Baccarat strategy is mostly selection and exposure control, not card prediction. Once the wagers are closed, the Player and Banker hands follow fixed drawing rules. The player does not hit, stand, split or choose which card is drawn. The useful decisions happen before the coup: which wager, which variant, how much, how often and when to stop adding action.
That makes honest baccarat strategy less glamorous than a roadmap system, but far more useful.
The strategy hierarchy
A sensible order of decisions is:
- Identify the exact variant and paytable. “No commission” is not a complete rule description.
- Compare the main wagers. In standard commission baccarat, Banker is normally the lowest-cost main bet and Player is close behind.
- Treat Tie and side bets separately. Their large payouts often come with much higher edges.
- Control total action. A low edge applied to excessive betting still produces substantial expected loss.
- Ignore shoe narratives. Roadmaps record outcomes; they do not alter the next drawing rule.
This hierarchy matters because players often reverse it. They begin with a pattern, increase the wager, add a side bet, and only later ask what the table actually pays.
Standard main-bet math
Under the familiar eight-deck commission game, the approximate all-coup outcome probabilities are:
| Outcome | Approximate probability |
|---|---|
| Banker wins | 45.86% |
| Player wins | 44.62% |
| Tie | 9.52% |
A winning Player wager normally pays even money and pushes on Tie. A winning Banker wager normally pays even money less 5% commission and also pushes on Tie. The commission offsets Banker’s slightly higher win frequency.
For one unit on Player, the expected value is approximately:
[ EV_P = P(\text{Player}) - P(\text{Banker}) ]
[ EV_P \approx 0.4462 - 0.4586 = -0.0124 ]
That corresponds to a house edge of about 1.24%.
For one unit on Banker with 5% commission:
[ EV_B = 0.95P(\text{Banker}) - P(\text{Player}) ]
[ EV_B \approx (0.95 \times 0.4586) - 0.4462 \approx -0.0106 ]
That is about 1.06%. Rounded probabilities make the displayed arithmetic approximate; full combinatorial calculations produce the familiar published figures.
The difference is real but small. It does not make Banker a prediction, and it does not guarantee Banker will win more often in a short shoe.
Lowest edge is not the same as best personal decision
Banker has the lowest standard main-bet edge, but that statement is easy to misuse. It compares equal amounts of action under a specific commission rule. It does not say a $100 Banker wager is cheaper than a $25 Player wager, or that a player should increase the stake to use the mathematically lower-cost side.
For example:
[ $100 imes 1.06% = $1.06 ]
[ $25 imes 1.24% = $0.31 ]
The first wager has the lower percentage edge but the higher expected dollar cost because four times as much was staked. A strategy recommendation that ignores wager size is incomplete.
Commission handling can also affect small-chip play. The theoretical Banker return assumes the full commission is collected accurately. If a property rounds each winning commission to the smallest available chip, the effective cost can depend on wager denomination and the rounding convention. A $20 winning Banker wager produces $1 commission at 5%. A $5 wager produces $0.25, which cannot be settled exactly if the table has no suitable fractional unit. Some properties track commission, some collect it immediately, and some use a no-commission rule instead. The player should verify the actual procedure rather than import the theoretical price from a different table.
This is why “always bet Banker” is too crude. The defensible version is: compare the exact settlement rule, then keep the total dollar exposure within the amount you intended to risk.
Strategy begins with the rule card, not the scoreboard
Standard commission baccarat is only one product. A no-commission table may pay Banker wins at 1:1 except for a special result such as Banker 6 paying half. Another variant may push a particular Banker result. Side bets and progressive wagers add further paytables.
The table name can therefore be less informative than the settlement rule. Ask:
- What does an ordinary Banker win pay?
- What happens when Banker wins with 6?
- Is any Banker result a push?
- What does Tie pay?
- Are commissions collected immediately, tracked, or rounded by chip denomination?
- Which side-bet paytable is in use?
The active Massachusetts table-game rules directory illustrates why exact game documents matter: Baccarat, Midi Baccarat and Mini Baccarat are maintained as approved rule sets, and only authorised games may be offered there. Rules and availability differ elsewhere.
Total action is the decision that compounds
Expected loss is driven by more than the headline edge:
[ \text{Expected loss} = \sum_{i=1}^{k}(A_i \times H_i) ]
where Aᵢ is the action placed on wager type i and Hᵢ is that wager’s house edge.
Suppose a player completes 80 coups with:
- $25 on Banker each coup at an illustrative 1.06% edge; and
- $5 on a side bet each coup at an illustrative 10% edge.
Main-bet action:
[ 80 \times $25 = $2{,}000 ]
Main-bet expected loss:
[ $2{,}000 \times 0.0106 = $21.20 ]
Side-bet action:
[ 80 \times $5 = $400 ]
Side-bet expected loss:
[ $400 \times 0.10 = $40 ]
The optional $5 wager contributes almost twice the expected loss of the $25 main bet. This is why “I only add a small side bet” can radically change the session price. Compare exact paytables through Baccarat side bets ranked rather than treating every bonus wager alike.
Flat betting is not magic, but it limits escalation
Betting the same amount does not remove the house edge. Its value is behavioural: it prevents a losing sequence from automatically forcing the next wager upward.
A Martingale-style progression beginning at $25 creates this exposure after six consecutive losses:
[ 25 + 50 + 100 + 200 + 400 + 800 = $1{,}575 ]
The sequence is vulnerable to table limits, bankroll limits and ordinary losing runs. A later win may recover previous stakes in the idealised sequence, but it does not change the expected value of any coup. The baccarat betting-systems analysis explains the same point without relying on one progression.
Roadmaps can organise attention, not create probability
The bead plate, Big Road and derived roads can help a player or dealer record what happened. They do not modify the shoe’s drawing rules. A continuation system and a reversal system can look at the same column and tell opposite stories, which is evidence of narrative flexibility—not predictive power.
There is one narrow operational distinction worth preserving: cards are dealt without replacement, so the shoe is finite rather than a sequence of independently replaced decks. In theory, composition changes probabilities slightly. In ordinary play, practical card-counting value is constrained by the game’s drawing rules, cut-card placement, limited information and the small size of exploitable changes. A roadmap is not a composition count.
A usable baccarat rule card
For a standard game, a disciplined approach is short:
- Confirm the variant and all special Banker settlements.
- Use Banker or Player if you choose to play; compare their exact cost rather than recent results.
- Avoid assuming Tie or a side bet is attractive because the payout is large.
- Set a stake that does not need to grow after losses.
- Estimate cost by total action, not by the buy-in alone.
- Treat the scoreboard as history.
The expected-loss-per-hour guide is more valuable than a prediction chart because it connects wager size, pace, duration and edge. Baccarat strategy cannot turn a negative game positive. It can stop the player from making the game more expensive than it needs to be.