Baccarat Strategy: There Are Only Three Bets and One Right Answer
Baccarat has the shortest strategy article in casino gaming, because there is exactly one decision and it never changes.
Bet banker. Every hand.
That's the whole strategy. The rest of this explains why, what the alternatives cost, and why the enormous body of baccarat "systems" is describing something that isn't there.
Why there's nothing to decide
In baccarat you don't play a hand. You bet on which of two hands — Player or Banker — will come closer to 9, and then the dealer executes a fixed set of drawing rules with no input from anyone at the table. Whether a third card is drawn is determined entirely by the rules and the cards already dealt. No one chooses anything.
This is why it's the least skill-dependent good bet in the building. There is no way to play it badly except by betting on the wrong thing.
The three bets
| Bet | House edge | Pays |
|---|---|---|
| Banker | 1.06% | 1:1 less 5% commission |
| Player | 1.24% | 1:1 |
| Tie | 14.4% | 8:1 (sometimes 9:1) |
Banker wins slightly more often than Player — about 45.9% vs 44.6% of non-tie hands. That's because the Banker hand acts after the Player hand and its drawing rules depend on what the Player drew, which is a real informational advantage baked into the rules.
The casino takes a 5% commission on Banker wins specifically to claw that advantage back. Even after the commission, Banker remains the better bet — 1.06% against Player's 1.24%.
The tie bet is a trap. At 8:1 the house edge is 14.4%, roughly fourteen times worse than the bet next to it on the same felt. Ties happen about 9.5% of the time, which needs roughly 9:1 to break even, and 8:1 is what's usually offered. At 9:1 it improves to about 4.8% — still four times worse than Banker. There is no version of the tie bet worth making.
Why the commission doesn't change the answer
A recurring piece of table-side reasoning: "the commission eats the advantage, so bet Player."
It doesn't. The 1.06% figure is calculated after the commission. Banker's raw win rate is high enough that a 5% cut on the wins still leaves it ahead. This is settled by exact enumeration of the full coup over an eight-deck shoe, not by argument — every possible way the hand can play out is counted, weighted by probability, and averaged.
No-commission baccarat, where Banker pays even money but a winning Banker 6 pays only 1:2, is not a better deal. That variant's house edge on Banker is about 1.46% — worse than the commission game. Watch for it.
Why the systems don't work
Baccarat attracts more betting systems than any other game, and casinos happily hand out scorecards to track them.
Pattern tracking. The scorecards, the "big road", the streaks — every hand is dealt from a shoe that neither knows nor cares what came before. Baccarat has slightly more memory than roulette, because cards are removed from the shoe as they're dealt, but the effect is far too small to act on. Card counting in baccarat has been studied thoroughly, and the achievable edge is a small fraction of a percent under conditions that essentially never occur. It's not a viable approach.
Martingale (double after each loss). Guarantees a small win until it doesn't. The sequence of losses that breaks you is inevitable given enough hands, and table limits ensure you'll hit the ceiling before recovering. What Martingale actually does is trade a high probability of winning a little for a small probability of losing enormously — which leaves the expected value exactly where it started, minus the house edge, with more risk attached.
Anything else. No betting pattern changes the house edge of the underlying bets. This isn't a claim about baccarat specifically; it's true of every negative-expectation game. Varying the size of your bets changes your variance, never your expectation.
What actually matters in baccarat
Three things, none of them about which hand to bet on:
1. Bet Banker. Every hand.
2. Never bet the tie. Or any of the side bets that have crept onto baccarat tables — Dragon Bonus, Perfect Pair, Panda 8. They run 3% to 10% edge.
3. Manage the pace. Because there's nothing to decide, baccarat is fast, and house edge applies per bet. The 1.06% is excellent per hand and gets expensive at 120 hands an hour. This is the only lever you actually control.
Where a trainer fits
Honestly: baccarat needs less practice than any other game in a casino, precisely because there's nothing to decide.
What's worth seeing is the arithmetic behind the recommendation. Casino Trainer enumerates the full coup over an eight-deck shoe and shows all three house edges before you bet, so the 1.06% / 1.24% / 14.4% comparison is in front of you at the moment the decision is being made rather than in an article you read last month. That's a more durable way to learn a fact than reading it.
The same solver approach runs the other ten games it covers — blackjack, video poker, Three Card Poker, Ultimate Texas Hold'em, Pai Gow, Let It Ride, Mississippi Stud, Caribbean Stud and more — where there genuinely is something to practise.
Virtual credits only, no real-money play.
The summary
- Banker, 1.06%. Every hand.
- Player is 1.24%. Slightly worse, and fine if the commission annoys you.
- Tie is 14.4%. Never.
- The commission is already in the numbers.
- No system changes any of this.
Baccarat is one of the best bets in a casino and requires you to learn precisely one word.
Want the exact edges shown before every bet, across eleven games? Get Casino Trainer — virtual credits only.
Want the correct play named for you, hand by hand? Get Casino Trainer — virtual credits only, never real money.