Baccarat RTP is the long-run percentage of wagered money that a specific bet is expected to return. It is the mirror image of house edge: a 1.06% house edge corresponds to roughly 98.94% RTP. The important words are long run and wagered money. RTP does not predict what happens to one buy-in, one shoe, one evening, or even one very long session.
The standard baccarat bets do not all have the same RTP
For common eight-deck punto banco rules, the main bets sit in very different places mathematically.
| Wager | Approximate house edge | Approximate RTP | What drives the number |
|---|---|---|---|
| Banker with 5% commission | 1.06% | 98.94% | Banker wins slightly more often, but commission reduces the payoff |
| Player | 1.24% | 98.76% | Even-money payoff with a slightly lower win probability than Banker |
| Tie paying 8:1 | 14.36% | 85.64% | The payout is too small for the rarity of a tie |
| Tie paying 9:1 | 4.84% | 95.16% | The extra unit of payoff improves value substantially |
These are theoretical figures for the stated rules. A different commission method, a no-commission variant, a different Tie paytable, or a side bet creates a different return.
That is why the phrase “baccarat has a 98.94% RTP” is incomplete. The bet has an RTP, not the table as a whole. A player who mixes Banker with high-edge side bets does not experience the Banker number on all of the money wagered.
For the casino-edge view of the same math, read baccarat house edge. For the underlying outcome frequencies, use baccarat odds and the baccarat odds chart.
RTP is measured against turnover, not against your starting bankroll
This is the point most often missed.
Suppose you arrive with a $1,000 bankroll and make 100 Banker bets of $50 each. Your total wagering turnover is:
100 × $50 = $5,000
Using an approximate Banker RTP of 98.94%, the long-run expected return attached to that $5,000 of action is:
$5,000 × 0.9894 = $4,947
The corresponding expected loss is about:
$5,000 × 0.0106 = $53
That does not mean the casino is expected to hand you $4,947 at the end. Most of that “return” is the repeated recycling of chips through wins, losses, pushes, and subsequent bets. You may have begun with only $1,000, yet your chips can generate several thousand dollars of turnover during the session.
This is why comparing RTP directly with the cash in your pocket leads to bad conclusions. A player can lose 20% of a bankroll in a short session even though the wager’s theoretical RTP is close to 99%. The RTP percentage describes the average relationship between all action and all returned value over a huge number of trials.
A useful regulator-facing explanation from the UK Gambling Commission makes the same broad point: return to player is an average measured over a significant volume of play, not a promise for an individual session. Testing organizations such as Gaming Laboratories International use RTP as an expected mathematical return under defined rules.
House edge and RTP are the same relationship viewed from opposite sides
The conversion is simple:
RTP% = 100% - House Edge%
and:
House Edge% = 100% - RTP%
If a wager has a 1.24% house edge, its theoretical RTP is 98.76%. If a wager has an 85.64% RTP, its house edge is 14.36%.
The wording changes the emotional reaction. “98.94% return” sounds generous. “1.06% expected casino advantage” sounds expensive. Mathematically, they describe the same expectation.
The distinction becomes more useful when you compare wagers:
| House edge | RTP | Approximate expected loss per $1,000 of turnover |
|---|---|---|
| 1.00% | 99.00% | $10 |
| 1.50% | 98.50% | $15 |
| 5.00% | 95.00% | $50 |
| 10.00% | 90.00% | $100 |
| 14.36% | 85.64% | $143.60 |
The table does not predict the next $1,000 of action. It shows what repeated action costs on average if the rules and betting mix stay the same.
For a more direct expected-value treatment, read baccarat expected value or use the house edge calculator.
Why Banker normally has the highest main-bet return
Banker wins a little more often than Player because the drawing rules are not symmetrical. In standard commission baccarat, the casino compensates for that advantage by taking commission on winning Banker bets, commonly 5%.
The resulting standard Banker house edge is about 1.06%, versus about 1.24% on Player. That difference is real but small. It does not make Banker a winning bet, and it does not mean Banker must outperform Player in a particular shoe.
The Wizard of Odds baccarat analysis publishes the familiar long-run house-edge figures for the main wagers.
A practical comparison helps:
- $10,000 of Banker turnover at 1.06% theoretical edge implies about $106 expected loss.
- $10,000 of Player turnover at 1.24% implies about $124.
- $10,000 of Tie turnover at a 14.36% edge implies about $1,436.
The first two figures are close enough that short-session variance will overwhelm the difference. The Tie comparison is not close.
Tie payout is a good example of why the paytable matters
Players often see the same word — Tie — and assume the same RTP. The payout proves otherwise.
A tie occurs rarely enough that changing the payoff from 8:1 to 9:1 has a large effect on expected return. Under common baccarat probabilities, the 8:1 version carries a very high house edge, while 9:1 is materially less bad but still well below Banker and Player.
This is a useful lesson beyond Tie bets: never evaluate a casino wager from its name alone. Read the settlement rule and the paytable.
The same principle applies to baccarat side bets. A side bet can have a high hit frequency, an exciting top prize, or a familiar name while still returning far less than the main bets over time.
No-commission baccarat needs its own RTP calculation
“No commission” does not mean the casino removed its mathematical advantage.
A common no-commission structure pays winning Banker bets at 1:1 except when Banker wins with a final total of 6; that result pays only 0.5:1. Nevada’s published Live Baccarat rules of play show exactly that type of settlement in no-commission mode.
Other variants use different adjustments. EZ Baccarat, for example, uses a push mechanism rather than simply copying the half-pay Banker 6 structure. Read no-commission baccarat, Super 6 Baccarat, and EZ Baccarat separately.
The practical rule is simple: if the payout or settlement rule changes, the RTP changes even if the cards are dealt by ordinary baccarat drawing rules.
RTP does not tell you how often a bet wins
Return percentage and win frequency answer different questions.
A bet can win frequently but pay too little. Another can win rarely but pay a large amount. RTP combines the probabilities and payoffs into one money-return expectation.
Baccarat Banker illustrates this clearly. Banker wins more hands than Player before ties are removed, yet winning Banker bets in standard baccarat do not receive full even-money profit because commission is taken. The eventual RTP depends on both how often the result occurs and how much is paid when it occurs.
This is also why a side bet that advertises a spectacular top payout can still have poor RTP. The top event may be so rare that the smaller wins and losses dominate the average.
What a 98.94% RTP can and cannot tell you about a session
Assume a player generates $20,000 of Banker turnover.
The theoretical figures are approximately:
Expected return = $20,000 × 98.94% = $19,788
Expected loss = $20,000 × 1.06% = $212
That $212 is an expectation, not a target the session is trying to reach.
A real session may finish:
| Result after $20,000 of turnover | Compatible with 98.94% RTP? | Why |
|---|---|---|
| +$2,000 | Yes | Short-run variance can overwhelm expectation |
| -$3,500 | Yes | A losing sequence can be much worse than average |
| +$100 | Yes | Near-even sessions are possible |
| -$212 exactly | Possible but not special | Expected value is not a scheduled settlement |
The longer and more often the same wager is played, the more the house edge matters. But “longer” should not be interpreted as a guarantee that an individual player’s result will smoothly converge on the theoretical number.
Use the expected loss calculator to see how wager size and total action change the dollar expectation. For pace, read baccarat expected loss per hour.
RTP is also part of casino ratings and theoretical win
Casinos normally express the same idea from the other side of the ledger.
A player rated at a $500 average bet for 60 decisions has approximately $30,000 in recorded action. If the property applies a 1.06% theoretical edge to that action, the theoretical win is about:
$30,000 × 0.0106 = $318
The player-side theoretical return is the complement of that same number.
Actual win may be far above or below $318. A player can leave thousands ahead while still generating positive theoretical value for the casino. That is one reason comps and host decisions are often linked to rated action and theoretical loss rather than one trip’s actual result alone.
RTP therefore connects three ideas that players often keep separate:
- the game’s mathematical edge;
- the cost of total betting turnover;
- the casino’s estimate of theoretical value.
Errors that make baccarat RTP look more useful than it is
The most common interpretation errors are not arithmetic errors.
Treating RTP as a refund percentage. A 98.94% RTP does not mean 98.94% of your initial bankroll must come back.
Ignoring turnover. Re-betting the same chips creates additional action and therefore additional expected loss.
Using one baccarat RTP for every bet. Banker, Player, Tie, and side bets have separate mathematics.
Ignoring the exact paytable. Tie at 8:1 is not the same wager mathematically as Tie at 9:1.
Assuming no commission means higher return. A replacement rule can cost as much or more than commission.
Confusing high RTP with low volatility. A bet can have a relatively high RTP and still produce severe short-run swings.
Thinking RTP can identify the next winner. RTP says nothing about whether Banker or Player will win the next coup.
Questions players ask about baccarat return percentages
Is baccarat RTP really close to 99%?
For the standard Banker and Player main bets, yes, under common rules. That statement should not be extended to Tie or side bets.
Which standard baccarat bet has the highest RTP?
Banker generally has the highest standard main-bet RTP when the usual 5% commission structure is used.
Does a 98.94% RTP mean I lose only 1.06% of my bankroll?
No. The 1.06% applies to betting turnover, not simply to the amount you first brought to the table.
Can a session return more than 100%?
Yes. Any winning session can finish above its starting bankroll. RTP is a long-run average across repeated wagers, so individual sessions can be far above or below 100%.
Does 100% RTP mean no risk?
A hypothetical 100% expected-return wager can still have variance. RTP measures expectation, not certainty. Standard casino baccarat wagers are below 100% under ordinary rules.
Why does Tie RTP change so much between 8:1 and 9:1?
Because ties are relatively rare. Adding one extra unit of profit on every winning Tie materially changes the average value.
Is online baccarat RTP automatically the same as a live table?
No. Standard drawing rules may be identical, but payout rules, commission treatment, side bets, and variant rules must be checked. The UK remote gambling technical standards are one example of a regulated framework for game information and technical controls, but an individual game’s return still depends on its actual rules.
The useful way to read any baccarat RTP figure
RTP becomes useful only after three questions are answered:
- Which wager is the percentage describing?
- What exact payout and settlement rules apply?
- How much total turnover will be generated?
Once those are clear, the calculation is straightforward.
Expected Return = Total Wagered × RTP
Expected Loss = Total Wagered × (1 - RTP)
For $10,000 of standard Banker action at 98.94% RTP:
Expected Return ≈ $9,894
Expected Loss ≈ $106
The calculation is not a forecast for tonight. It is a pricing tool. It tells you what the wager costs on average as action accumulates.
For a broader comparison, continue with baccarat house edge, baccarat expected value, and baccarat expected loss per hour. Those pages convert the same long-run mathematics into edge, dollars, and playing time.