BAC 110: Baccarat Odds

In standard eight-deck Punto Banco baccarat, the long-run outcome probabilities are approximately:

• Banker win: 45.8597%
• Player win: 44.6247%
• Tie: 9.5156%

Those percentages describe all completed coups. Banker and Player wagers normally push on a Tie, so the chance that Banker wins a resolved Banker-versus-Player decision is higher than 45.86%.

All-coup probability and resolved-hand probability are different

Let:

• (P_B) be the probability of a Banker win;
• (P_P) be the probability of a Player win;
• (P_T) be the probability of a Tie.

Using the standard eight-deck values:

[ P_B=0.458597,\quad P_P=0.446247,\quad P_T=0.095156 ]

Banker and Player decisions resolve on:

[ P_B+P_P=0.904844 ]

or about 90.4844% of coups.

Conditional on a non-Tie result:

[ P(Banker\mid resolved)=\frac{P_B}{P_B+P_P} ]

[ P(Banker\mid resolved)=\frac{0.458597}{0.904844}=50.6824% ]

Player wins the remaining 49.3176% of resolved decisions.

This is why two statements can both be correct:

• Banker wins fewer than half of all coups.
• Banker wins slightly more than half of resolved Banker-versus-Player coups.

The baccarat odds chart provides a compact reference. This page explains what the percentages mean and how payout rules convert them into expected value.

Why Banker wins slightly more often

Punto Banco does not allow the bettor to choose hits or stands. The Player hand follows one drawing rule. The Banker hand follows a rule that can depend on the Player total and, when Player draws, the value of Player’s third card.

That informational structure gives Banker a small probability advantage. It is not a prediction that Banker must win the next coup, and it does not mean Banker wins most of every shoe.

The exact drawing sequence is explained in the third-card rule guide. A practical bettor does not need to memorize the chart to place a wager, because the dealer or game system executes it. The chart matters for understanding why Banker and Player are not perfectly symmetrical.

Fair payout and casino payout

A bet is fairly priced when its expected value is zero.

Banker without commission

If Banker paid even money and Tie pushed:

[ EV_B=P_B(1)-P_P(1)+P_T(0) ]

[ EV_B=0.458597-0.446247=+0.01235 ]

That would be a 1.235% player advantage. Standard commission baccarat removes that advantage by paying less than even money on Banker wins.

The break-even Banker profit payout (r) satisfies:

[ P_Br-P_P=0 ]

[ r=\frac{P_P}{P_B}=0.97307 ]

A fair Banker win would therefore pay about 0.97307 to 1. Standard 5% commission pays 0.95 to 1, leaving the casino with the familiar edge.

Player at even money

Player wins pay 1 to 1 and Tie pushes:

[ EV_P=P_P-P_B ]

[ EV_P=0.446247-0.458597=-0.01235 ]

The Player house edge is about 1.235%.

Tie at 8 to 1

Tie wins about 9.5156% of coups and loses on Banker or Player outcomes. At 8 to 1:

[ EV_T=P_T(8)-(1-P_T) ]

[ EV_T=0.095156(8)-0.904844=-0.143596 ]

The house edge is about 14.3596%.

A fair Tie payout would be:

[ \frac{1-P_T}{P_T}=9.5091\text{ to }1 ]

The common 8-to-1 award is therefore well below fair odds. A 9-to-1 Tie payout is better, but it still does not fully reach the approximately 9.51-to-1 break-even level for this standard model.

Standard expected values per 100 units wagered

The baccarat house-edge guide compares these prices with no-commission variants and side bets. The baccarat payouts page focuses on chip settlement, commission, and total return.

A 100-coup example is not a forecast

If 100 coups followed the long-run proportions exactly, they would produce roughly:

• 46 Banker wins;
• 45 Player wins;
• 9 or 10 Ties.

A real 100-coup sample will not normally match those rounded counts. Banker could win 55 times, Player could dominate, or Ties could cluster. The probabilities describe the distribution generating outcomes, not a quota the shoe must fill.

Suppose a player wagers 50 units on Banker for 80 coups:

[ Total\ action=50\times80=4{,}000\text{ units} ]

Using a 1.0579% edge:

[ Expected\ loss\approx4{,}000\times0.010579=42.32\text{ units} ]

That is a long-run average. The actual session can move hundreds of units in either direction because each resolved hand is close to an even-money result.

Ties reduce decisions but not time or exposure automatically

A Tie normally returns the Banker or Player stake. It does not create a win, but it also does not create a loss for those main wagers.

A player counting “80 hands” may mean:

• 80 total coups, including pushes; or
• 80 resolved Banker/Player decisions.

Those are different volumes. On average, obtaining 80 resolved decisions requires about:

[ \frac{80}{0.904844}=88.41\text{ coups} ]

This matters when estimating game speed, session duration, commission handling, and the number of actual win-or-loss decisions.

Tie frequency is low per coup but common across a session

A 9.5156% Tie probability means most individual coups are not Ties. Across many coups, however, seeing at least one Tie becomes likely.

For (n) coups:

[ P(at\ least\ one\ Tie)=1-(1-P_T)^n ]

Using (P_T=0.095156):

This does not make a Tie wager attractive. It only shows the difference between “a Tie occurs somewhere in the session” and “this specific Tie bet is fairly paid.” A player can correctly expect to observe Ties and still face a large negative expectation by betting on every coup.

Commission procedure can affect practical chip handling

Standard Banker mathematics assumes a 5% commission on every winning Banker wager. Casinos collect that amount in different approved ways: immediately from each win, through a commission box or ledger, or at another defined settlement point.

For a 25-unit Banker win:

[ Gross\ profit=25 ]

[ Commission=25 imes0.05=1.25 ]

[ Net\ profit=23.75 ]

A live table must have a procedure for fractional commission and chip denominations. Rounding rules, deferred collection, and minimum commission units can affect the exact result of a small wager. Those operational details should be read from the posted rule rather than assumed from the headline “5% commission.”

The theoretical 1.0579% figure uses exact 5% commission. A table that rounds commission differently should be analyzed from its actual settlement convention.

Card removal changes probabilities slightly

Baccarat is dealt from a finite shoe. As cards are exposed, the composition of the remaining shoe changes. The exact next-coup probabilities therefore are not mathematically identical after every deal.

That fact does not validate scoreboards or roadmaps. A pattern of Banker and Player labels does not contain enough composition information to calculate the remaining shoe. Two shoes can show the same visible outcome history while having different remaining card ranks because the underlying cards that produced those outcomes differed.

Composition-sensitive analysis requires tracking card values and applying a verified model. Even then, practical value depends on penetration, rules, accuracy, bet limits, and whether the available wager becomes positive after commission. The standard percentages on this page are baseline full-shoe values, not a claim that every shoe state is identical.

Rule variants change payout before they change probability

“No commission” baccarat is not one universal rule. Common structures include:

• Banker 6 paying half;
• Banker 6 pushing;
• selected totals or card counts receiving special settlement;
• side bets attached to the main game.

The underlying Banker, Player, and Tie result may be generated by familiar drawing rules while the wager settlement changes. House edge must be recalculated from the actual paytable.

Do not apply the standard 1.06% Banker figure to a table that pays Banker wins differently. The help screen or placard must identify the commission and special-result rules.

Operational rules define what is being priced

Official rules separate the three main wagers and define when each wins, loses, or pushes. The Massachusetts baccarat rules provide one regulated example of the dealing and settlement framework. Other jurisdictions and approved variants may use different commission collection, dealing formats, or side wagers.

For a live table, useful checks include:

• commission rate and collection method;
• Tie payout;
• number of decks;
• no-commission exceptions;
• whether side bets have separate posted paytables;
• whether main wagers push on Tie.

A mathematically accurate percentage attached to the wrong rule is still the wrong answer.

What the odds support as a decision

The standard ranking of the three main wagers is straightforward:

1. Banker is usually the lowest-cost main bet after 5% commission.
2. Player is close, but slightly more expensive.
3. The common 8-to-1 Tie bet is much more expensive.

That ranking does not guarantee a session result. It identifies the long-run price of repeated action.

Roadmaps, streak labels, and recent outcomes do not improve the baseline payout. The useful choices are to verify the rule version, prefer lower-cost wagers, control total action, and keep probability separate from prediction.

Use the baccarat odds calculator to compare standard outcomes and the variance simulator to see how widely short samples can move around expectation. For the individual main wagers, continue to Banker bet, Player bet, and Tie bet.