Banker carries a lower house edge than Player. That single fact is the most repeated — and most misused — statistic in baccarat, so it is worth putting the important half first.
With that established, here is where the numbers actually come from, what they cost, and what they do not entitle you to conclude.
The numbers
For the standard eight-deck game with 5% commission on Banker wins:
| Bet | Pays | House edge | Expected cost per $100 |
|---|---|---|---|
| Banker | 1:1 less 5% commission | ≈ 1.06% | −$1.06 |
| Player | 1:1 | ≈ 1.24% | −$1.24 |
| Tie | 8:1 | ≈ 14.36% | −$14.36 |
Where the edge comes from
Across all possible eight-deck outcomes, the three results occur at roughly:
- Banker wins — 45.86% of hands
- Player wins — 44.62% of hands
- Tie — 9.52% of hands, about one in ten
Ties push on Banker and Player bets, so ignore them and look at decided hands only: Banker takes about 50.68% of them and Player about 49.32%. That is the entire asymmetry — roughly a 1.4-point lead, and it exists for a structural reason.
Why Banker wins slightly more often
The drawing rules are not symmetrical. The Player hand acts first and draws by a fixed rule. The Banker hand acts second, and its drawing decision depends on what the Player hand drew. Acting with that extra information is worth a little over one percentage point of decided hands.
Nothing about this is negotiable or seat-dependent — it is baked into the rules of the game, and it is the same at every table.
Why the commission exists
Paid at even money, that structural advantage would make Banker a positive-expectation bet, and no casino offers those. The 5% commission on winning Banker bets exists to take the advantage back:
- Without commission, Banker would favour the bettor.
- With 5% commission, Banker’s house edge is about 1.06%.
- Player needs no commission, because it is already behind.
So the commission is not a fee for a good bet. It is the mechanism that converts a good bet into a slightly-less-bad one.
The Tie bet
At 8:1, the Tie bet carries a house edge of about 14.36% — more than thirteen times Banker’s. Ties arrive on roughly 9.5% of hands, about one in ten, while an 8:1 payout would need them closer to one in nine just to break even.
Some tables pay 9:1, which reduces the edge substantially but leaves it far above the two main bets. However dramatic a tie looks when it lands, the arithmetic behind it does not improve.
What the difference is actually worth
Banker’s edge is about 0.18 percentage points lower than Player’s. Concretely, per $100 wagered, the long-run expected cost is about $1.06 against about $1.24 — a difference of roughly 18 cents.
It is real, and it is small, and it only exists as an average over a very large number of hands. Over a single session it is thoroughly buried by ordinary variance. Anyone presenting that 18 cents as an opportunity is describing a rounding error as a strategy.
A board that looks nothing like the odds
Here is a real 24-hand stretch in which Banker won 20 times, including a seven-hand streak long enough to bend along the bottom of the grid. The board is drawn by the same engine that powers our free Road Decoder:
A shoe like this is what a “hot Banker table” looks like, and it is entirely ordinary: across the thousands of hands a table deals, runs this lopsided appear regularly. The board records that they happened. It does not make them more or less likely to continue.
Can a system overcome the edge?
No — and the reason is worth understanding rather than taking on trust.
The house edge is a property of the payout structure, not of the sequence of hands. Banker pays 0.95 on a bet that wins 50.68% of decided hands; that shortfall is fixed, and it applies identically to the first hand of a shoe, the hand after a nine-hand streak, and the hand after any scoreboard pattern you can name.
Betting progressions change the size of your bets, not the edge on any of them. A sequence of negative-expectation bets stays negative-expectation regardless of the order or the amounts — the arithmetic has no mechanism to turn around.
And the scoreboards cannot help either, because they contain no information about undealt cards. The derived roads in particular are computed from the Big Road, so they add nothing the Big Road did not already have. If that is new, the complete guide to reading baccarat roads walks through exactly what each board is and is not measuring.
What the numbers are actually good for
Knowing the edge will not put you ahead of the house. It does three useful things:
- It sets expectations honestly — every bet on the table has a negative expected value, and now you know each one’s size.
- It makes the Tie bet’s cost visible, which is the one genuinely large number on the board.
- It gives you an immediate test for anything sold as a strategy: if a claim requires the edge to change, the claim is wrong.
Read the board for what it is
The scoreboards are worth learning — as literacy. They let you follow the table, understand what other players are reacting to, and recognise immediately when someone is describing a pattern that is not there.
Our Road Decoder builds all five boards live as you enter results, free and without an account. Enter the Banker-heavy shoe above and watch what the roads do with it — then remind yourself what the next hand’s odds are.