Odds & House Edge
The mathematics behind every bet on the table.
Every casino game can be reduced to a set of numbers. In baccarat, those numbers are unusually transparent and unusually favorable compared to most table games. This guide gives you every probability, every house edge calculation, every expected value formula, and every comparison you need to understand exactly what you are risking and why. No approximations, no vague reassurances β just the math, explained completely.
Core Probabilities: 8-Deck Shoe
The foundation of all baccarat odds is the probability of each possible outcome in an 8-deck shoe (416 cards). These numbers are derived from exhaustive combinatorial analysis of all possible two- and three-card hand combinations under the standard Punto Banco rules. They are not estimates β they are precise calculations.
| Outcome | Probability | Percentage | Odds Against |
|---|---|---|---|
| Banker Win | 0.458597 | 45.86% | 1.18:1 |
| Player Win | 0.446247 | 44.62% | 1.24:1 |
| Tie | 0.095156 | 9.52% | 9.51:1 |
| Total | 1.000000 | 100% | β |
Banker wins more often than Player β 45.86% vs. 44.62%. This 1.24 percentage point difference is the entire basis for the 5% commission on Banker bets. Without the commission, Banker would be a player-advantageous bet. The commission is precisely calibrated to convert Banker's natural mathematical edge into a 1.06% house edge β low enough to attract serious players who understand the math, but still profitable for the casino at scale.
Non-Tie Probabilities: The Decisive Split
Many analysts prefer to examine probabilities excluding ties, since Tie outcomes push on both Player and Banker main bets:
| Outcome | Probability (all hands) | Probability (non-tie hands only) |
|---|---|---|
| Banker Win | 45.86% | 50.68% |
| Player Win | 44.62% | 49.32% |
Excluding ties, Banker wins just over 50% of decisive hands. This is where Banker's mathematical advantage becomes stark: in a pure head-to-head comparison of decisive outcomes, Banker is actually the stronger side by more than one percentage point. The 5% commission on Banker wins exists to correct this imbalance in the casino's favor.
House Edge: The Three Main Bets
House edge is the casino's mathematical advantage expressed as a percentage of each dollar wagered. It represents the expected loss per dollar bet over an infinite number of hands β the long-run tax the game extracts from every wager.
Calculating the Banker Bet House Edge
Banker Bet: Full Expected Value Calculation
When Banker wins, Player's bet is lost. When Banker wins, the bettor receives 0.95:1 (after the 5% commission). When a Tie occurs, Banker bets push (return original stake).
EV = (Prob. Banker Wins Γ Payout) + (Prob. Player Wins Γ β1) + (Prob. Tie Γ 0)
= (0.458597 Γ 0.95) + (0.446247 Γ β1) + (0.095156 Γ 0)
= 0.435667 β 0.446247 + 0
= β0.010579
House Edge = 1.0579% β 1.06%
For every $100 wagered on Banker, expect to lose $1.06 in the long run.
Calculating the Player Bet House Edge
Player Bet: Full Expected Value Calculation
Player bets pay 1:1 (even money) with no commission. When Player wins, the bettor receives $1 for every $1 wagered. When Banker wins, the Player bettor loses their wager. Ties push.
EV = (Prob. Player Wins Γ 1) + (Prob. Banker Wins Γ β1) + (Prob. Tie Γ 0)
= (0.446247 Γ 1) + (0.458597 Γ β1) + (0.095156 Γ 0)
= 0.446247 β 0.458597
= β0.012350
House Edge = 1.2350% β 1.24%
For every $100 wagered on Player, expect to lose $1.24 in the long run.
Calculating the Tie Bet House Edge
Tie Bet (8:1 payout): Full Expected Value Calculation
The Tie bet wins only when both Player and Banker end the hand with the same point total. The standard payout is 8:1 (some casinos offer 9:1). When Tie does not occur, the entire Tie bet stake is lost.
EV = (Prob. Tie Γ 8) + (Prob. No Tie Γ β1)
= (0.095156 Γ 8) + (0.904844 Γ β1)
= 0.761248 β 0.904844
= β0.143596
House Edge = 14.36%
For every $100 wagered on Tie, expect to lose $14.36. This is one of the worst bets in any casino table game.
Note: At 9:1 payout, the Tie bet house edge reduces to approximately 4.84% β still nearly 5 times worse than the Banker bet, but dramatically better than the standard 8:1 offering.
| Bet | Payout | House Edge (8 decks) | Expected Loss per $100 | Expected Loss per $1,000 |
|---|---|---|---|---|
| Banker | 0.95:1 | 1.06% | $1.06 | $10.60 |
| Player | 1:1 | 1.24% | $1.24 | $12.40 |
| Tie (8:1) | 8:1 | 14.36% | $14.36 | $143.60 |
| Tie (9:1) | 9:1 | ~4.84% | $4.84 | $48.40 |
Why Banker Has a Lower House Edge Than Player
The mathematical reason Banker wins more often is the Banker's optimized third card rules. When Player draws a third card, Banker's drawing decision depends on both its own total AND the value of Player's third card. This conditional strategy β drawing on 3 unless Player drew an 8, drawing on 4 only when Player drew 2-7, and so on β is mathematically calibrated to maximize Banker's win rate across all possible card combinations.
Player's rules are simpler and fixed: draw on 0-5, stand on 6-7. They don't adapt to what Banker is holding. This asymmetry gives Banker a permanent structural advantage that compounds over millions of hands β and is why Banker wins more than 50% of decisive hands.
What the 5% Commission Actually Does
Without the 5% commission, Banker would be a player-favorable bet β the house would be giving away money on every Banker wager. The commission structure converts Banker's natural 50.68% win rate on decisive hands into a house edge of 1.06%. Here's the math on why 5% specifically was chosen:
| Commission Rate | Banker House Edge (8 decks) | Is this favorable to player or house? |
|---|---|---|
| 0% (no commission) | β1.24% (player advantage!) | Player-favorable |
| 1% | β0.78% (player advantage) | Player-favorable |
| 2% | β0.32% (player advantage) | Player-favorable |
| 2.75% | ~0.03% | Nearly break-even |
| 3% | 0.14% | House edge (small) |
| 5% | 1.06% | Standard house edge |
At a 5% commission, the house extracts a meaningful but not excessive edge. At 2.75%, the game would be nearly break-even β which is why some high-roller casinos briefly experimented with 2.75% commission to attract whale players. The 5% standard balances player appeal with casino profitability.
How Number of Decks Affects Baccarat Odds
| Decks | Banker Win % | Player Win % | Tie % | Banker Edge | Player Edge | Tie Edge (8:1) |
|---|---|---|---|---|---|---|
| 1 | 45.96% | 44.68% | 9.36% | 1.01% | 1.29% | 15.75% |
| 6 | 45.87% | 44.63% | 9.51% | 1.056% | 1.237% | 14.44% |
| 8 | 45.86% | 44.62% | 9.52% | 1.058% | 1.235% | 14.36% |
The practical difference between deck counts is minimal for the main bets. Single-deck baccarat slightly favors the player (lower Banker edge of 1.01%), but single-deck games are extremely rare. The 6-to-8 deck difference is $0.02 per $100 wagered on Banker β effectively negligible. If you have a choice, prefer fewer decks, but it should not drive table selection decisions.
Notably, single-deck Tie bet house edge (15.75%) is actually higher than 8-deck (14.36%) β fewer decks make ties slightly less likely due to fewer duplicate cards. Counterintuitively, single-deck baccarat is worse for Tie bet players.
No-Commission Baccarat Odds
No-commission baccarat (Super 6) changes Banker payouts: wins pay 1:1 normally, but Banker winning with a 6 pays only 1:2. Despite appearing player-friendly (no commission tracking), the math shows a higher Banker house edge:
| Bet | Standard Punto Banco | No Commission (Super 6) | Difference |
|---|---|---|---|
| Banker House Edge | 1.06% | ~1.46% | +0.40% worse for player |
| Player House Edge | 1.24% | 1.24% | No change |
Why does No Commission produce a higher edge? Banker wins with a 6 approximately 5% of the time. The reduced payout on these wins extracts more revenue from players than a blanket 5% commission on all wins. The math doesn't lie: "no commission" costs you more than the standard commission.
EZ Baccarat: The Best Banker Odds Available
EZ Baccarat eliminates commission but uses a different mechanism: when Banker wins with a three-card total of exactly 7, the hand is a push for Banker bettors. All other Banker wins pay 1:1. The result is the lowest Banker house edge of any standard game:
| Event | Payout | Probability | House Edge Contribution |
|---|---|---|---|
| Banker wins (3-card 7) | 0 (push) | 2.25% | +0.0225 |
| Banker wins (all other) | 1:1 | 43.61% | β0.4361 |
| Player wins | β1 | 44.62% | +0.4462 |
| Tie | 0 | 9.52% | 0 |
| Total Banker House Edge | 1.02% | ||
EZ Baccarat's 1.02% Banker edge is the best available on any standard baccarat main bet. The three-card Banker 7 push is a more surgical commission mechanism than the blanket 5% β it targets a specific, relatively rare outcome rather than taxing every single Banker win.
Side Bet Odds and House Edges
Player Pair and Banker Pair
Pair bets win when the first two cards of either side form a matching pair (e.g., two Kings). The payout is typically 11:1. In an 8-deck shoe, the probability that any two specific cards are a pair involves drawing from 416 cards:
Pair Bet Probability Calculation (8-deck shoe)
The first card is dealt. The probability the second card matches its rank:
Number of cards of the same rank remaining: 31 (since one of the 32 cards of that rank has been dealt)
Cards remaining: 415
Probability = 31/415 = 7.47%
At 11:1 payout: EV = (0.0747 Γ 11) + (0.9253 Γ β1) = 0.8217 β 0.9253 = β0.1036
House Edge = 10.36%
| Side Bet | Payout | Probability (8 decks) | House Edge | Relative to Banker Bet |
|---|---|---|---|---|
| Player Pair | 11:1 | 7.47% | 10.36% | 9.8Γ worse |
| Banker Pair | 11:1 | 7.47% | 10.36% | 9.8Γ worse |
| Dragon Bonus β Player | Up to 30:1 | Various | ~2.65% | 2.5Γ worse |
| Dragon Bonus β Banker | Up to 30:1 | Various | ~9.37% | 8.8Γ worse |
| Dragon 7 (EZ) | 40:1 | ~2.25% | 7.61% | 7.2Γ worse |
| Panda 8 (EZ) | 25:1 | ~3.45% | 10.19% | 9.6Γ worse |
Dragon Bonus Payout Table
| Winning Margin | Dragon Bonus Payout | Approximate Probability |
|---|---|---|
| Natural winner (non-tie) | 1:1 | ~16.3% |
| Win by 9 points | 30:1 | ~2.3% |
| Win by 8 points | 10:1 | ~2.7% |
| Win by 7 points | 6:1 | ~3.6% |
| Win by 6 points | 4:1 | ~4.5% |
| Win by 5 points | 2:1 | ~5.8% |
| Win by 4 points | 1:1 | ~7.7% |
| Win by 3 points or fewer | Loss | ~57.1% |
The Dragon Bonus is a losing bet more than 57% of the time β and that's before even accounting for the house edge. The appealing 30:1 payout for a 9-point blowout win occurs less than 2.3% of hands.
Expected Value: Worked Session Calculations
Session 1: 200-Hand Session, $50 Banker Bets
Total wagered: 200 Γ $50 = $10,000
Expected loss: $10,000 Γ 1.06% = $106
Expected Banker wins: 200 Γ 45.86% = 91.7 β 92 wins
Expected Player wins: 200 Γ 44.62% β 89 (where Banker bettor loses)
Expected Ties (push): 200 Γ 9.52% β 19
Net decisive hands: approximately 181. On those 181 hands: win ~92, lose ~89. Net β +3 units. But each Banker win pays only 0.95 units, so: +3 Γ $50 = +$150 minus 3 Γ $2.50 commission = +$142.50 net on average... wait β the EV already accounts for commission. Expected loss of $106 on $10,000 wagered is correct.
Session 2: Comparing Banker vs. Player Over 200 Hands at $50
Banker bets: $10,000 Γ 1.06% = $106 expected loss
Player bets: $10,000 Γ 1.24% = $124 expected loss
Annual difference (2 sessions/week, 52 weeks): ($124 β $106) Γ 104 = $1,872 per year
Betting Player instead of Banker costs nearly $1,900 per year in additional expected losses for a player doing two $50/hand sessions per week. The math compounds into real money.
Session 3: The Tie Bet Cost
A player places one $25 Tie bet per hand alongside $25 Banker bets (200 hands total):
Banker bets: $5,000 Γ 1.06% = $53 expected loss
Tie bets: $5,000 Γ 14.36% = $718 expected loss
Total expected loss: $771
Without Tie bets: $53. With Tie bets: $771. Adding one $25 Tie bet per hand increases expected session loss by 1,354% for the same total stake.
Session 4: EZ Baccarat Advantage Over Standard
200 hands at $100 Banker, EZ Baccarat vs. standard:
Standard Punto Banco: $20,000 Γ 1.058% = $211.60 expected loss
EZ Baccarat: $20,000 Γ 1.02% = $204 expected loss
Saved: $7.60 per 200-hand session
At two $100/hand sessions per week for 52 weeks: $7.60 Γ 104 = $790 saved annually by choosing EZ Baccarat over standard.
Variance and Standard Deviation in Baccarat Sessions
Expected value tells you the average outcome over many sessions. Variance tells you how much any individual session will deviate from that average. For baccarat (Banker bet), the standard deviation per hand is approximately 0.93 units β derived from the near-50/50 outcome distribution combined with the commission structure.
Session Variance Formula
Standard deviation for N hands: SD_session = 0.93 Γ unit_size Γ βN
| Session Length | Unit Size | Expected Loss | Standard Deviation | 68% Range | 95% Range |
|---|---|---|---|---|---|
| 80 hands | $25 | -$21.20 | $207.70 | -$229 to +$187 | -$436 to +$394 |
| 80 hands | $50 | -$42.40 | $415.40 | -$458 to +$373 | -$873 to +$788 |
| 200 hands | $25 | -$53 | $328.30 | -$381 to +$275 | -$709 to +$603 |
| 200 hands | $50 | -$106 | $656.60 | -$763 to +$551 | -$1,419 to +$1,207 |
| 500 hands | $50 | -$265 | $1,038.50 | -$1,304 to +$774 | -$2,342 to +$1,812 |
The high standard deviation relative to expected value explains why baccarat players often win significant amounts in any given session β even though the long-run expectation is negative. A β$106 expected loss for a 200-hand session means that roughly 44% of those sessions will end with a profit (because the distribution skews right of the expected value by about one standard deviation on positive outcomes).
This variance is your short-term opportunity and long-term constraint. Short sessions maximize your probability of walking away ahead. Long sessions allow the expected value to dominate over variance, grinding the player toward the predicted loss. See the Bankroll Management guide for how to structure session length based on this mathematics.
Baccarat vs. Other Casino Games: House Edge Comparison
| Game / Bet | House Edge | Strategy Required? | Hands/Decisions per Hour | Expected Loss/Hour ($25 unit) |
|---|---|---|---|---|
| Baccarat Banker | 1.06% | No | 60β80 | $16β$21 |
| Baccarat Player | 1.24% | No | 60β80 | $19β$25 |
| Blackjack (perfect basic strategy) | 0.40β0.60% | Yes β full strategy memorization required | 60β80 | $6β$12 |
| Craps (Pass Line) | 1.41% | No | 30β40 rolls | $11β$14 |
| Craps (Pass + maximum odds) | ~0.30% | Yes β odds structure knowledge required | 30β40 rolls | $2β$4 |
| European Roulette (single zero) | 2.70% | No | 30β40 spins | $20β$27 |
| American Roulette (double zero) | 5.26% | No | 30β40 spins | $39β$53 |
| Three Card Poker (Ante/Play) | 2.00% | Basic strategy needed | 40β50 hands | $20β$25 |
| Pai Gow Poker | 2.50% | Yes β setting strategy | 20β30 hands | $12β$19 |
| Baccarat Tie (8:1) | 14.36% | No | 60β80 | $215β$287 |
| Slot Machines (typical) | 2β15% | No | 400β600 spins | Highly variable |
Baccarat Banker at 1.06% ranks third-lowest of any casino bet available to average players β behind only blackjack with perfect basic strategy and craps with maximum odds. The critical distinction: both blackjack and craps with optimal strategy require significant knowledge and execution effort. Baccarat's 1.06% edge is achieved with zero strategic knowledge. Just bet Banker.
Card Counting in Baccarat: Why It Doesn't Work
Card counting is effective in blackjack because specific card removals β particularly Aces and 10s β shift game probabilities by meaningful amounts. In baccarat, the effect of removed cards is orders of magnitude smaller. Here is why:
- In blackjack, removing an Ace from the deck reduces the natural (blackjack) probability by approximately 0.36% and affects soft totals significantly
- In baccarat, removing an Ace shifts Banker-vs-Player win probability by approximately 0.0003% β one tenth of one percent of the blackjack effect
- Most baccarat card removal effects are so small they cannot overcome the minimum bet increment (e.g., a $5 table forces you to bet in $5 increments; the edge gained never justifies the precision required)
- The only theoretically positive counting situation in baccarat involves the Tie bet at specific extreme shoe compositions β but these situations are so rare that waiting for them produces enormous negative expected value from the Banker/Player bets placed while waiting
Edward Thorp β who invented card counting for blackjack β analyzed baccarat and concluded that counting provides no practical advantage. Anyone claiming a baccarat card counting system is either mistaken or selling something. The game cannot be beaten through card counting.
Practical Implications: Five Rules from the Math
- Always bet Banker: The 0.18% edge difference (Banker 1.06% vs. Player 1.24%) compounds significantly over a year of regular play. There is no legitimate mathematical reason to prefer Player.
- Never bet Tie: A 14.36% house edge makes the Tie bet economically irrational. The fair payout for a Tie would be approximately 9.5:1; at 8:1 you're paying 6 percentage points of pure premium to the house.
- Prefer EZ Baccarat when available: The 1.02% Banker edge beats standard 1.06%. On serious money, this saves real dollars.
- Avoid pair bets and most side bets: Pair bets (10.36%), Dragon Bonus Banker (9.37%), Panda 8 (10.19%) all carry house edges that make the Tie bet look competitive. None belong in a disciplined bankroll.
- Keep sessions short: Variance is your friend in short sessions; the house edge is your enemy in long ones. The longer you play, the more expected value dominates over variance, grinding toward predicted losses.
Frequently Asked Questions
What is the house edge in baccarat?
Banker bet: 1.06%. Player bet: 1.24%. Tie bet (8:1 payout): 14.36%. In an 8-deck shoe, these are the house edges for standard Punto Banco baccarat. EZ Baccarat offers a Banker edge of 1.02%; No Commission (Super 6) Banker edge is approximately 1.46%.
What are the odds of winning in baccarat?
In an 8-deck shoe: Banker wins 45.86% of hands, Player wins 44.62%, and Ties occur 9.52% of the time. Excluding ties, Banker wins 50.68% of decisive hands. No betting position offers better-than-even odds against the house when accounting for payouts and commissions.
Why is the Banker bet better than the Player bet in baccarat?
Banker wins more often (45.86% vs. 44.62%) due to its optimized third card drawing rules. Banker's conditional strategy accounts for the value of Player's third card, giving it a mathematical win-rate advantage. The 5% commission converts this advantage into a 1.06% house edge; Player carries 1.24% because of its lower win rate.
Is baccarat the best odds in the casino?
Baccarat Banker at 1.06% is among the best available β third-lowest behind blackjack with perfect basic strategy (0.4-0.6%) and craps with maximum odds (~0.3%). Unlike those games, baccarat's optimal edge requires zero strategic knowledge. Just bet Banker.
Does baccarat pay even money?
Player bets pay 1:1 (even money). Banker bets pay 0.95:1 β you receive $0.95 for every $1 wagered, due to the 5% commission. Tie bets pay 8:1 at most casinos (some offer 9:1, reducing the house edge to approximately 4.84%). The commission on Banker is why its house edge (1.06%) is lower than Player's (1.24%) despite both being near-50/50 wagers.
Can you count cards in baccarat?
Technically possible but practically worthless. The effect of removed cards on baccarat probabilities is so small β typically hundredths of a percent β that no counting strategy produces a meaningful, exploitable edge. Edward Thorp, who invented blackjack card counting, analyzed baccarat and confirmed it cannot be beaten through counting. Anyone claiming otherwise is wrong.
What is the probability of a natural in baccarat?
Either Player or Banker receives a natural (two-card 8 or 9) in approximately 32.6% of all hands in an 8-deck shoe. When either side has a natural, the hand ends immediately β no third cards drawn for either side. The natural check always takes precedence over all third card rules.
What is variance in baccarat and why does it matter?
Variance is the statistical measure of how much actual results deviate from the expected value in any given session. In baccarat, the standard deviation per hand is approximately 0.93 units β meaning individual session results can vary significantly from the expected loss. A 200-hand session at $50/hand has an expected loss of $106 but a standard deviation of $657, meaning most sessions fall anywhere from -$763 to +$551. High variance is why short sessions are often profitable despite a negative expected value.