Why Side-Bet Counting Sometimes Works in Baccarat (And Pure Counting Doesn’t)

Baccarat has a reputation as the unbeatable game. Card counters who built fortunes at blackjack tried for decades to crack it, and the math kept refusing to cooperate. The story isn’t quite that simple, though. Pure baccarat counting really is a dead end at the main bets, but a handful of side bets break the pattern. They pay big multiples on rare outcomes, and rare outcomes are exactly what depleted shoes start to favor. This is where baccarat side bet counting stops being a fantasy and starts looking like arithmetic.

Why Counting the Main Game Earns You Almost Nothing

The Banker and Player bets in baccarat are governed by a fixed drawing rule. Neither side has a choice. You can’t stand on 16 against a 10, you can’t double, you can’t surrender. The dealer just follows the chart. Because of that, the only way the count can help you is by shifting the underlying probability of Banker winning, Player winning, or a Tie, and then betting the side that’s currently favored.

Edward Thorp and Peter Griffin ran this calculation in the 1960s and 70s. Even with perfect knowledge of every card removed from an eight-deck shoe down to the cut card, the maximum theoretical advantage on the main game comes out to roughly 0.0007 per hand at the deepest penetration. That’s seven hundredths of one percent. In practical terms, a counter playing a full shoe might find one or two hands where Player or Banker swings into a tiny positive expectation, and the gain over a whole shoe is on the order of 0.01% of total action.

To put that in dollars: if you bet $100 a hand for 70 hands, you’re risking $7,000 of action, and your expected gain from counting is under a dollar. The house edge on Banker (about 1.06%) and the cost of cover bets eat that instantly. This is why nobody serious counts the main game.

Why Side Bets Are a Different Animal

Side bets in baccarat pay multiples. Dragon Bonus pays 30 to 1 on a natural winning by 9. Lucky 6 pays 12 to 1 (sometimes 20 to 1 for a three-card 6). Tiger Pair pays 4 to 1 or higher. These payouts depend on specific cards remaining in the shoe — a 30:1 outcome that needs particular ranks to show up can swing wildly as the deck depletes.

The math here is the same idea that makes blackjack countable. When the cards you need are over-represented in what’s left, the bet’s expected value goes up. When they’re depleted, it drops. Side bets in baccarat have much higher variance and much steeper payout curves, so the swings are bigger. A side bet sitting at -10% house edge with a full shoe can flip to +10% player edge in the last 20 cards.

Dragon Bonus and the 30-to-1 Hook

Dragon Bonus pays based on the margin of victory and whether the win was a natural (two-card 8 or 9). The headline payout is 30 to 1 on a non-natural winning by 9. The full table depends on the casino, but the structure is always the same: bigger margins pay more, naturals pay separately.

To win by 9, one side typically needs to draw a third card and land on a specific total while the other side stays low. That requires particular cards to still be in the shoe. A count that tracks small cards versus large cards (because small cards help complete big-margin hands after a draw) can identify shoes where the 30:1 outcome is meaningfully more likely than its baseline frequency.

The catch is the baseline. A 30:1 payout that hits about 1 in 50 hands is already close to break-even. Push the frequency to 1 in 40 and you’ve got a healthy edge. Drop it to 1 in 60 and you’re being skinned. The count tells you which world you’re in.

Lucky 6 and the Real Edge

Lucky 6 pays when Banker wins with a total of 6. Most casinos pay 12 to 1 for a two-card 6 and 20 to 1 for a three-card 6. The Wizard of Odds has published full analyses of this bet at wizardofodds.com/games/baccarat/card-counting, and the conclusion is consistent across simulations: Lucky 6 is the most countable side bet in baccarat by a wide margin.

The reason is the card composition needed. A Banker total of 6 happens most often when specific small cards are still available to be drawn as the third card. When the shoe depletes 4s, 5s, and 6s, the bet becomes structurally less likely to hit. When those cards are over-represented late in the shoe, the bet becomes more likely to hit, and the 12:1 payout overwhelms the house edge.

At a typical 50-card cut (meaning the dealer reshuffles when 50 cards are left), the deepest playable counts on Lucky 6 produce a player advantage of roughly 10%. That number gets quoted a lot, and it’s not marketing — it shows up in independent simulations. The trade-off is that you only see those counts in a small fraction of shoes, and the variance is brutal.

Tiger Pair and Smaller Side Bets

Tiger Pair pays when the Banker’s first two cards form a pair. Variants include Big Tiger (Banker wins with a 6 made of three cards), Small Tiger (Banker wins with a two-card 6), and Tiger Tie. Each one is countable in principle because each one depends on the rank composition of what’s left in the shoe.

The edges on Tiger variants are smaller than Lucky 6, usually peaking at 2% to 5% at extreme counts. They’re more useful as part of a multi-side-bet strategy than as a standalone target. A counter who tracks Lucky 6 and Tiger Pair together gets more playable hands per shoe than someone tracking just one bet.

Panda 8 is the same idea on the Player side: it pays 25 to 1 when Player wins with a three-card 8. The countable conditions are similar to Lucky 6 but mirrored, and the edges run a couple of points lower. If your table offers both Lucky 6 and Panda 8, you can sometimes find a shoe where one or the other is playable on nearly every hand of the last 30 cards, though that combination is rare.

Approximate Advantage by Count

The exact numbers vary by paytable and penetration, but the rough shape looks like this:

Practical Bet Spread and Detection

The hard part isn’t the math, it’s the execution. You sit through dozens of hands counting cards while making minimum bets on the main game. When the count goes positive on the side bet you’re tracking, you spike your side-bet wager. Then the count crashes again and you go back to flat. Pit bosses notice this pattern fast.

The standard detection signals casinos watch for include:

• Players who only bet side bets when the shoe is half depleted or later.
• Wagering that jumps by 10x or more between hands with no obvious trigger.
• Sitting out hands and rejoining late in the shoe.
• Watching the discard tray instead of the table action.
• Avoiding the main bets entirely or betting them at minimum while side bets are large.

Cover play matters. Counters who want to last more than a few sessions mix in losing-EV main bets, spread their side-bet action over multiple shoes, and avoid the obvious tells. Most don’t bother — the edges are small enough and the variance high enough that getting backed off after a few winning sessions is just part of the cost of doing business.

Legal Status and What Casinos Actually Do

Card counting is not illegal anywhere in the United States. It’s not cheating. You’re using public information (the cards that have been dealt) to make better decisions. The Nevada Supreme Court settled this for blackjack decades ago, and the same logic applies to baccarat side bets.

What casinos can do, and routinely do, is refuse your action. They’ll bar you from the side bet, reduce your maximum, increase the cut card depth so penetration drops, or just ask you to leave. None of that is a legal problem for the player, but it ends the play. Some jurisdictions outside the U.S. treat counting differently, and a few have prosecuted players who used electronic devices to track counts, but mental counting is universally legal.

The techniques in this article are well-published. They appear in books by Stanford Wong, James Grosjean, and Don Schlesinger. They show up in academic papers and on public gambling-math sites. None of this is a secret. If you’re learning baccarat side-bet counting expecting to find hidden knowledge, you won’t — the work is in the practice, the bankroll, and the discipline to walk away.

FAQ

Is Lucky 6 always the best baccarat side bet to count?
In most paytables, yes. Lucky 6 has the steepest count-to-edge curve because Banker totals of 6 depend heavily on small-card composition. If your local casino offers a 20:1 three-card-6 payout, the edge climbs even higher.

How deep does the shoe need to go before counting matters?
You typically need to be past the halfway point before any side-bet count moves meaningfully. The strongest edges show up in the last quarter of the shoe, which is why deep cut cards (50 cards or fewer remaining) are essential. A shoe cut at 100 cards is unplayable for this strategy.

Can I count Dragon Bonus and Lucky 6 at the same time?
Yes, but the two bets respond to different card compositions, so you’d need to track two separate running counts. Most players pick one and stay disciplined. Trying to track both adds error and slows your decisions, and errors eat edge fast.

Do online live-dealer baccarat games allow this?
The shoes are usually shuffled too often or cut too shallow to give a count any chance to develop. A few live-dealer tables do use eight-deck shoes with deeper penetration, but most online operators have read the same math and set the parameters to defeat the strategy.

How much bankroll does side-bet counting require?
The variance is much higher than blackjack counting because the payouts are larger and the playable hands are rarer. A reasonable rule of thumb is 1,000 to 2,000 units of your maximum side-bet wager, which is roughly double what serious blackjack counters target.

Baccarat side-bet counting is a real strategy with real edges, but it sits in a narrow corner of the game. If you want to dig deeper into the underlying probability math, the free explainers at EffortlessMath are a good starting point for the conditional-probability framework that drives every counting system. Gambling outcomes are uncertain; no strategy guarantees profit.