The Probability Behind Baccarat, in Detail

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

Working
Let p(Banker win) = 0.4586, p(Player win) = 0.4462, p(Tie) = 0.0952.

Banker bet expected return per unit:

Working
E = (0.4586 x 0.95) + (0.4462 x -1) + (0.0952 x 0) = 0.43567 - 0.4462 = -0.01053

House edge: 1.06%.

Player bet expected return per unit:

Working
E = (0.4462 x 1) + (0.4586 x -1) + (0.0952 x 0) = 0.4462 - 0.4586 = -0.0124

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

ProbabilityValue
Banker win per hand45.86%
Player win per hand44.62%
Tie per hand9.52%
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.