Sic Bo Complete Guide — Rules, Odds, Bets & House Edge

Sic Bo Complete Guide: Dice Rules, Bet Types, Probabilities, House Edge and Online Variants

Sic Bo is a three-dice casino game built around one simple event: three dice are rolled and the player bets on characteristics of the result. The apparent simplicity hides a large menu of bets with very different probabilities and house edges.

That makes Sic Bo a good example of why a casino guide should not talk about one game-level house edge as if every wager were equivalent.

A “Small” or “Big” wager can have a very different expected cost from a specific triple, a two-dice combination or a total bet. The interface may place all of those wagers on one table, but mathematically they are different products.

For NivaroBet, a strong Sic Bo page should explain the dice model, identify the common bet families, show how probabilities are derived and keep provider-specific payouts separate from generic rules.

The guiding editorial principle is simple: probability comes from the three-dice outcome space; expected return comes from probability plus the exact payout table.

The three-dice outcome space

With three fair six-sided dice, there are 6 × 6 × 6 = 216 ordered outcomes.

The word ordered matters.

A result of 1-2-3 can appear as:

1-2-3;

1-3-2;

2-1-3;

2-3-1;

3-1-2;

3-2-1.

Those are six distinct ordered outcomes even though the visible set of dice values is the same.

Probability calculations should therefore start from the 216 equally likely ordered combinations when the dice are fair and independent.

Totals range from 3 to 18

The lowest possible total is 3, created only by 1-1-1.

The highest is 18, created only by 6-6-6.

Middle totals can be formed in many more ways.

For example, total 10 can occur through many ordered combinations, while total 4 has only three: 1-1-2, 1-2-1 and 2-1-1.

This creates the familiar bell-shaped distribution around the central totals.

The payout table normally compensates rare totals with larger returns.

Big and Small

A common Sic Bo table offers “Big” and “Small.”

A widely used rule defines:

Small: total 4 through 10, excluding triples;
Big: total 11 through 17, excluding triples.

The triple exclusion is essential.

If the three dice show 4-4-4, the total is 12, but a conventional Big bet loses because any triple is excluded.

Likewise, 2-2-2 totals 6 but usually defeats Small.

Always read the exact game rules because providers can label or settle variants differently.

Why Big and Small are close to even-money but not 50/50

If triples were not excluded, the low and high ranges would be symmetric.

The triple rule removes six outcomes from the Big/Small winning space.

There are 216 total ordered outcomes and six triples.

Each of Big and Small wins on 105 of the 216 outcomes under the conventional rule.

That gives a win probability of 105/216, approximately 48.61%.

At an even-money payout, the house edge is approximately 2.78%.

This is one of the lower-house-edge common Sic Bo bets, although the exact payout and rule set should always be verified.

Odd and Even

Odd and Even typically work similarly to Big and Small.

The player bets on whether the total of the three dice is odd or even.

Conventional Sic Bo usually excludes triples from these wagers too.

Because there are equal numbers of odd and even totals across the full three-dice space and three odd triples plus three even triples, each side again has 105 winning outcomes under the common rule.

At 1:1, the house edge is therefore the same approximately 2.78%.

Do not assume every digital variant uses the triple exclusion unless the rules confirm it.

Single-number bets

A single-number bet selects one face, such as 5.

The payout can depend on how many dice show that number.

A common structure might pay:

one matching die: 1:1;
two matching dice: 2:1;
three matching dice: 3:1.

Other tables can use different returns.

The probability of at least one chosen face appearing among three fair dice is:

1 - (5/6)^3 = 91/216, approximately 42.13%.

But the expected value depends on the separate one-match, two-match and three-match payouts.

NivaroBet should therefore store the whole payout ladder.

Exactly one match

For a selected face, exactly one of the three dice must show that face.

Choose which die matches in 3 ways.

The other two dice each have 5 non-matching faces.

That gives 3 × 5 × 5 = 75 ordered outcomes.

Probability: 75/216, approximately 34.72%.

This is the most common winning state for a single-number bet.

Exactly two matches

Choose which one of the three dice is the non-matching die.

There are 3 positions.

That die can show any of the five other faces.

So there are 3 × 5 = 15 ordered outcomes.

Probability: 15/216, approximately 6.94%.

Three matches

All three dice must show the selected face.

There is exactly one ordered outcome.

Probability: 1/216, approximately 0.463%.

This very low probability is why a triple-related prize needs a higher payout to compensate.

The actual return still depends on the paytable.

Specific doubles

A specific double bet predicts that at least two dice will show a chosen number, for example double 4.

The common rule usually wins on both exactly two 4s and triple 4s, although some tables handle triples separately.

If both states count, the number of winning ordered outcomes is 15 + 1 = 16.

Probability: 16/216, about 7.41%.

If the table excludes triples from the double, the probability is 15/216.

That small rule difference changes the correct expected value.

Specific triples

A specific triple predicts all three dice will show one exact face.

Examples:

1-1-1;
2-2-2;
6-6-6.

Each specific triple has exactly one winning outcome out of 216.

Probability: 1/216, approximately 0.463%.

The headline payout can look huge, but it should be compared with the tiny probability.

A 150:1 payout, for example, is still below fair odds for a 1-in-216 event.

The exact house edge can be calculated from the actual return rule.

Any Triple

“Any Triple” wins if all three dice match, regardless of face.

There are six winning outcomes: one for each face.

Probability: 6/216 = 1/36, approximately 2.78%.

The payout is much lower than for a specific triple because the event is six times more likely.

A provider may label this bet “Any Triple,” “Any Triples” or another equivalent phrase.

NivaroBet should preserve the provider label and map it to a normalised bet type.

Two-dice combinations

A combination bet selects two different faces, such as 2 and 5.

The wager wins if at least one die shows 2 and at least one die shows 5.

The third die can be another 2, another 5 or any of the four remaining faces.

Count the outcomes carefully.

There are 30 ordered outcomes containing both selected numbers at least once.

Probability: 30/216, approximately 13.89%.

A common payout such as 5:1 produces a particular house edge, but providers can vary the return.

The combination should not be confused with predicting exactly two dice.

Total bets

A total bet predicts the sum of the three dice.

The probabilities vary sharply.

The number of ordered outcomes for totals 3 through 18 is symmetric:

3: 1

4: 3

5: 6

6: 10

7: 15

8: 21

9: 25

10: 27

11: 27

12: 25

13: 21

14: 15

15: 10

16: 6

17: 3

18: 1

These counts add to 216.

That table explains why 10 and 11 are common while 3 and 18 are extremely rare.

Total 4 example

Total 4 has three ordered outcomes:

1-1-2;

1-2-1;

2-1-1.

Probability: 3/216 = 1/72, approximately 1.389%.

If a table pays 50:1, the expected return can be calculated from that probability and the stated payout convention.

Do not compare “50:1” directly with another table that quotes total return as “51x” without normalising the terminology.

Total 10 example

Total 10 has 27 ordered outcomes.

Probability: 27/216 = 12.5%.

Because it is much more common than total 4, the payout is correspondingly lower.

This is why a Sic Bo total table looks uneven.

The different payout figures are attempting to price very different underlying probabilities.

The operator’s margin is embedded in the difference between fair odds and offered odds.

House edge belongs to the bet, not the table

A Sic Bo page that says “house edge: 2.78%” is incomplete.

That figure might describe only Big, Small, Odd or Even under one conventional rule.

A specific triple can have a much higher house edge.

A total bet can have another value.

A two-dice combination can differ again.

NivaroBet should therefore store house edge by bet family where it is verified or mathematically derived from a complete, current payout table.

Fair odds versus casino odds

If an event has probability p, fair profit odds are approximately:

(1 - p) / p to 1.

For a specific triple with p = 1/216, fair profit odds would be 215:1.

A casino paying 150:1 retains a substantial margin.

This does not mean the game is unfair in the regulatory sense.

It means the offered price is lower than actuarially fair odds, which is how casino games create a house edge.

The player should be able to see the payout before wagering.

Payout notation needs care

Three formats commonly appear:

1:1;
2x;
“pays 2 for 1.”

They can mean different things depending on the interface.

A 1:1 profit payout normally returns stake plus one unit of winnings.

A “2x total return” can mean the same thing.

A careless review can double-count or undercount the stake.

NivaroBet should normalise profit payout and total return separately in the data model.

Live Sic Bo and RNG Sic Bo

Online Sic Bo can be offered as:

RNG software;
live dealer or live studio game;
automated physical-dice system.

The visual interface can be similar while the result-generation method differs.

For RNG Sic Bo, the software determines the outcome under the certified game model.

For live Sic Bo, physical dice or equipment produce the result and cameras show the event.

A review should label the implementation rather than assuming all online versions use RNG.

Live dealer presentation does not change the bet mathematics

If the live table uses three fair dice and the same settlement rules, the combinatorial probabilities remain the same as the corresponding digital game.

What can differ:

betting speed;
payout table;
side bets;
limits;
interface;
result verification;
studio procedure.

The house edge therefore needs the exact table rules even when the dice model is familiar.

Dice-shaking devices can be automated

Some live Sic Bo products use transparent dice shakers or automated dice equipment.

The device may tumble the dice rather than a human dealer throwing them by hand.

That does not by itself change the mathematical model.

The relevant questions are equipment integrity, result recognition, game rules and regulatory testing.

NivaroBet should not imply that an automated shaker is less random merely because it is mechanical.

Game speed changes exposure

Sic Bo can settle quickly.

A live table can run many rounds per hour.

An RNG version can be even faster.

Expected cost per hour depends not only on house edge but also on:

stake;
number of bets per round;
rounds per hour.

A player placing five simultaneous side bets can have much higher turnover than someone making one Small bet.

This is why “low house edge” does not automatically mean low spending.

Multiple simultaneous bets can overlap

Sic Bo allows several bets on one roll.

For example, a player could bet Small, a specific number and a total.

Those wagers are not independent because the same dice result settles all of them.

A winning result on one bet can coincide with losses on others.

Do not add the individual win probabilities and call the total a chance of profit.

Portfolio-style combinations require joint-outcome analysis.

A consumer guide should avoid suggesting hedging systems create guaranteed profit.

Betting both Big and Small does not remove the house edge

A player might think betting both sides guarantees one winner.

Triples can cause both to lose under the conventional rules.

Even without that issue, paying commissions or non-even odds can preserve the casino advantage.

The two opposite bets also increase total money staked.

A “cover the board” strategy should therefore be analysed by total expected value, not by the psychological comfort of frequent winning tickets.

Martingale does not change expected value

Doubling the stake after a loss does not change the probability of the next dice result.

The progression increases exposure.

Table limits, bankroll limits and losing streaks prevent indefinite doubling.

The expected percentage loss remains tied to the bet rules.

NivaroBet should never present a staking system as a way to overcome the house edge.

It can explain the mechanics and risk.

Hot dice and cold dice are not predictive

Recent totals do not change the next fair roll.

A sequence of Big results does not make Small due.

A recent triple does not make another triple less possible.

Historical roadmaps can be useful visual records but should not be described as forecasting tools.

This is the same independence principle found in roulette and other repeated random games.

Electronic history boards can create false pattern confidence

Live Sic Bo tables often display recent outcomes in charts.

Those charts help users see what happened.

They do not alter the probabilities.

A reviewer can describe the feature as history visualisation.

It should not call it a strategy aid unless the game rules contain a genuinely state-dependent mechanism, which conventional Sic Bo does not.

Side bets vary by provider

Modern Sic Bo titles can add:

four-number combinations;
special totals;
domino-style bets;
pair-plus-total combinations;
multiplier features.

These are not universal.

NivaroBet should not create one “Sic Bo odds” table that mixes bets from several providers.

Each side bet needs:

rule;
probability;
payout;
house edge where calculated;
version and date.

That preserves search value without inventing standardisation.

Random multipliers create another layer

Some live and RNG titles attach random multipliers to selected bets.

A multiplier feature changes the payout distribution.

The headline base payout no longer tells the full story.

The provider should disclose the game RTP or sufficient rule information.

NivaroBet should not calculate an exact house edge from the base table if random multipliers materially change the return and their distribution is unknown.

Maximum payout caps matter

A table can cap the total win per round or per bet type.

That can reduce the effective value of high-payout combinations for large stakes.

The cap should be shown before a user places a wager.

NivaroBet should store it with the table-limit evidence.

Do not publish unlimited-looking multipliers when the rules impose a monetary maximum.

The rules should explain the likelihood of winning

Great Britain’s RTS 3 requires applicable rules and information about the likelihood of winning to be available before gambling.

See UK Gambling Commission — RTS 3.

For Sic Bo, the paytable and bet definitions are essential.

Other jurisdictions can use different technical standards.

NivaroBet should attach the correct market source rather than making UK rules global.

Game testing matters

Regulated online games are normally subject to technical testing and performance monitoring under the relevant framework.

In Great Britain, the Gambling Commission explains that online gambling games must be tested in accordance with its testing strategy.

See UK Gambling Commission — How we ensure online gambling games are fair.

A provider logo alone does not prove which exact build was tested.

Title-level evidence is stronger.

Example: Big at even money

Under conventional rules:

105 winning outcomes;
111 losing outcomes including triples.

For a 1-unit wager paying 1 unit profit on a win:

Expected profit =

105/216 × 1

111/216 × 1

= -6/216

≈ -2.78%.

That is the house edge for that exact structure.

If a provider changes the triple rule or payout, recalculate.

Example: specific triple at 150:1

Probability of a chosen triple = 1/216.

A 1-unit wager wins 150 units profit on one outcome and loses 1 unit on 215 outcomes.

Expected profit =

1/216 × 150

215/216 × 1

= -65/216

≈ -30.09%.

That demonstrates why a huge-looking payout can still carry a large house edge.

Always verify the provider’s actual specific-triple payout before quoting the number.

Example: Any Triple

Probability = 6/216 = 1/36.

If the table pays 24:1 profit:

Expected profit =

1/36 × 24

35/36 × 1

= -11/36

≈ -30.56%.

Again, the exact return depends on the offered payout.

This is why NivaroBet should never show “Sic Bo house edge” as a single number.

Five-minute Sic Bo audit workflow

Confirm provider and title.
Record RNG, live or automated implementation.
Open full rules and payout table.
Confirm Big/Small and Odd/Even triple treatment.
Record single-number payout ladder.
Record doubles, triples and combinations.
Record total-bet payouts.
Identify side bets and random multipliers.
Record stake limits and maximum win.
Capture RTP or house-edge information where published.
Record version, market and checked date.
Keep calculated values tied to the exact paytable.

This produces a page that another reviewer can reproduce.

NivaroBet Sic Bo data model

Game identity

Provider, title, version, market and implementation.

Core rules

Three dice, Big/Small range, triple exclusions and settlement.

Bets

Bet family, event definition, probability and payout.

Mathematics

House edge, RTP and multiplier contribution where supported.

Limits

Minimum stake, maximum stake and payout cap.

Evidence

Rules URL, provider source, casino source and checked date.

A structured model avoids one generic paragraph being reused across dozens of versions.

What NivaroBet should never claim

Do not say all Sic Bo bets have a 2.78% house edge.

Do not ignore triple exclusions.

Do not treat live history boards as predictive.

Do not present staking systems as mathematical advantages.

Do not combine side bets from different providers into one standard table.

Do not infer random-multiplier RTP from base payouts alone.

Do not publish a house-edge calculation without the exact payout convention.

Do not assume every live and RNG version has identical rules.

Sic Bo is useful to compare with games that price risk in different ways. Blackjack complete guide covers a decision-based table game where player choices can change expected value, while Roulette complete guide shows another fixed-outcome betting table where each wager has its own probability and payout structure.

Those comparisons should not be used to rank one game as “better.” They help explain why a single casino-level house-edge number is not meaningful across different products.

The bottom line

Sic Bo is simple at the dice level and complex at the betting level.

Three fair dice create 216 ordered outcomes. Every wager selects a different subset of those outcomes, and the payout table determines the house edge.

Big, Small, Odd and Even can be relatively low-edge under common rules. Specific triples and many side bets can be far more expensive.

For NivaroBet, the right page is not a generic list of wagers. It is a version-specific mathematical audit that preserves rules, probability, payout, house edge, implementation, market and evidence date.

Related guides and tools

Sic Bo Complete Guide: Dice Rules, Bet Types, Probabilities, House Edge and Online Variants

The three-dice outcome space

Totals range from 3 to 18

Big and Small

Why Big and Small are close to even-money but not 50/50

Odd and Even

Single-number bets

Exactly one match

Exactly two matches

Three matches

Specific doubles

Specific triples

Any Triple

Two-dice combinations

Total bets

Total 4 example

Total 10 example

House edge belongs to the bet, not the table

Fair odds versus casino odds

Payout notation needs care

Live Sic Bo and RNG Sic Bo

Live dealer presentation does not change the bet mathematics

Dice-shaking devices can be automated

Game speed changes exposure

Multiple simultaneous bets can overlap

Betting both Big and Small does not remove the house edge

Martingale does not change expected value

Hot dice and cold dice are not predictive

Electronic history boards can create false pattern confidence

Side bets vary by provider

Random multipliers create another layer

Maximum payout caps matter

The rules should explain the likelihood of winning

Game testing matters

Example: Big at even money

Example: specific triple at 150:1

Example: Any Triple

Five-minute Sic Bo audit workflow

NivaroBet Sic Bo data model

Game identity

Core rules

Bets

Mathematics

Limits

Evidence

What NivaroBet should never claim

Related NivaroBet game guides

The bottom line

Primary sources