The real probabilities and the house edge
Baccarat has three main bets: Banker, Player, Tie. With eight decks and standard drawing rules their probabilities are fixed. They do not depend on the road, on the previous hand, or on where you sit.
Three numbers
| Outcome | Probability |
|---|---|
| Banker wins | 45.86% |
| Player wins | 44.62% |
| Tie | 9.52% |
You do not have to take anyone’s word for these. We simulated 40,000 shoes, 3.28 million hands with eight decks and the full drawing rules (Player draws on 0–5; Banker follows the table keyed on Player’s third card). The result: Banker 45.87%, Player 44.64%, Tie 9.49% — matching the published figures to two decimals.
One number that rarely gets mentioned: a shoe averages 82 hands, about 74 of which are decided. When we compute streak probabilities later, 74 is the denominator — not the 60 that circulates in a lot of writing.
Why Banker wins more often
Because Banker acts last. Banker sees whether Player drew a third card, and what it was, before deciding. That information is worth roughly 1.24 percentage points. The casino knows this, which is why it charges 5% on Banker wins — the commission exists precisely because Banker has the edge.
How the commission eats it
Excluding ties (a tie pushes):
- Banker 50.68% / Player 49.32%
Banker wins more than half the decided hands. But each win pays only 0.95:
Banker EV = 0.95 × 45.86% − 44.62% = −1.06%
Player EV = 1.00 × 44.62% − 45.86% = −1.24%
Our 3.28-million-hand simulation returns −1.060% and −1.233%.
That is the whole mechanism. You do win more often than you lose; you just collect less on wins than you pay on losses. Every 100 units of turnover evaporates 1.06 units. No betting pattern changes that number, because it is attached to each individual bet.
Side bets: the worst seat at the table
| Bet | Payout | House edge |
|---|---|---|
| Banker (5% commission) | 0.95:1 | −1.06% |
| Player | 1:1 | −1.24% |
| Tie | 8:1 | −14.36% |
| Tie | 9:1 | −4.84% |
| Banker / Player pair | 11:1 | −10.36% |
The Tie costs roughly 13 times what Banker costs. That is not “higher risk, higher reward” — it is the same risk priced far worse.
Pair bets (7.47% paying 11:1) work the same way. They exist to raise the house’s take per hand, not to give you an opening.
No-commission is not cheaper
No-commission baccarat (Super 6) drops the 5% but pays only half when Banker wins with 6. That works out to about a 1.46% house edge, worse than the standard 1.06%. What you save is the bookkeeping, not the money.
The conclusion
All three main bets are negative expectation; the only question is how negative. This is not a solvable problem. It is the structure of the rules: payouts are set slightly below the odds, and that slight gap is the entire revenue of the game.
Once that is clear, roads and staking plans fall into place — they change the order and size of your bets, and the expectation of each bet was fixed before you placed it.
Read next: How common are long streaks · Why staking plans cannot work · How to read baccarat roads · The third card rule · Every side bet, ranked
FAQ
Is Banker really the better bet?
Yes, but barely. Banker loses about 1.06 per 100 staked over the long run; Player about 1.24. Both are negative — Banker just bleeds more slowly.
Is no-commission baccarat a better deal?
No. It replaces the 5% commission with a half payout when Banker wins on 6, which works out to roughly a 1.46% house edge — worse than the standard 1.06%.
Why does the Tie pay so well?
Because it rarely lands. A 9.52% chance paying 8:1 gives −14.36% expectation, the most expensive bet on the table.