Baccarat Strategy
Baccarat leaves you one real decision: which bet to make. Every card after that is dealt by a fixed rule, so there is no playing strategy to learn the way there is in blackjack. What remains is how much to stake and when, and that changes the shape of a session without changing its price.
This guide ranks the bets by cost, shows how often long streaks turn up by pure chance, puts numbers on what a doubling system risks, and explains how much the shoe really shifts as cards are dealt. None of it turns the game in your favour, and any method that claims to is wrong about something on this page.
The Only Decision: Which Bet
The house edge is what a bet gives up over time as a share of everything staked on it. Picking Banker over the Tie at 8 to 1 is worth more than any betting system could ever be, because it is the one choice that changes the price. Prices at other deck counts are on the baccarat odds page.
Streaks Happen More Often Than You Think
Ties are left out: a tie neither extends nor breaks a run here, and the rounds counted are rounds with a winner. The middle column covers only the hands about to be dealt: the chance that every one of them goes to Banker, which is small. The last two columns are the chance that one turns up somewhere across a long stretch of play, which is not small at all. A run of five for Banker is more likely than not over that stretch, and a run of ten turns up in about one such stretch in 25.54. A streak is what chance looks like up close, not a sign that something has changed.
Roads and Scoreboards
Most baccarat tables, live and digital, show a scoreboard of recent results: a bead plate with one mark per round, and derived roads that group the same results into columns and patterns. They are an accurate record of what happened in this shoe.
They do not say what happens next. The drawing rule does not look at earlier rounds, and the only way the past affects the next round is through the cards it has used up, an effect measured below and far too small to act on. A road that shows a long Banker column is a picture of a streak, and the table above shows how ordinary streaks are. Following a road, or betting against one, leaves the edge where it was: 1.06% on Banker and 1.24% on Player, whatever the board looks like.
What a Doubling System Risks
Martingale on the Player bet, eight decks: start at one unit, double after every loss, return to one unit after a win, and a tie leaves the bet where it is. Every series that succeeds wins exactly one unit. A series that fails loses the amount in the fourth column, all at once.
Why Doubling Does Not Beat the Edge
Martingale feels safe because it almost always works. At a limit of 500 base bets, a series fails only 0.22% of the time. But each success wins one unit and each failure loses 511. Play 100 series and the chance that at least one of them fails is 19.82%. Add the frequent one-unit gains to the occasional collapse and the average is unchanged: each unit that reaches a decision still gives up 1.36%, the same as Player with ties set aside at a flat stake.
The same holds for every progression. D'Alembert adds one unit after a loss and takes one away after a win. Paroli doubles after wins instead of losses. Fibonacci, Labouchère and 1-3-2-6 follow their own sequences. Each is a schedule for bet sizes, and bet sizes decide how the result of a session is spread out, not what the game charges for each bet. A larger total staked simply means a larger expected cost.
Card Counting
A dealt shoe does change as it goes: every card that leaves alters, slightly, the chances for the next round. That is real, and it is also tiny. In an eight-deck shoe, removing a single card moves the Banker edge by at most 0.011572 of a percentage point and the Player edge by at most 0.011952. Removing a 4 helps a Banker bettor the most; removing a 6 hurts the most. Even with a single deck, where each card weighs more, the largest effect on Banker is 0.093154 of a point.
Against an edge of 1.06%, shifts of that size add up slowly: a great many cards of the same kind would have to leave the shoe before Banker or Player tipped over into a bet worth making. The Tie reacts more strongly, because a tie needs matching totals: a shoe short of 6s and 7s makes it dearer, one short of 8s, 9s or 10-value cards a little cheaper. But it starts from 14.36% at 8 to 1, a long way from break-even. Counting changes what you know about the shoe, not the bet you are making.
Bankroll and Session Length
Over time, what a session costs is the total you stake multiplied by the house edge of the bet. That total grows with every round, so the two things you control are the bet you choose and how many rounds you play. A player who stakes the same amount per round on Banker for twice as long pays, on average, twice as much, and faster tables add rounds to every hour.
In the short run the swings are far larger than the edge, in both directions. A session can end well ahead or well behind, and that says nothing about the price of the game. Decide the amount you are willing to spend before you start, treat it as the cost of the evening, and stop when it is gone. One table rule does reorder the bets, and the no-commission guide explains it; to feel how quickly rounds go by before money is involved, try the free baccarat games.
Baccarat Myths
"Banker is due after a run of Player wins." The shoe keeps no score. A run of Player wins does not make the next Banker win more likely in any way you could use. Banker's chance on the next round stays close to 45.86%, moved only by the cards already dealt, and by far too little to act on.
"A tie is due late in the shoe." A tie comes up in 9.52% of rounds with eight decks. Going many rounds without one is common, and it does not bring the next one closer.
"Always bet with the streak." Or against it: either way the edge on each bet is the one in the table above. A streak is only visible once it has happened.
"The right system turns the edge around." A system chooses bet sizes. The price of each bet is set by the cards and the payout, and no sequence of bet sizes can change it.
Baccarat Strategy FAQs
Where the 5% commission is charged, it is the lowest-cost habit available, sometimes called the one-sided approach: back Banker every round and ignore the scoreboard. It does not beat the game; it only keeps every stake on the cheapest bet. At a no-commission table the same logic points to Player instead.
It wins small amounts often and loses a large amount occasionally. On Player with a table limit of 500 base bets, one series in 453.19 fails and costs 511 units. Over time the result is the same edge as flat betting.
Not to any profit. How often a digital game reshuffles is the game's own setting, stated only in its rules, and even a shoe dealt to the end shifts too little to act on, as the card counting section above measures.
Not on price. At either payout it is several times dearer than Banker or Player, and the odds page shows how much the 9 to 1 version saves. Back it for the big return if you like, knowing that is what you are paying for.