Level 5 · Lesson 18 of 4 · Advanced Player
The Probability Behind Baccarat, in Detail
The shoe
An 8-deck baccarat shoe contains 416 cards. Each rank appears 32 times (8 decks x 4 suits). Cards 2 through 9 count face value. Ace counts 1. Tens, Jacks, Queens, Kings all count zero.
Cards worth zero make up 128 of the 416 cards: 30.77% of the shoe. Every other rank, including Aces, is 7.69% of the shoe.
The most important result of the two-card distribution: the combined probability of a natural (an 8 or 9 on the first two cards, for either Player or Banker) is approximately 32.2%. Natural hands end immediately. No third cards. Highest natural wins, or it ties.
Why Banker wins more than Player
Banker wins 45.86% of hands. Player wins 44.62%. The 1.24% gap is structural.
Player's drawing rule is simple: draw on totals 0-5, stand on 6 or 7. If Player draws, Banker's decision then depends on both Banker's current total and the value of Player's third card.
The full Banker drawing tableau:
- Total 0-2: always draw
- Total 3: draw unless Player's third card was an 8
- Total 4: draw if Player's third card was 2-7
- Total 5: draw if Player's third card was 4-7
- Total 6: draw only if Player's third card was 6 or 7
- Total 7: always stand
That conditioning on Player's third card is the key. Banker draws only when the known third card is likely to improve its position, and avoids draws that would worsen it. This informational advantage tilts the distribution slightly in Banker's favour. Without any commission, Banker bets would carry a player edge of roughly 1.24%. The 5% commission converts most of that structural advantage into casino margin, leaving the house at 1.06%.
At the Venetian Macao's high-limit rooms, croupiers deal this tableau without consulting a card. It's memorised. It has to be: the room runs at pace, and each hand's third-card decision must be instant.
The edge calculation from first principles
Banker bet expected return per unit:
House edge: 1.06%.
Player bet expected return per unit:
House edge: 1.24%.
These are exact derivations from the probability distribution of the 8-deck shoe and the drawing tableau. They match Eliot Jacobson's published tables and every other serious mathematical treatment of the game.
Naturals
A natural ends the hand before any third cards are dealt. The probability of a natural 8 or 9 on either side is roughly 32.2% of hands. When both sides have naturals, the higher wins; equal naturals tie.
When Player receives a natural, Banker's tableau advantage is irrelevant: no third card is drawn, and the structural conditioning that gives Banker its edge doesn't apply. Natural hands are the scenario where the probabilities are closest to symmetric.
The Tie bet probability in detail
The 9.52% Tie frequency is the basis of the 14.36% house edge on the 8 to 1 payout. For a fair 8 to 1 bet, the tie would need to occur on 1 in 9 hands, or 11.11%. The actual 9.52% falls short, and the gap between 9.52% and 11.11%, amplified across the payout structure, produces the 14.36% edge.
A 9 to 1 payout on Tie reduces the edge to 4.85%. This version exists at some tables. It is still 4.5 times the Banker edge. The Tie bet is not competitive with the main bets at any standard payout.
What the probability structure tells you about the game
The 45.86%-44.62%-9.52% split is not the casino's invention. It follows mathematically from the rules of the drawing tableau applied to a standard 8-deck shoe. The drawing rules were codified over decades of play. The commission rate of 5% was calibrated against that probability structure to produce a 1.06% house edge. All of this can be derived from the rules.
At the Bellagio Salon Prive in Las Vegas, a streak scoreboard tracks Banker and Player runs for the session. Regulars study it. The probability structure guarantees that streaks of six or more Banker wins will appear in a typical 70-hand session. Streaks are not evidence of anything. They are the expected mathematics of a 50.68% Banker win rate in head-to-head comparisons.
Probability of runs and streaks
A common question from students of baccarat probability: what's the probability of Banker winning six consecutive hands?
Treating ties as pushes (both Banker and Player bets survive the tie):
- Adjusted Banker win rate (excluding ties): 45.86% / (45.86% + 44.62%) = approximately 50.68%
- Probability of six consecutive Banker wins: 50.68%^6 = approximately 1.73%
That's roughly one in 58 six-hand sequences. In a 70-hand session, you should expect at least one such streak on average. Streaks are not aberrations. They are expected outcomes of the probability distribution.
The Bellagio Salon Prive in Las Vegas has a Banker streak scoreboard that regulars consult. The streaks they see are not evidence of anything unusual. They are the mathematics behaving exactly as expected.
Key numbers
| Probability | Value |
|---|---|
| Banker win per hand | 45.86% |
| Player win per hand | 44.62% |
| Tie per hand | 9.52% |
| Natural (8 or 9) probability per side | ~32.2% |
| Banker win rate (ties excluded) | ~50.68% |
| Probability of 6 consecutive Banker wins (ties excluded) | ~1.73% |
Sources: Eliot Jacobson probability analysis, Macau DICJ gaming statistics, Bellagio casino table games, Venetian Macao baccarat, UK Gambling Commission technical standards.
Welcome to the lesson on the probability behind baccarat, in detail.
I'm Annabel, and this lesson is for those of you who want to understand where the numbers come from. Not just that the Banker edge is one point zero six percent, but why. If you understand the structure, the edge figures stop being assertions and become inevitable results.
Start with the shoe.
An eight-deck baccarat shoe contains four hundred and sixteen cards. Each rank appears thirty-two times. Cards two through nine count face value. Aces count one. Tens, Jacks, Queens, and Kings count zero.
Cards worth zero make up thirty point seven seven percent of the shoe. Every other rank, including Aces, is seven point six nine percent.
The most important result from this: the combined probability of a natural, an eight or nine on the first two cards for either side, is approximately thirty-two point two percent. Natural hands end immediately. Highest natural wins, or it ties. No third cards.
Now, why does Banker win more than Player?
Banker wins forty-five point eight six percent of hands. Player wins forty-four point six two percent. The gap is structural.
Player's drawing rule is simple: draw on totals zero through five, stand on six or seven. Banker's drawing rule is conditioned on Player's third card when Player draws. The Banker tableau in full:
Total zero, one, or two: always draw. Total three: draw unless Player's third card was an eight. Total four: draw if Player's third card was two through seven. Total five: draw if Player's third card was four through seven. Total six: draw only if Player's third card was six or seven. Total seven: always stand.
That conditioning on Player's third card is the key. Banker draws only when the known third card is likely to improve its position. This informational advantage tilts the distribution in Banker's favour. Without any commission, Banker bets would carry a player edge. The five percent commission is calibrated to convert most of that structural advantage into casino margin, leaving the house at one point zero six percent.
The croupiers at the Venetian Macao's high-limit rooms deal this tableau from memory, without a reference card. At the pace those rooms run, the decision has to be instant.
Now the edge calculation from first principles.
Banker expected return per unit: forty-five point eight six percent multiplied by zero point nine five, minus forty-four point six two percent multiplied by one, plus nine point five two percent multiplied by zero. Result: zero point four three five six seven minus zero point four four six two, equals negative zero point zero one zero five three. House edge: one point zero six percent.
Player expected return per unit: forty-four point six two percent minus forty-five point eight six percent, equals negative zero point zero one two four. House edge: one point two four percent.
These are exact derivations from the probability distribution of the eight-deck shoe and the drawing tableau. They match Eliot Jacobson's published tables at apheat.net and every other serious mathematical treatment.
A note on streaks. The question comes up constantly: what's the probability of Banker winning six consecutive hands?
Treating ties as pushes, the adjusted Banker win rate is approximately fifty point six eight percent, which is Banker's forty-five point eight six divided by the sum of forty-five point eight six and forty-four point six two. The probability of six consecutive Banker wins is fifty point six eight percent to the power of six: approximately one point seven three percent. Roughly one in fifty-eight six-hand sequences.
In a seventy-hand session, you should expect at least one streak of that length on average. Streaks are not aberrations. They are the mathematics behaving exactly as expected.
Here is the thing I most need you to take from this lesson.
The one point zero six percent Banker edge is not a casino estimate or a marketing figure. It is the exact result of the probability distribution of a standard eight-deck shoe. The probabilities were fixed by the rules. The commission was calibrated to those probabilities. The edge that results is one point zero six percent. It doesn't move.
Understanding that doesn't improve your play in any mechanical sense. There are no decisions to make during the hand. But it means the figure isn't something you're asked to trust. It's something you can derive. That's a different kind of confidence.