How to Read Baccarat Roads: Bead Plate, Big Road, Big Eye Boy, Small Road and Cockroach Pig

2026-09-17

The scoreboard next to a baccarat table shows five “roads”. They look like a code, but there are really only two kinds: roads that record what happened (Bead Plate, Big Road) and roads that are calculated from the Big Road (Big Eye Boy, Small Road, Cockroach Pig).

The short answer

• The Bead Plate logs every hand, ties included. The Big Road stacks consecutive wins for the same side into one column.
• Big Eye Boy, Small Road and Cockroach Pig don’t show Banker or Player at all. They show whether the Big Road is repeating its shape: red = it matches an earlier column, blue = it doesn’t.
• We rebuilt all three derived roads on 1.8 million real Evolution hands. Red appears 50.02%, 49.93% and 49.93% of the time. After three reds in a row, the next mark is red 49.96%, 49.88% and 50.17% of the time.
• Betting on “whichever side makes the road red” returns −1.11% to −1.28% per bet, slightly worse than simply betting Banker every hand (−1.06%).

Roads are a genuinely useful way to keep records. They carry no information about the next hand. Here’s how each one is drawn and read.

One real shoe from our dataset (first 36 hands). All five roads record the same results.

The five roads at a glance

Bead Plate: the raw log

Start at the top-left corner and fill downward, six cells per column, then move one column right. One circle per hand: red for Banker, blue for Player, green for Tie.

The Bead Plate never drops information. Ties get their own cell, so it’s the fastest way to count ties in a shoe. What it can’t show you is streaks, which is why the Big Road exists.

Big Road: one column per streak

The Big Road only draws Banker and Player, using three rules:

1. Same side again → next cell down in the same column. A different side starts a new column to the right, at the top row.
2. Ties don’t take a cell. Draw a green slash through the previous circle; some tables add a number for multiple ties.
3. If a column reaches the sixth row and the streak continues, turn right along the bottom row. This is called the “dragon tail”.

So every Big Road column is a streak, and its length is the streak length. How often does each length show up in real play?

Based on 800,870 complete columns. The last column in each shoe may be cut off, so it’s excluded.

The pattern is simple: each extra cell roughly halves the frequency, exactly what you’d expect when every hand is independent and Banker vs Player is close to 50/50. Columns of 6+ make up only about 2.9% of columns, but a shoe averages 30 columns. In our data, 61% of shoes contain at least one streak of 6, and 20% contain a streak of 8. Long streaks are routine. For the full maths, see How common are long streaks, really.

How the derived roads are calculated

Big Eye Boy, Small Road and Cockroach Pig follow the same rules. The only difference is how many columns back they look:

• Big Eye Boy: 1 column
• Small Road: 2 columns
• Cockroach Pig: 3 columns

Call that gap k (Big Eye Boy k=1, Small Road k=2, Cockroach Pig k=3). Each time a result is added to the Big Road, each derived road gets a mark:

Case 1: the result starts a new Big Road column Compare the previous column with the column k further to its left. Same length → red. Different length → blue.

Case 2: the result extends the current column Look at the cell in the same row, k columns to the left:

• That cell has a result → red
• That cell is empty but the cell directly above it has a result (that column ended exactly here) → blue
• That cell and the one above it are both empty → red

When a road starts: each road needs enough columns to compare. Big Eye Boy starts at the second hand of the Big Road’s 2nd column (or the first hand of the 3rd column, if the 2nd column is a single hand). Small Road starts the same way from the 3rd column, Cockroach Pig from the 4th. That’s why the derived roads are blank for the first few hands of a shoe.

Derived road marks are laid out like the Big Road: same colour goes down the column, a colour change starts a new column.

What red and blue actually mean

Red is not Banker and blue is not Player. Red means the Big Road’s shape is repeating (consecutive columns of equal length, or streaks breaking at the same row). Blue means it isn’t. When players say a shoe looks “orderly”, they mean lots of red.

One property matters a lot here: for any position, a Banker result and a Player result always produce opposite colours on the same derived road. So “I want Big Eye Boy to go red” always translates into betting Banker or betting Player. That lets us test the idea directly.

The test: do derived roads help you pick a side?

We rebuilt the Big Road and all three derived roads, hand by hand, for 27,510 real shoes, and checked three things.

1. Does either colour show up more often?

All three are a coin flip.

2. After three marks of the same colour, does it continue?

Every cell is based on more than 170,000 cases, and every result sits at 50%. The colours don’t form streaks, and they’re never “due” to switch.

3. What if you always bet the side that turns the road red?

Ties push, and Banker wins pay 0.95 after the 5% commission:

Following a derived road lands between always-Banker and always-Player. In other words, it just switches between the two sides at random, adding nothing. The differences are within statistical noise.

So what are roads good for?

Roads do have three real uses:

• Record-keeping. One glance tells you how the shoe has run and how many ties there were.
• Sanity-checking a table. Over large samples, Banker, Player and Tie should sit near 45.86% / 44.62% / 9.52%. That’s exactly how we validated this dataset.
• Discipline. If you’ve decided in advance to watch rather than chase, the road makes it obvious when you start wanting to chase.

What they can’t do is tell you the next hand. Each hand is dealt from the cards left in the shoe, and marks on a scoreboard don’t change those cards.

Try any road in the road-reading engine on our homepage to see how common that shape is in random shoes, or put a pattern you believe in through the pattern tester to see whether it actually beats chance over a large sample.

Data and method

• Source: Evolution live baccarat, 107 tables, September 3–11, 2026: 27,510 shoes, 1,801,127 hands (171,839 ties).
• Excluded: the Always 8 Baccarat and Always 9 Baccarat tables (different rules that change Banker/Player odds), plus shoes flagged as incompletely recorded.
• Validation: after exclusions, Banker 45.85%, Player 44.61%, Tie 9.54%. All within 0.03 percentage points of the exact 8-deck values (45.86% / 44.62% / 9.52%).
• Road reconstruction: Big Road and derived roads rebuilt programmatically with the standard rules described above. Column lengths use logical columns, so the dragon-tail display doesn’t affect the calculation.
• Expected value: 5% commission on Banker wins, ties push, no side bets.

This site provides probability calculations and statistical analysis only, not betting advice. 18+ only. If gambling is affecting your life, please reach out to a support service early.

Keep reading: How common are long streaks · Real probabilities and the house edge · Why staking plans cannot work · The third card rule