Baccarat Strategy: Why the Banker Bet Wins on Math and Scoreboard Systems Don't
Quick answer
The best baccarat strategy is simple: bet banker, skip the tie, and flat bet a small share of your bankroll. With 8 decks the banker bet has a 1.06% house edge after the 5% commission, player 1.24%, and tie 14.36% at 8 to 1. Scoreboard patterns and progressive systems don't change these numbers.
| Lowest-edge bet | Banker: 1.06% (8 decks, 5% commission) |
|---|---|
| Player bet | 1.24% |
| Worst bet | Tie at 8 to 1: 14.36% |
| Banker without commission | Would give players a 1.24% advantage |
| Do patterns work? | No. Each hand is independent of the scoreboard |
The best baccarat strategy is also the simplest: bet banker, never bet the tie, and keep your stakes small. With 8 decks, the banker bet has a house edge of 1.06% after the standard 5% commission. The player bet costs 1.24%, and the tie costs 14.36% at 8 to 1.
That's the whole edge story. Everything else sold as baccarat strategy (scoreboards, road maps, streak-following, progressive betting) changes which side you bet or how much, but not the house edge. This guide shows the math behind each claim using our exact enumeration of every possible baccarat deal.
How baccarat works (quick rules)
You bet on banker, player or tie. Two cards go to each hand; tens and face cards count 0, aces 1, and only the last digit of a total counts. A natural 8 or 9 ends the hand. Otherwise a fixed table decides whether each hand takes a third card. The hand closest to 9 wins.
Because the rules decide every draw, your only decisions are which bet and how much. The full drawing table is in our baccarat rules and odds guide.
The banker bet: lowest house edge in baccarat
Our exact calculation for an 8-deck shoe gives these probabilities per hand dealt:
| Result | Probability (8 decks) | Pays | House edge |
|---|---|---|---|
| Banker wins | 45.8597% | 1 to 1, minus 5% commission | 1.06% |
| Player wins | 44.6247% | 1 to 1 | 1.24% |
| Tie | 9.5156% | 8 to 1 | 14.36% |
These match the Wizard of Odds 8-deck figures (banker 0.458597, player 0.446247, tie 0.095156; edges 1.06%, 1.24% and 14.36%).
Banker wins more often because it acts second: its third-card decision takes the player's third card into account. That's why the casino charges a commission on banker wins, and it's why banker still comes out ahead of player after paying it.
Commission math: how the 5% works
Start with what banker would be worth with no commission. It would win 45.8597% and lose 44.6247%, with ties pushing:
banker advantage with no commission = P(banker) − P(player) = 0.458597 − 0.446247 = +0.012350Without commission, banker bettors would have a 1.24% edge over the casino.
Now take 5% from each win. The commission costs the bettor 5% of the stake on the 45.86% of hands that banker wins:
commission cost per hand = 0.05 × P(banker) = 0.05 × 0.458597 = 0.022930About 2.29% of each banker bet goes to commission on average.
banker edge = 0.022930 − 0.012350 = 0.010580 → 1.06%Same result as P(player) − 0.95 × P(banker). The commission more than cancels banker's natural advantage.
The commission at which banker would be exactly fair is 1 − P(player) ÷ P(banker) = 1 − 0.446247 ÷ 0.458597 ≈ 2.69%. Anything above that gives the house an edge; the standard 5% gives it 1.06%.
Banker vs player in USDT. Per 100 USDT wagered, banker costs about 1.06 USDT and player about 1.24 USDT, a difference of 0.18 USDT. Small per bet, but it adds up over thousands of hands, and there's no reason to give it away.
The tie bet: the worst bet on the table
A tie happens in 9.52% of 8-deck hands. A fair payout would be the true odds:
fair tie payout = (1 − 0.095156) ÷ 0.095156 ≈ 9.51 to 1An 8 to 1 payout, the figure the Wizard of Odds uses, falls well short of that.
tie edge at 8 to 1 = 1 − 9 × 0.095156 = 0.143596 → 14.36%At 9 to 1, the edge would be 1 − 10 × 0.095156 = 4.84%.
| Shoe | Banker | Player | Tie (8 to 1) | Tie (9 to 1) |
|---|---|---|---|---|
| 8 decks | 1.06% | 1.24% | 14.36% | 4.84% |
| 6 decks | 1.06% | 1.24% | 14.44% | 4.93% |
| 1 deck | 1.01% | 1.29% | 15.75% | 6.38% |
The tie gets worse as decks are removed: from 14.36% with 8 decks to 15.75% with one.
Why scoreboard and pattern strategies don't work
Baccarat tables often display "roads" or scoreboards that record past results, and many systems tell you to follow a streak, bet against it, or wait for a pattern. They fail for a simple reason: the cards don't know what's on the board.
Britannica calls the belief behind these systems the doctrine of the maturity of the chances, better known as the gambler's fallacy: it falsely assumes that independent plays depend on each other. OpenStax defines two events as independent when the outcome of one doesn't change the probability of the other, which is exactly the situation here. The chance of a banker win on the next hand is set by the cards in the shoe and the drawing rules, not by the pattern on the scoreboard.
Here's the decisive point: whatever a pattern tells you, you end up betting banker or player. Those two bets have fixed edges of 1.06% and 1.24%. A pattern system can only shift your mix between them, and every hand it moves to player costs you an extra 0.18 points.
Betting systems: Martingale and friends
Progressive systems change bet size after wins or losses. The Wizard of Odds explains that Martingale doesn't dent the house edge: you win a small amount often and lose a large amount occasionally.
We don't have a baccarat-specific system simulation, but our roulette test of even-money betting shows the effect. From a 100 USDT bankroll, betting red on a single-zero wheel:
| System | Reached +50 USDT goal | Average result per session | Loss per 100 USDT wagered |
|---|---|---|---|
| Flat 1 USDT | 2.4% | -26.81 USDT | 2.71 USDT |
| Martingale | 48.7% | -6.46 USDT | 2.77 USDT |
| D'Alembert | 52.3% | -16.98 USDT | 2.84 USDT |
| Fibonacci | 48.5% | -12.43 USDT | 2.62 USDT |
Martingale reached the profit goal in about half of sessions, but the other half ended busted or unable to make the next bet. Its smaller average loss per session only reflects shorter sessions: it wagered 233 USDT on average against 989 USDT for flat betting. Per 100 USDT wagered, every system lost close to the wheel's 2.70% edge. Baccarat's banker bet has a lower edge, but the same logic holds: the system can't turn a negative-edge bet into a positive one. See our roulette strategy simulations for the full test.
What each bet costs over a session
Bet choice is the one lever that changes your expected cost. Here is what 100 hands at 10 USDT (1,000 USDT wagered) costs on average with each bet, using our exact edges:
| Shoe | Banker | Player | Tie (8 to 1) |
|---|---|---|---|
| 8 decks | 10.58 USDT | 12.35 USDT | 143.60 USDT |
| 6 decks | 10.56 USDT | 12.37 USDT | 144.38 USDT |
| 1 deck | 10.12 USDT | 12.86 USDT | 157.46 USDT |
With 8 decks, sticking to banker instead of player saves about 1.77 USDT per 1,000 USDT wagered. Putting the same money on the tie costs roughly 13.6 times as much as banker. These are averages; any one session will land above or below them, which is exactly why a budget and a stop point matter more than any pattern.
A practical baccarat strategy
- Set a session budget in USDT and only play with money you can afford to lose.
- Bet banker every hand. It's the lowest-cost bet at every deck count.
- Never bet the tie, and treat any side bet with suspicion unless its edge is published.
- Flat bet. Keep each bet to a small, fixed share of your budget.
- Estimate your cost: total wagered × 0.01058. For 100 hands at 10 USDT, that's about 10.58 USDT on average.
- Stop when you hit your limit, win or lose.
Playing baccarat with USDT
A stablecoin bankroll keeps the math clear: the house edge is the main force on your balance, without a coin's price moving underneath you. Network fees on deposits and withdrawals are a fixed cost per transfer, so fewer, larger transfers keep them down. Our USDT payment guides compare networks. If you want to compare baccarat's edge with crypto-native games, our dice guide shows a game with a published 1% edge, and the dice odds calculator shows how payout and win chance trade off.
Common mistakes
- Switching to player to "avoid the commission".
- Betting the tie because the payout looks big.
- Following road maps or streaks.
- Raising stakes after losses to win it back.
For how baccarat compares with other games, see the house edge comparison, or return to the baccarat hub.
Frequently asked questions
What is the best strategy for baccarat?
Bet banker on every hand, avoid the tie, and keep each bet small relative to your bankroll. The banker bet has the lowest house edge at 1.06% with 8 decks. No strategy can push that below zero, so the goal is to pay as little as possible for your play, not to find a winning system.
Should I bet player to avoid the commission?
No. The commission is already included in the banker's 1.06% house edge, and that's still lower than the player bet's 1.24%. Banker wins more often than player, enough to more than make up for the 5% taken from each win. Over 100 USDT of bets, banker costs about 0.18 USDT less on average.
Do baccarat patterns or road maps work?
No. Scoreboards record past results, but the next hand comes from the cards left in the shoe, not from the pattern on the board. Believing a streak must continue or end is the gambler's fallacy. Pattern systems change which side you bet, but every banker or player bet keeps its fixed edge.
Does the Martingale system work in baccarat?
No. The Wizard of Odds explains that Martingale doesn't dent the house edge. Doubling after losses wins small amounts often and loses large amounts occasionally. Our roulette simulation of Martingale lost about the same share of money wagered as flat betting, and the same logic applies to baccarat's even-money bets.
Why is the tie bet so bad?
A tie happens in only 9.52% of hands with 8 decks. For a fair bet, that would need to pay about 9.5 to 1. Paying 8 to 1 leaves a house edge of 14.36%, so you lose about 14.36 USDT per 100 USDT bet on average, more than 13 times the banker bet's cost.
Sources
- Baccarat Basics — Wizard of Odds. 8 decks: banker 0.458597, player 0.446247, tie 0.095156; edge banker 1.06%, player 1.24%, tie 14.36% at 8 to 1; drawing rules
- Doctrine of the maturity of the chances — Encyclopaedia Britannica. Gambler's fallacy: falsely assumes plays are dependent
- Martingale Betting System — Wizard of Odds. Martingale does not dent the house edge
- Two Basic Rules of Probability — OpenStax, Rice University. Multiplication rule; independent events