loosen them.

The maximum rebate is not a target. In the hypothetical 20% program capped at $25, the first $125 of eligible loss reaches the cap. Losing another $100 earns nothing more. Increasing stakes to move into a higher tier changes the outcome distribution and can cost more in expected loss than the extra rebate is worth.

Set the session and wager plan before opting in. Calculate cashback only on that planned activity. If a tier requires more gambling than the plan allows, value the unreachable tier at zero. Stop at the original time or loss limit even when the settlement window remains open.

Cashback can soften a defined losing branch. It cannot remove variance, convert essential money into risk capital, or make a loss desirable. A rebate claimed after controlled activity is value; activity created to chase the rebate is a new risk decision.