This page collects every original baccarat number calculated for this site, with the method behind each one. The main-bet odds come from exact combinatorial analysis of an eight-deck shoe: Banker wins 45.86% of hands, Player 44.62% and Tie 9.52%. Shoe, streak and betting-system numbers come from simulations of up to 2 million eight-deck shoes.
You're welcome to use any of them. A link back to this page is all I ask.
Baccarat odds, house edge and deck count
Standard rules, 5% commission on Banker wins, ties push Banker and Player bets.
| Win probability | Payout | House edge | Standard deviation per unit | |
|---|---|---|---|---|
| Banker | 45.86% | 0.95 to 1 | 1.06% | 0.93 |
| Player | 44.62% | 1 to 1 | 1.24% | 0.95 |
| Tie | 9.52% | 8 to 1 | 14.36% | 2.64 |
| Tie (9 to 1 tables) | 9.52% | 9 to 1 | 4.84% | – |
Leave ties out and Banker wins 50.68% of decided hands, Player 49.32%. Every bet pays a little less than a fair price, and that shortfall is the house edge:
| Pays | Fair payout | |
|---|---|---|
| Banker | 0.95 to 1 | 0.97 to 1 |
| Player | 1 to 1 | 1.03 to 1 |
| Tie | 8 to 1 | 9.51 to 1 |
Deck count barely matters for the main bets. Pair bets are the exception:
| 1 deck | 6 decks | 8 decks | Infinite deck | |
|---|---|---|---|---|
| Banker wins | 45.96% | 45.87% | 45.86% | 45.84% |
| Player wins | 44.68% | 44.63% | 44.62% | 44.61% |
| Tie | 9.36% | 9.51% | 9.52% | 9.54% |
| Banker edge | 1.01% | 1.06% | 1.06% | 1.06% |
| Player edge | 1.29% | 1.24% | 1.24% | – |
| Tie edge (8 to 1) | 15.75% | 14.44% | 14.36% | 14.12% |
| Tie edge (9 to 1) | 6.38% | 4.93% | 4.84% | – |
| Pair probability | 5.88% | 7.40% | 7.47% | 7.69% |
| Pair edge (11 to 1) | 29.41% | 11.25% | 10.36% | 7.69% |
Online RNG baccarat behaves like the infinite deck column. The commission is what keeps Banker negative:
| Banker house edge | |
|---|---|
| 5% commission | 1.06% |
| 4% commission | 0.60% |
| About 5.39% commission | Banker and Player cost the same |
| No commission, pays 1 to 1 | −1.24% (edge to the player) |
More in baccarat odds explained, how many decks in baccarat and baccarat Banker vs Player odds.
Banker's 1.06% is the lowest house edge on a standard baccarat table, and it is still a negative expected value.
How baccarat hands play out
About a third of hands end on the first four cards.
| Share of hands | |
|---|---|
| Natural 8 or 9, no one draws | 34.30% |
| No natural, both stand | 3.58% |
| Only Player draws | 18.57% |
| Only Banker draws | 11.77% |
| Both draw (6 cards) | 31.77% |
| Player gets a natural | 18.94% |
| Natural tie | 1.79% |
A hand uses 4.94 cards on average. Because 10s and face cards count as zero, 0 is by far the most common two-card total:
| Probability | |
|---|---|
| 0 | 14.74% |
| 1, 3, 5, 7 or 9 | 9.49% each |
| 2, 4, 6 or 8 | 9.45% each |
Player's final total almost decides the hand on its own:
| Player wins | Banker wins | Tie | |
|---|---|---|---|
| Player ends on 0 | 0.00% | 93.83% | 6.17% |
| Player ends on 3 | 16.90% | 77.13% | 5.97% |
| Player ends on 5 | 32.98% | 56.36% | 10.66% |
| Player ends on 6 | 46.95% | 38.61% | 14.44% |
| Player ends on 7 | 61.07% | 23.70% | 15.22% |
| Player ends on 9 | 91.75% | 0.00% | 8.25% |
Ties are not spread evenly across totals. A tie on 7 happens 2.04% of the time and a tie on 6 1.92%, while a tie on 2 is only 0.40%. Full table in baccarat tie odds by total, and every total in baccarat odds by point total.
Where Banker's edge comes from: if Banker drew by Player's simple rule, both sides would win 45.31%. Most of the real gap comes from a handful of drawing decisions.
| Share of the gap | |
|---|---|
| Banker has 5, Player's third card is 0 | 45.78% |
| Banker has 4, Player's third card is 0 | 11.43% |
| Banker has 5, Player's third card is 9 | 10.08% |
| Banker has 5, Player's third card is 1 | 8.50% |
| Banker has 4 or 5, Player's third card is 8 | 8.81% |
| All other cases | about 15% |
Details in why Banker has an advantage and baccarat third card rules.
How each card changes the odds
Removing cards shifts the odds, but only slightly. This is the change in house edge after four cards of one value leave a fresh eight-deck shoe:
| Banker | Player | Tie (8 to 1) | |
|---|---|---|---|
| Four aces | −0.018 | +0.018 | −0.05 |
| Four 2s | −0.021 | +0.022 | +0.09 |
| Four 3s | −0.025 | +0.026 | +0.07 |
| Four 4s | −0.046 | +0.047 | +0.11 |
| Four 5s | +0.033 | −0.033 | +0.10 |
| Four 6s | +0.046 | −0.046 | +0.46 |
| Four 7s | +0.034 | −0.033 | +0.43 |
| Four 8s | +0.020 | −0.022 | −0.29 |
| Four 9s | +0.010 | −0.010 | −0.19 |
| Four 10-value cards | −0.007 | +0.007 | −0.21 |
Small cards leaving help Banker, and 5s through 9s leaving help Player, by a few hundredths of a point. That's why card counting barely touches the main bets. See baccarat card removal effects and how the remaining cards change baccarat odds.
Removing four cards of any value moves the Banker edge by less than 0.05 percentage points.
The effect only gets large when very few cards are left. Two ways to see it:
| Banker | Player | Tie | |
|---|---|---|---|
| Nothing removed | 45.86% | 44.62% | 9.52% |
| Remove 20 zero-value cards | 45.81% | 44.55% | 9.64% |
| Remove five each of A, 2, 3, 4 | 45.94% | 44.56% | 9.49% |
| Remove ten 8s and ten 9s | 45.73% | 44.58% | 9.69% |
| Remove 20 cards from 5, 6, 7 | 45.85% | 44.82% | 9.32% |
| Banker | Player | Tie | |
|---|---|---|---|
| 20 left, all worth 0 | 0.00% | 0.00% | 100.00% |
| 20 left, five each of A, 2, 3, 4 | 30.15% | 47.64% | 22.21% |
| 20 left, ten 8s and ten 9s | 31.73% | 31.73% | 36.53% |
| 20 left, mix similar to a full shoe | 45.93% | 44.46% | 9.61% |
Shoe and streak statistics
The numbers below come from 2 million simulated eight-deck shoes, with the cut card 16 cards from the end. Ties neither extend nor break a streak.
| Result | |
|---|---|
| Hands per shoe | 79.9 on average |
| Decided hands per shoe | 72.3 |
| Ties per shoe | 7.6 |
| Shoes where Banker won more hands than Player | 52.30% |
| Shoes where Player won more hands than Banker | 43.05% |
| Banker's share of decided hands, middle 80% of shoes | 43% to 58% |
| Share of shoes | |
|---|---|
| 5 or more | 92.10% |
| 6 or more | 68.91% |
| 7 or more | 42.51% |
| 8 or more | 23.41% |
| 9 or more | 12.16% |
| 10 or more | 6.15% |
| 11 or more | 3.07% |
| 12 or more | 1.52% |
| Most common longest streak | 6 (26.40% of shoes) |
| Share of shoes | |
|---|---|
| 77 hands or fewer | 6.60% |
| 78 hands | 13.00% |
| 79 hands | 21.60% |
| 80 hands | 24.40% |
| 81 hands | 19.00% |
| 82 hands | 10.30% |
| 83 hands or more | 5.10% |
Betting Banker every hand, the chance of at least one losing run in a single shoe is higher than most players expect:
| Chance | |
|---|---|
| 4 or more losses in a row | 91.32% |
| 5 or more | 67.07% |
| 6 or more | 40.54% |
| 7 or more | 21.91% |
| 8 or more | 11.19% |
| Banker | Player | |
|---|---|---|
| 2 in a row | 25.69% | 24.32% |
| 3 in a row | 13.02% | 12.00% |
| 4 in a row | 6.60% | 5.92% |
| 6 in a row | 1.69% | 1.44% |
| 10 in a row | 0.11% | 0.09% |
| After a Banker streak | After a Player streak | |
|---|---|---|
| 1 in a row | 50.70% | 50.68% |
| 2 in a row | 50.69% | 50.69% |
| 3 in a row | 50.67% | 50.74% |
| 4 in a row | 50.65% | 50.69% |
| 5 in a row | 50.63% | 50.67% |
| 6 in a row | 50.60% | 50.68% |
Streaks don't predict anything. After six Bankers in a row, the next hand went Banker 45.88% of the time. After a tie it went Banker 45.89%, Player 44.61% and Tie 9.51%. After seven or more Bankers in the last ten decisions, Banker's share of the next decided hand was 50.68%, exactly the baseline. See baccarat streak probability, is baccarat rigged, hot baccarat tables and baccarat simulation results.
Player wins more hands than Banker in about 43% of shoes, even though Banker has the edge.
Scoreboard methods and betting systems tested
Every road and prediction rule we tested lands on the same expected value as betting at random, or slightly worse.
| Hit rate | Expected value per bet | |
|---|---|---|
| Follow the last result | 50.01% | −1.27% |
| Bet against the last result | 49.99% | −1.27% |
| Majority of the last 5 | 50.01% | −1.28% |
| Follow streaks of 3+, otherwise bet the chop | 49.99% | −1.29% |
| Follow Big Eye Boy | – | −1.27% |
| Follow Cockroach Pig | – | −1.29% |
| Always Banker | 50.69% | −1.16% |
| Next hand Banker | Next hand Player | |
|---|---|---|
| Big Eye Boy red | 45.90% | 44.58% |
| Big Eye Boy blue | 45.82% | 44.66% |
| Small Road red | 45.90% | 44.57% |
| Small Road blue | 45.83% | 44.67% |
| Cockroach Pig red | 45.88% | 44.61% |
| Cockroach Pig blue | 45.84% | 44.64% |
| Any hand, exact | 45.86% | 44.62% |
| Finished ahead | Worst 1% of sessions | Average result (units) | |
|---|---|---|---|
| Flat 1 unit | 42.7% | −23 or worse | −1.24 |
| Paroli | 43.1% | −42 or worse | −2.10 |
| 1-3-2-6 | 43.1% | −50 or worse | −2.45 |
| D'Alembert | 62.2% | −291 or worse | −7.29 |
| Fibonacci | 91.6% | −378 or worse | −5.25 |
| Chance of this many losses in a row | Lost so far | Next bet | |
|---|---|---|---|
| 3 losses | 13.02% | $70 | $80 |
| 5 losses | 3.34% | $310 | $320 |
| 6 losses | 1.69% | $630 | $640 |
| 7 losses | 0.86% | $1,270 | $1,280 |
The systems that finish ahead most often are the ones with the worst losing sessions. An eight-step Martingale on Banker won 99.64% of its sequences, but the other 0.36%, about 1 in 280, lost 255 units, and overall it lost 1.14% of everything wagered. Labouchère reached its goal in 94.99% of 1 million sessions and Oscar's Grind in 99.14%, and both still averaged a loss. Details in baccarat prediction software, baccarat derived roads, baccarat betting systems tested and Labouchère and Oscar's Grind.
A system that wins most sessions is not a system that wins money.
Bankroll and session statistics
Flat 1-unit Banker bets, 8 decks.
| Chance of being ahead | Average loss (units) | |
|---|---|---|
| After 10 hands | 44.46% | 0.11 |
| After 100 hands | 44.74% | 1.06 |
| After 1,000 hands | 35.76% | 10.58 |
| After 10,000 hands | 12.69% | 105.80 |
| 100 hands | 500 hands | 1,000 hands | |
|---|---|---|---|
| 20-unit bankroll | 4.41% | 43.48% | 62.74% |
| 50-unit bankroll | <0.01% | 3.07% | 16.09% |
| 100-unit bankroll | <0.01% | <0.01% | 0.22% |
| Chance of reaching the goal | Average result (units) | |
|---|---|---|
| Win +5, stop −5 | 46.28% | −0.36 |
| Win +10, stop −10 | 43.68% | −1.33 |
| Win +5, stop −20 | 73.46% | −1.44 |
| Win +20, stop −5 | 15.95% | −1.26 |
Betting 1 unit on Banker every hand of a shoe, 47.6% of shoes end ahead, but the share of players ahead drops to 37.5% after 10 shoes, 15.6% after 100 and fewer than 1 in 1,000 after 1,000. See why you lose at baccarat, baccarat bankroll and risk of ruin and baccarat stop-loss and win goal.
Baccarat side bet statistics
Side bets pay more and cost more. Exact figures for eight decks, using the most common pay tables:
| Win probability | Payout | House edge | |
|---|---|---|---|
| Dragon Bonus on Player | by margin | 1 to 1 up to 30 to 1 | 2.65% |
| Big (5 or 6 cards) | 62.11% | 0.54 to 1 | 4.35% |
| Small (4 cards) | 37.89% | 3 to 2 | 5.28% |
| Dragon 7 (EZ Baccarat) | 2.25% | 40 to 1 | 7.61% |
| Dragon Bonus on Banker | by margin | 1 to 1 up to 30 to 1 | 9.37% |
| Panda 8 (EZ Baccarat) | 3.45% | 25 to 1 | 10.19% |
| Player Pair or Banker Pair | 7.47% | 11 to 1 | 10.36% |
| Perfect Pair (either hand) | 3.35% | 25 to 1 | 13.03% |
| Small Tiger | 3.72% | 22 to 1 | 14.33% |
| Big Tiger | 1.66% | 50 to 1 | 15.25% |
| Super 6 side bet | 5.39% | 12 to 1 | 29.98% |
| Tiger Tie | 1.92% | 35 to 1 | 30.74% |
| Average hits per shoe | Chance of no hits in the whole shoe | |
|---|---|---|
| Dragon 7 | about 1.8 | 16.1% |
| Panda 8 | about 2.8 | 6.0% |
| House edge | |
|---|---|
| 10 to 1 | 17.83% |
| 11 to 1 | 10.36% |
| 12 to 1 | 2.89% |
| Pays | Player side | Banker side | |
|---|---|---|---|
| Win by 9 | 30 | 0.368% | 0.308% |
| Win by 8 | 10 | 0.682% | 0.566% |
| Win by 7 | 6 | 1.792% | 1.591% |
| Win by 6 | 4 | 2.826% | 2.385% |
| Win by 5 | 2 | 3.324% | 3.147% |
| Win by 4 | 1 | 3.737% | 4.024% |
| Natural win | 1 | 16.259% | 16.259% |
| Natural tie | Push | 1.787% | 1.787% |
| Loss | −1 | 69.224% | 69.933% |
| House edge | 2.65% | 9.37% |
Every side bet starts a fresh shoe with a negative expected value. As cards leave the shoe, the mix can briefly push some of them positive. A separate simulation of 20,000 eight-deck shoes, with the cut card 16 cards from the end, found:
| Positive hands per shoe | Shoes with at least one | |
|---|---|---|
| Dragon 7 | 9.39 | 72.44% |
| Panda 8 | 4.15 | 55.67% |
| Player Pair or Banker Pair | 1.70 | 46.74% |
| Tie (8 to 1) | 0.11 | 6.42% |
Those positive moments are short and mostly come late in the shoe. A positive expected value is a long-run average, and the next hand can still lose. Details in how often baccarat side bets turn positive, baccarat side bets expected value, baccarat Tiger bet and baccarat Big and Small.
Baccarat variants and casino hold
| House edge | |
|---|---|
| EZ Baccarat Banker (three-card 7 pushes) | 1.02% |
| Super 6 Banker (win on 6 pays half) | 1.46% |
| Three Card Baccarat Ante, 1 deck, 5% commission | 2.41% |
| Dragon Tiger, Dragon or Tiger (half lost on a tie) | 3.73% |
| Dragon Tiger Tie at 11 to 1 | 10.36% |
| Dragon Tiger Tie at 8 to 1 | 32.77% |
Player stays at 1.24% in the no-commission games. See no commission baccarat and Dragon Tiger vs baccarat.
Casinos report hold, not house edge. Hold is the share of chips bought that the casino keeps. Nevada's statewide baccarat hold ranged from 11.47% to 16.71% between 2017 and 2025, and was 16.07% in 2025, according to the UNLV Center for Gaming Research. The gap from 1.06% comes from the same chips being bet many times. The full explanation is in baccarat hold vs house edge.
A 16% hold and a 1.06% house edge describe the same game measured two different ways.
Methodology and how to cite these numbers
All figures assume standard Punto Banco drawing rules, eight decks unless stated, a 5% Banker commission and ties pushing Banker and Player bets.
Exact figures enumerate every possible order of cards that can affect the hand, weighted by how many ways each can be dealt from the shoe.
Shoe and streak figures come from simulated complete eight-deck shoes with a standard burn and the cut card 16 cards from the end. Their overall results match the exact odds to within 0.03 percentage points.
Betting system and scoreboard figures come from 1 million simulated shoes or sessions, as noted in each table.
Side bet house edges use the most common published pay tables. Other pay tables change the numbers.
To cite: Baccarator, "Baccarat Statistics," www.baccarator.com/blog/baccarat-statistics. Charts on this site can be reused with the same credit and a link.
Watch these numbers move in a real shoe
Every number on this page describes a fresh shoe or a long run of shoes. A shoe in progress is different.
After each hand you enter the cards and tap Calculate, and Baccarator recalculates the odds and expected value for Banker, Player, Tie, pairs and side bets from the cards still in the shoe. If a bet briefly turns positive, it shows a Kelly sizing suggestion. That only appears when expected value is positive, it is a long-run average, and the next hand can still lose. Positive moments are rare and short. You can choose 6 or 8 decks and practice with virtual chips, with no real money involved.
Baccarat statistics FAQ
What are the odds of winning in baccarat?
With eight decks, Banker wins 45.86% of hands, Player 44.62% and Tie 9.52%. Leaving ties out, Banker wins 50.68% of decided hands.
How many hands are in a baccarat shoe?
About 80 with eight decks and the cut card 16 cards from the end, of which about 72 are decided and 7 or 8 are ties.
How often does a 6-hand streak happen in baccarat?
A streak of six or more on Banker or Player shows up in about 69% of eight-deck shoes, and six is the most common longest streak.
Do betting systems change baccarat odds?
No. In 1 million simulated sessions each, every system tested still lost money on average. Some win more sessions, but they lose much more when they fail.
Baccarator is a math tool for education. No real-money gambling. 18+ only.
