The two-dice combination bet in Sic Bo wins on 30 of the 216 possible rolls. Paid at 5 to 1 it carries a 16.6667% house edge; paid at 6 to 1 on a different legal paytable it carries 2.7778%. One increment of payout, a six-fold difference in price, and nothing about the dice has changed. Learning how to play Sic Bo is straightforward; learning what a given table charges for it is the part that takes reading.
What Sic Bo is
Sic Bo is a Chinese casino dice game played with three dice. The name is 骰寶, usually rendered as “precious dice”; the same game appears as tai sai or dai siu — “big small” — and as hi-lo in the Philippines. It is a staple in Macau and appears in many Western casinos, and licensed UK casinos have offered it since 2002.
The closest relative on a Western floor is chuck-a-luck, the birdcage game where a dealer rotates a wire hourglass and the main wager is a single number paid 1 to 1, 2 to 1 or 3 to 1. Sic Bo keeps that single-number idea and adds totals, pairs, triples and two-dice combinations on one large illuminated layout. The wider arc from dice to casino floor is in CasinoDrop’s history of gambling guide.
The contrast with craps is worth holding on to, because it removes most of the complexity people expect. Craps uses two dice and several of its bets stay open across rolls, waiting on a point. Sic Bo uses three dice and every wager is an immediate win or loss on the shake in front of you. There is no banker hand, no point number, and no working bet. A chip still sitting on a box when the next round opens is a new wager.
The 216 outcomes everything is priced from
Three six-sided dice give 6 × 6 × 6 = 216 ordered results, and every price on the layout is a count of those 216. The totals cluster hard in the middle, which is what makes the outer totals look attractive and cost so much.
| Total | Ways | Probability | Common payout | House edge |
|---|---|---|---|---|
| 4 or 17 | 3 | 1.39% | 60:1 | 15.28% |
| 5 or 16 | 6 | 2.78% | 30:1 | 13.89% |
| 6 or 15 | 10 | 4.63% | 17:1 | 16.67% |
| 7 or 14 | 15 | 6.94% | 12:1 | 9.72% |
| 8 or 13 | 21 | 9.72% | 8:1 | 12.50% |
| 9, 10, 11 or 12 | 25–27 | 11.57–12.50% | 6:1 | 18.98% |
Totals run 4 through 17 rather than 3 through 18, because three 1s and three 6s are already sold on the triple boxes. The most likely totals are the worst-priced: 9 through 12 come up most often and carry the highest edge on the board at 18.98%, while the mid-range 7 or 14 at 12 to 1 is the cheapest total on offer at 9.72%. Payouts rise as the ways fall, but never quite enough.
How a round is played
Live-dealer versions keep the same loop with a studio dealer, a sealed tumbler and a digital grid highlighting the winners. History boards showing recent Small, Big and Triple results are decoration — each shake is independent of the last, and a run of Smalls does not make a Big more likely on the next one.
Every bet, priced
The layout is busy because it is several games sharing one shake: single numbers, two-number combinations, three-dice totals, triples, pairs, and Big and Small. Enumerating all 216 outcomes against a standard schedule gives the spread.
Small and Big are the cheapest bets on the board by a wide margin, and understanding why explains the whole game. Small wins on totals 4 through 10, Big on 11 through 17, and both pay even money — but both lose when all three dice match.
those 210 split evenly: 105 Small, 105 Big
win probability = 105 ÷ 216 = 48.6111%
house edge = (111 − 105) ÷ 216 = 1/36 = 2.7778%
Where a table offers Odd and Even, they use the same even-money price and the same triple-loses rule, so treat them as the same product rather than a second way to approach the dice. The single-number bet, sometimes called one of a kind, wins if a chosen face appears on at least one die: 75 ways with one die showing it, 15 with two and 1 with all three, so 91 of 216 shakes hit. Paid 1 to 1, 2 to 1 and 3 to 1 respectively, that is a 7.8704% edge — and where a table lifts the three-of-a-kind leg to 12 to 1, it falls to 3.7037%.
The paytable is the product
Sic Bo is unusual among casino games in how far its prices move between jurisdictions. Pennsylvania publishes two legal schedules in its table-game rules, and the difference between them is not decorative. The same enumeration run against each payout gives this:
| Bet | Ways in 216 | Cheapest seen | Dearest seen |
|---|---|---|---|
| Two-dice combination | 30 | 2.78% at 6:1 | 16.67% at 5:1 |
| Specific double | 16 | 11.11% at 11:1 | 33.33% at 8:1 |
| Specific triple | 1 | 16.20% at 180:1 | 30.09% at 150:1 |
| Any triple | 6 | 13.89% at 30:1 | 30.56% at 24:1 |
| Single number | 91 | 3.70% with 12:1 trips | 7.87% with 3:1 trips |
| Small or Big | 105 | 2.78% at 1:1 | 2.78% — does not vary |
The bottom row is the practical takeaway. Small and Big are the only bets whose price does not move between paytables, because they pay even money everywhere and their edge comes entirely from the triple rule. Every other bet on the layout is a negotiation you are not part of, and the same chip can cost between three and eleven times as much depending on which room you are standing in.
One consequence is worth stating plainly: published Sic Bo house-edge tables are only as good as the paytable they assume, and they are not always right. Ontario’s PlaySmart odds page gives the total 4 or 17 bet a house edge of “approximately 12.43%” at 60 to 1. Enumerated over the 216 outcomes, three winning ways paid at 60 to 1 gives (3 × 60 − 213) ÷ 216 = 15.2778%. That is the figure Wizard of Odds publishes, and it is what our own enumeration returns. Check the arithmetic against the paytable in front of you rather than a table on a website, including this one.
Live Super Sic Bo adds a further wrinkle. Evolution’s version applies random multipliers to selected bets after betting closes, on a harsher base schedule than a standard board. The multipliers make the advertised maximums much larger; they do not make the base bets cheaper, and the published rules are where the actual prices are.
Does the enumeration hold up in practice?
Everything above rests on one assumption: that the dice are fair and all 216 outcomes are equally likely. CasinoDrop does not track a live Sic Bo table, so we cannot show that assumption holding on our own tape. What we can show is the same counting method applied to a game we do record — single-zero roulette, where 37 pockets give every bet an identical 2.7027% edge, and where we log every settled spin on the original Lightning Roulette table.
All ten groupings we checked across 5,600 recorded spins landed inside the range chance produces at that sample size, including the zero at 2.61% against a designed 2.70%. That says nothing about any Sic Bo shaker. It says that counting a finite outcome space and comparing it against a real table is a method that survives contact with reality, and it is the method behind every figure in this guide.
- Which edge you pay. Small, Big, Odd and Even sit at 2.78% while the totals reach 18.98%.
- Which table you sit at, once you can read a paytable — the same bet ranges from 2.78% to 16.67%.
- Exposure per shake, by not covering families you have no price for.
- Volatility, since a triple at 180 to 1 and Small at even money are different games on one layout.
- The 216 outcomes. Every shake is independent and no result is ever due.
- The sign of the expected value. Every bet on the board is priced below true odds.
- The history board. A run of Smalls tells you what happened, not what is next.
- The triple rule, which is where the entire edge on the even-money bets comes from.
We do not track a live Sic Bo table, so what we can hand you is the method rather than a Sic Bo feed. The Lightning Roulette board carries every settled spin we record against the 37-pocket wheel — the same finite-outcome arithmetic used above, and the one place on the site you can watch it hold up round by round against a real table.
Open the Lightning Roulette tracker →
- CasinoDrop does not track Sic Bo. Every Sic Bo figure here is enumerated arithmetic over the 216-outcome space, not a measurement of a real table, and none of it is evidence about any particular shaker.
- The enumeration assumes fair dice and equally likely outcomes. That is what regulators test for, and it is an assumption here rather than something we verified.
- Every house edge is only as good as the payout beside it. The paytable ranges in this guide are the schedules published in Pennsylvania’s rules and the ones commonly documented for Macau, Atlantic City and Australia; your table may use none of them. Read the board in front of you.
- The “cheapest and dearest seen” columns are not a survey. They are the range across the published schedules we could source, so a worse or better price than either end may well exist somewhere.
- The roulette comparison checks the method, not Sic Bo. It shows enumeration matching what a real table produces on a game we record, and says nothing about dice.
- That roulette sample is the original Lightning Roulette table only, 5,600 spins over about three days. It describes those spins, not the game’s lifetime or any other table of it.
- House edge is a long-run average per unit staked. It does not describe a session, and the spread around it over an evening is far wider than the edge itself.
- Nothing here predicts a shake. Each is independent of the last, and a history board showing a run of Smalls does not make a Big more likely next.
FAQs
Place chips on the boxes you want before the dealer shakes three dice in a covered container. When the cover comes off, every bet on the table is settled immediately against that one result. There is no point number and no bet that waits for a later roll, so a chip left on a box when the next round opens is a new wager.
Small or Big, at a 2.7778% house edge. Three dice give 216 equally likely outcomes; six are triples, which lose both bets, and the remaining 210 split evenly into 105 Small and 105 Big. Paying even money on a 48.6111% chance gives an edge of exactly 1/36. Where a table offers Odd and Even, they are the same product at the same price.
On a standard board it is the totals 9 through 12 at 6 to 1, carrying an 18.9815% house edge. Those are the most likely totals on the layout, which is what makes them look reasonable. On harsher paytables a specific double at 8 to 1 reaches 33.33% and any triple at 24 to 1 reaches 30.56%, so the worst bet depends on which schedule the table uses.
Because the payouts do. Pennsylvania alone publishes two legal schedules. The two-dice combination wins on 30 of 216 rolls either way, but at 6 to 1 that is a 2.7778% edge and at 5 to 1 it is 16.6667%. A specific double ranges from 11.11% at 11 to 1 to 33.33% at 8 to 1. Only Small and Big hold their price everywhere.
Because that rule is the entire house edge on those bets. Without it, 105 outcomes would win and 105 would lose and the bet would be an exactly fair coin. Removing the six triples from the winning side turns a 50% chance into 48.6111%, and paying even money on that gives the house 1/36, or 2.7778%.
Out of 216 outcomes, totals 10 and 11 come up 27 ways each, 9 and 12 come up 25, 8 and 13 come up 21, 7 and 14 come up 15, 6 and 15 come up 10, 5 and 16 come up 6, and 4 and 17 come up 3. Totals 3 and 18 are not sold as totals because they are already covered by the triple boxes.
You pick a face from 1 to 6 and win if it appears on at least one die, with the payout scaling to how many dice show it. Of 216 outcomes, 75 show your face once, 15 twice and 1 three times, so 91 in 216 win something. Paid 1 to 1, 2 to 1 and 3 to 1 the house edge is 7.8704%; where a table pays 12 to 1 on all three dice it falls to 3.7037%.
Its best bet is competitive and the rest of the board is not. Small or Big at 2.7778% sits just behind single-zero roulette at 2.7027% and behind the craps pass line at 1.4141%. But most of a Sic Bo layout runs from 9% to 19% on a standard board, which is well above anything on a craps or roulette table except the proposition bets.
No. Each shake is independent of the last, so a run of Smalls does not make a Big more likely on the next one. The boards showing recent Small, Big and Triple results are decoration. The 216 outcomes are the same on every shake regardless of what the previous ones did.
Evolution’s live version, which applies random multipliers to selected bets after betting closes. The multipliers raise the advertised maximum payouts substantially, but they sit on a harsher base schedule than a standard board — its specific triple pays 150 to 1 rather than 180 to 1, for instance. The published rules carry the actual prices.
Craps uses two dice and 36 outcomes, and several of its bets stay open across rolls waiting on a point. Sic Bo uses three dice and 216 outcomes, and every bet settles on the shake you just watched. Craps also has the cheaper board: its pass line is 1.4141% and free odds carries no edge at all, while Sic Bo’s cheapest bet is 2.7778%.
Playing within a limit
Every box on a Sic Bo layout is priced against you, and the game settles a round every shake with no waiting bets, so turnover builds faster than the pace feels. Even the cheapest bet on the board keeps 2.78 pence of every pound over the long run, and most of the layout keeps far more. Decide the loss limit before the first shake, and treat the illuminated longshots as the most expensive entertainment on the table rather than the best value. CasinoDrop’s responsible gambling guide covers deposit caps, timeouts and self-exclusion, and support is listed at GambleAware.
- Everything comes from 216 outcomes. Six of them are triples. Every price on the layout is a count of the rest, and you can rebuild any figure here yourself.
- Small and Big are the only stable price on the board. 2.7778% everywhere, because they pay even money universally and their edge is entirely the triple-loses rule.
- The paytable is the product. The same two-dice combination costs 2.78% at 6 to 1 and 16.67% at 5 to 1. A specific double ranges from 11.11% to 33.33%.
- The likeliest totals are the worst-priced. Totals 9 to 12 come up most often and carry the highest edge on a standard board at 18.98%; the mid-range 7 or 14 at 12 to 1 is the cheapest total at 9.72%.
- Published edge tables can be wrong. Ontario’s official odds page states 12.43% for the total 4 or 17 at 60 to 1; enumerated over 216 outcomes it is 15.2778%.
- CasinoDrop Lightning Roulette tracker — first-party, the 5,600 recorded spins used to check the enumeration method; original table, 2026-08-15 to 2026-08-18
- Lightning Roulette round feed — first-party, the JSON behind the observed grouping shares
- 58 Pa. Code § 625a.3 — primary, Pennsylvania Code: Sic Bo wagers and the two legal payout schedules
- 58 Pa. Code § 625a.5 — primary, Pennsylvania Code: procedures for shaking and revealing the dice
- 58 Pa. Code § 625a.6 — primary, Pennsylvania Code: payout odds and settlement
- Super Sic Bo game rules (PDF) — primary, Evolution help sheet hosted by Loto-Québec: the base paytable and the random multipliers
- Evolution launches Super Sic Bo — primary, Evolution: the live multiplier format
- Sic Bo analysis — reported, Wizard of Odds: independent house-edge figures across paytables that the enumeration here reproduces
- House edge summary — reported, Wizard of Odds: comparison figures for other casino games
- Sic Bo: how to play — primary, PlaySmart, Ontario Lottery and Gaming: table procedure and bet families
- Sic Bo: odds — primary, PlaySmart, Ontario Lottery and Gaming: the source of the 12.43% figure for total 4 or 17 that this guide’s enumeration puts at 15.2778%
- GambleAware — primary, safer-gambling information and support
