Cashback on net losses: why 10% pays more than a tenth of the average loss

Bonus value

Over 200 bets of 5 on red, 1,000 staked, the average loss is 27.03. A 10% cashback on net losses pays 4.37 on average, more than 2.70, because it pays in losing sessions and takes nothing from winning ones; the average loss is still 22.65.

Asymmetric Payment Structure

The lead shows how a cashback rate on net losses yields a higher average payout than simply taking that share of the total expected loss. This occurs because the calculation treats winning and losing sessions differently. Cashback applies only when the net result is negative. If a session ends with a profit, the cashback contribution is zero. The average payout therefore reflects the frequency and size of losing outcomes specifically, rather than the overall average of all outcomes. This distinction creates a value gap between the simple percentage calculation and the actual expected return. For further context on interpreting these averages, see help and support.

200 bets of 5 on red at European roulette, 10% back on a net loss
Average over many sessionsAmount
Total staked1,000
Loss before cashback27.03
Cashback paid4.37
Loss after cashback22.65

Calculating the True Value

The table above breaks down the figures for a session involving many bets. The average loss before cashback is the largest figure shown. The cashback paid is smaller than this initial loss but larger than the simple share calculation. This happens because the cashback averages out losses only from sessions that end down. Winning sessions do not dilute this average by contributing positive balances. Consequently, the cashback amount sits higher than the naive calculation would suggest. The final average loss after cashback remains positive, indicating a net cost to the player despite the refund.

Impact on Expected Loss

Even with cashback, the average session still ends with a loss. The refund reduces this loss but does not eliminate it. The house edge applies to every bet placed, regardless of the bonus structure. Cashback acts as a partial rebate on losses incurred during unfavorable sessions. It does not change the underlying probability of winning individual bets. The average loss after cashback is smaller than the initial average loss, yet it remains a cost. Players should view this refund as a reduction in cost, not as a profit generator or a way to overcome the house edge.

Variability in Individual Sessions

Averages describe long-term behavior over many sessions. Individual results vary significantly from this average. One session might end with a large profit, receiving no cashback. Another might end with a small loss, receiving a proportional refund. The average smooths these variations into a single expected value. This figure helps compare different bonus structures but does not predict a single outcome. Understanding this distinction prevents misinterpreting the average as a guaranteed result. The cashback mechanism stabilizes the average loss slightly but leaves the inherent variance of the game intact for each individual player.

Questions

Why is cashback worth more than a simple share of the average loss?

Cashback applies only to losing sessions, ignoring winning ones. This concentrates the payout on negative outcomes, raising the average payout compared to a flat percentage of the total average loss.

Does cashback eliminate the house edge?

No. The house edge applies to every bet. Cashback refunds a portion of losses but does not change the underlying probability of winning individual bets or the long-term expected loss.

Will my individual session match the average loss?

Not necessarily. Individual sessions vary widely. The average represents a long-term expectation over many sessions, while single results can swing significantly higher or lower than this figure.

How is the cashback amount calculated?

It is a percentage of the net loss in sessions that end down. Sessions ending up contribute nothing to the cashback average, which raises the overall expected payout compared to simple methods.

Every figure on this page is computed by code from an exact calculation over every possible balance after every bet, checked by simulation. See the methodology.

Updated: