Baccarat streaks · Scoreboards · Conditional probability · Gambler’s fallacy
A baccarat run is a sequence of Banker or Player wins. Runs are real after they occur, but a scoreboard does not tell you that the next hand will continue or reverse one. Trend betting converts a description of past cards into a forecast the evidence does not support.
The next hand is not exactly 50–50
In a conventional eight-deck game, Banker wins about 45.86%, Player 44.62%, and Tie 9.52%. Ignoring ties, Banker wins about 50.68% of decisive hands. That small structural difference comes from the drawing rules—not from the recent streak.
How rare is a 15-hand run?
Under a simplified fair 50–50 model with ties removed, 15 specified Banker wins in a fixed 15-hand window has probability (1/2)15, or 1 in 32,768. A run of either all Banker or all Player in that fixed window is 2/32,768, or 1 in 16,384.
That is not the same as “one run every X shoes.” A shoe contains many overlapping 15-hand windows, baccarat is not exactly 50–50, ties complicate what counts as a broken run, and cards are dealt without replacement. A within-shoe calculation requires the cut, burn, number of hands, and run definition.
Does a physical shoe make runs predictable?
No. The order in a shuffled shoe is fixed before dealing, but hidden. “Predetermined” does not mean known. Card removal changes probabilities slightly, yet a simple Banker/Player scoreboard omits the ranks and composition needed to quantify that change. An RNG game may simulate a shoe or generate rounds under another approved model; do not assume every online game reshuffles after every hand.
Card removal is not the same as trend reading
Because cards leave a physical shoe, the exact probabilities are conditional on the unseen composition that remains. That observation is mathematically true but does not validate a Banker/Player road. A scoreboard records only the winning side and sometimes the margin or pairs; it does not reveal every rank removed. Even a full composition calculation would estimate a small change in expectation, not declare that a side must win.
What the baccarat roads actually show
The bead plate and Big Road arrange completed Banker, Player, and Tie results. Derived roads such as Big Eye Boy, Small Road, and Cockroach Pig use display rules based on patterns in the Big Road. Their marks do not directly mean Banker or Player, and the algorithm does not add information about unseen cards. They are recordkeeping and presentation systems, not evidence that the next hand has become favorable.
Three common streak systems
| System | Rule | Failure |
|---|---|---|
| Follow | Bet the side that just won | The observed run can end immediately |
| Fade | Bet the opposite side because reversal is due | The run can continue |
| Wait for confirmation | Enter after two or three same-side results | Classifies the run after part of it has already occurred |
Flat betting keeps exposure stable; loss progressions make the cost of a wrong streak call grow. Neither changes the underlying Banker, Player, or Tie expectation.
A useful test for a streak claim
- Write the entry rule before viewing the shoe.
- Specify how ties are treated and when a run begins or ends.
- Use the same rule on every eligible shoe rather than selecting memorable examples.
- Track total dollars wagered, commission, and every skipped hand.
- Compare results with flat Banker or Player betting and with random entries.
- Test on new shoes that were not used to invent the rule.
A short winning record can show that the sequence occurred; it cannot show that the rule caused it. If the claimed advantage disappears when the definition is frozen or the sample is extended, the original pattern was fitted to noise.
What scoreboards are good for
- Checking that completed results were recorded correctly.
- Describing the shoe and discussing table culture.
- Auditing whether you changed sides or stakes impulsively.
- Studying how easily random clusters create convincing stories.
Use the baccarat odds table for the actual prices and the craps gambler’s-fallacy guide for the same independent-outcome mistake in dice language.
Audit a short follow-the-run sequence
Suppose six decisive results are Banker, Banker, Player, Player, Player, Banker. A $10 follow-the-last-winner system starts on hand two. It wins $9.50 on the second Banker after 5% commission, loses $10 when hand three changes to Player, wins $10 on each of hands four and five, then loses $10 when hand six changes back to Banker.
The five wagers finish at +$9.50: $29.50 of profit on three wins minus $20 on two losses. That outcome can feel like proof because the system captured the middle of a three-Player run. Reverse the timing of the changes or add a loss progression and the same rule can produce a very different result.
The ledger shows what the scoreboard cannot: the entry rule pays only after a streak has already begun and loses at each reversal. Evaluate every eligible hand, deduct Banker commission, state how ties are handled, and keep the next shoe out of the sample used to choose the rule.
State what happens on a tie before following a run
For the sample follow-the-last-decisive-winner rule, insert a tie after the second Banker. A $10 Banker wager pushes: it returns $10 and earns no winning profit or standard commission. Keeping the last decisive side as Banker means the next Player still produces the sample’s $10 loss. The tie adds one wagered round to the record even though it does not change that sample’s net result.
A rule that resets after a tie would make a different next entry. Neither choice is supplied by the road’s appearance, so state it beforehand. Keep a separate count of physical hands, decisive results and accepted wagers. Ignore ties only for the stated run definition; do not delete their action from the betting ledger. Adding a Tie side bet also changes both cost and settlement and requires its own record.
