Rakeback Is a Discount on the House Edge, Not a Bonus

Put two players at the same table for four hours, running identical basic strategy at identical stakes, and one of them can pay less than half as much for the privilege. Luck has nothing to do with it. One of them was playing somewhere that hands back part of the edge on every settled hand, and that rebate rewrites the arithmetic of the session in a way no welcome offer can.

The house edge is a price, not a forecast

Every game charges a fixed percentage of everything you push through it. Blackjack under decent rules sits close to half a percent against a basic-strategy player. That figure is not a prediction about tonight. It is a price per unit staked, and the bill scales with turnover.

Turnover is where most players lose the thread. It is not the money you deposited. It is the sum of every wager you made, including the same chips recycled twenty times over. A $200 buy-in that survives three hours can easily generate $6,000 of turnover, and the edge is charged on the $6,000.

The rebate maths is one line

If the underlying edge is e and a rebate returns a fraction r of it, your effective edge becomes e × (1 − r).

Take a shoe game running at 99.45% RTP, so e = 0.55%. Rebate 60% of that and you are left paying 0.55 × 0.40 = 0.22%. Flip it back into RTP and you get 99.78%. That is the whole mechanism: the rebate does not add a return stream, it deletes 60% of the cost line.

Cost per hour, worked properly

Assume a heads-up-ish table dealing 70 hands an hour and an average bet of $40. That is $2,800 of turnover per hour.

	No rebate (0.55%)	60% rebate (0.22%)
Cost per hour	$15.40	$6.16
Four-hour session	$61.60	$24.64
100 hours a year	$1,540	$616

Nine dollars an hour sounds trivial until you annualise it. The player putting in two sessions a week is keeping roughly $900 a year that would otherwise have gone to the felt, without changing a single decision at the table.

Live tables at Duel Casino run this model directly: Duel Blackjack Live applies 60% Instant Rakeback to every hand, credited the moment the hand settles rather than pooled and released later, which is what pushes the effective RTP on that product to 99.78%. The felt rules matter as much as the rebate, and there they are the ones you want — blackjack pays 3 to 2, dealer stands on 17, insurance pays 2 to 1 — with crypto stakes running from $10 to $100,000. There are no wagering requirements and no conditions attached to the rebate itself.

Why this is structurally different from a bonus

Run the numbers on a typical bonus and the gap becomes obvious. Say you take $200 with 25x wagering on the bonus amount. That is $5,000 of required turnover — except table games usually contribute at 10%, so unlocking it means $50,000 of blackjack action. At 0.55%, the expected cost of clearing the requirement is $275. You are paying $275 in expected value to liberate $200, and that is before any max-bet clause or cashout cap.

The structural differences:

• A bonus is a one-off with a deadline. A rebate is a permanent change to the rate you pay.

• A bonus is conditional on turnover you may not want to generate. A rebate is earned by the turnover you were generating anyway.

• A bonus is usually locked until a threshold clears. An instant per-hand rebate is in your balance before the next hand is dealt.

• A bonus can be forfeited by a rule breach. There is nothing to breach with a settled rebate.

What a rebate does not do

It does not narrow the distribution. Standard deviation in blackjack is roughly 1.15 betting units per hand, so 280 hands at $40 gives a session standard deviation of 40 × 1.15 × √280 ≈ $769. Your expected cost across those four hours was $24.64. The noise is thirty times the price.

That has a practical consequence: you will not feel the rebate. A losing night at a 0.22% edge feels exactly like a losing night at 0.55%, because both are dominated by variance. The rebate shows up in the ledger over hundreds of hours, not in the memory of any single session.

It also does not make the game positive. A 0.22% edge is still an edge. On $2,800 an hour you are still expected to be down, just slowly.

Five questions worth asking before you count on it

1. Rebate on turnover or on theoretical loss? A percentage of turnover and a percentage of expected loss are wildly different numbers. Rebating 0.3% of turnover on a 0.55% game is a 55% cut of the edge. Rebating 60% of theoretical loss is a 60% cut. Read which base is used.

2. Instant or deferred? A rebate credited per hand is money you can withdraw. A weekly pool that must be claimed inside a window is a rebate you may forget to collect.

3. Is the rebate itself wagered? A rebate with a playthrough attached is a bonus wearing a rebate's clothes.

4. Do the table rules survive? This is the one that catches people. A 6:5 blackjack payout adds about 1.39% to the edge. Rebating 60% of a 1.94% edge leaves you paying 0.78% — materially worse than a plain 3:2 table with no rebate at all. Check the payout before you check the promotion.

5. Is your strategy actually basic strategy? Every published edge figure assumes correct play. Deviating on a few borderline hands per hour can cost more than the rebate returns.

The takeaway

Rakeback is best understood as a price cut on a service you were going to buy anyway. It compounds quietly across every hand, it needs no opt-in behaviour, and it can be evaluated with a single multiplication. A bonus, by contrast, is a conditional credit whose real value only emerges after you model the turnover it demands.

None of this turns gambling into a source of income. Even at 0.22% the expected direction is down, and the hour-to-hour swings will dwarf the edge in both directions. Decide what a session is worth to you as entertainment, set the stake so the arithmetic above stays inside that budget, and treat the rebate as what it is — a smaller bill, not a wage.