Are Craps Dice Random?

Yes. In a properly run casino craps game, the two dice are intended to produce outcomes that are effectively random for betting purposes. Each die has six faces, so two fair dice create 36 equally likely ordered combinations. Those combinations are what give totals such as 7, 6, 8, 2, and 12 their familiar probabilities.

The casino does not rely on wishful thinking to make the dice random. Physical dice specifications, table procedures, supervision, surveillance, dice changes, and rules for valid throws are all part of protecting the game from manipulation. None of those controls guarantee that every short sequence will “look random.” Random play naturally includes streaks, repeats, long runs without a number, and clusters that can feel suspicious.

Two dice create 36 combinations, not 11 equally likely totals

The first misconception to remove is that totals 2 through 12 have equal chances. They do not.

Think of the dice as ordered: a 1 on die A and a 6 on die B is a different physical combination from 6 on die A and 1 on die B. There are 6 × 6 = 36 possible ordered results.

Seven is the most common total because it can be made six ways: 1-6, 2-5, 3-4, 4-3, 5-2, and 6-1. Two and twelve each have only one combination.

This distribution is the foundation of craps odds and the reason different bets carry different probabilities. The dice combinations page shows the combinations in more detail.

Casino dice are designed for controlled, inspectable use

Casino craps dice are not ordinary board-game dice. Properties use precision dice and formal handling procedures because the game depends on the dice behaving consistently and because tampering with a die could change the risk of the table.

Operational controls vary by jurisdiction and house procedure, but common protections include:

• controlled storage and issue of dice;
• inspection before or during use;
• matching dice sets;
• serial or identifying marks where required;
• transparent material that makes some forms of loading easier to detect;
• square, sharp edges intended for casino table use;
• scheduled or event-driven dice replacement;
• supervision of dice leaving the table surface or player control.

The goal is not to make every throw mechanically identical. It is the opposite: the casino wants the dice to be legitimate randomizing devices while preventing substitution, weighting, shaving, loading, or other interference.

The back-wall rule is a game-protection procedure

At a traditional craps table, the shooter is normally required to throw both dice so they travel across the table and strike the far wall. The wall is covered with a textured or pyramidal rubber surface designed to disturb a controlled landing.

That rule serves several practical purposes. It makes extremely soft “place the dice” throws harder, creates a visible standard for dealers and supervisors, and helps distinguish a proper throw from a suspicious attempt to limit dice movement.

A throw that fails to reach the wall is not automatically proof of cheating. Players make weak throws. Dice hit chips. One die may stop short. The crew applies the property’s valid-roll procedure and may warn the shooter, call no roll where appropriate, or require a different throw.

For the procedural side, read valid rolls from the main craps guide.

Random does not mean the totals must alternate neatly

A common player test for randomness is: “We already had three sevens, so another seven would be too strange.” That intuition is wrong.

If the dice are fair and throws are independent, the next-roll probability of seven remains 6/36, or 1/6, regardless of whether the previous five rolls were sevens or whether seven has not appeared for twenty rolls.

Random sequences often contain patterns that humans interpret as meaningful:

• three or four identical totals close together;
• a point repeated several times;
• a long hand with no seven-out;
• several short hands in succession;
• alternating high and low totals;
• a cluster of hardways;
• a shooter who appears “hot” for a period.

None of those observations, by itself, demonstrates a physical bias. Probability describes long-run frequencies, not a schedule that every short sample must obey.

Michael Shackleford’s dice and craps probability material at Wizard of Odds provides an accessible independent reference for two-dice probabilities.

A long roll is compatible with random dice

Players often become suspicious when a shooter keeps the dice for a long time. But the game would be very strange if random dice never produced unusually long hands.

Once a point is established, the shooter continues until making the point or rolling seven. Some points are more difficult to make before seven than others, and a hand can contain many neutral rolls that do neither. Long hands are uncommon, not impossible.

The reverse is also true. A table can produce repeated short hands without anything being wrong. Randomness does not promise entertainment balance. It permits uncomfortable clusters.

The variance simulator is useful for seeing how much short-run results can move even when the underlying probabilities stay fixed.

Dice control claims need evidence beyond a good session

Some players claim that a repeatable grip and controlled throw can influence the dice enough to reduce sevens or favor certain outcomes. The difficult question is not whether a person can throw dice with a particular motion. The question is whether that technique produces a measurable, repeatable probability shift under casino conditions large enough to overcome the house edge.

A few successful sessions do not establish that. Neither does a video of a shooter landing the dice in a small area. To demonstrate a real betting advantage, the result would need to survive large samples, controlled testing, ordinary table conditions, and statistical comparison with the expected distribution.

That is why ChipsAndTruths separates technique claims from demonstrated edge. Read Dice Control Myth for the evidence problem rather than assuming that a controlled-looking motion creates controlled outcomes.

If a casino believes a shooter is not making a legitimate throw, management can enforce the house procedure. A player does not have a right to bypass game-protection rules merely because the technique is described as “skill.”

Fair dice can still produce a house edge

Randomness and fairness of the randomizer do not make every wager fair to the player. The casino advantage comes from the payout rules.

For example, the Pass Line is built around genuine random dice probabilities, but the payment structure gives the house a small edge. Proposition wagers can carry much larger edges even though they use the same dice.

This distinction matters:

Random outcome source ≠ zero house edge

A fair six-sided die can be used in a wager with a fair payout, an unfavorable payout, or even a player-favorable promotion. The random device determines the event probability; the paytable determines how that probability is priced.

Use the craps house edge page and craps odds calculator to compare the bets rather than judging a wager by how exciting its outcome looks.

Physical randomness and electronic craps are different implementations

Traditional craps uses physical dice thrown by a player. Other formats can use different outcome sources.

Bubble craps may use physical dice agitated inside a machine. RNG craps uses software-generated random outcomes. Each format has a different game-protection and testing model, but the displayed bet probabilities should still follow the approved rules for that implementation.

Read Bubble Craps and RNG Craps for those distinctions.

A player should not assume that an electronic presentation is secretly using the same physical process as a live table. Nor should “computer generated” automatically mean non-random. The relevant questions are what outcome mechanism the approved game uses and whether it is operating under the applicable technical controls.

What casino crews watch during live dice play

From the operating side, dice integrity is a continuous procedure rather than a one-time inspection. Dealers and supervisors watch for matters such as:

• both dice remaining in view;
• the shooter using the issued dice;
• abnormal handling or repeated setting behavior that violates house rules;
• dice leaving the table;
• damage, marks, or contamination;
• unusual landing methods or attempts to slide rather than throw;
• interference from chips, hands, or objects;
• correct calls when a die is cocked or a roll is invalid.

Surveillance adds another layer when a disputed or suspicious event needs review. The point is not that every unusual throw is a cheating attempt. Consistent procedures prevent the crew from having to invent a response after something unusual happens.

The easiest randomness test is the wrong one

Players sometimes keep a short tally and compare it with the perfect 36-roll distribution. If there are not exactly six sevens, five sixes, five eights, and so on, they conclude the dice are “off.” That is not how probability works.

The 36 combinations describe the probability structure of each independent roll, not a requirement that every block of 36 rolls reproduce the theoretical table exactly. A sample of 36 throws can deviate substantially. Even hundreds of throws can show noticeable imbalances.

A real bias investigation would need an appropriate sample, a clear hypothesis, correct data collection, and statistical testing. It would also need to rule out recording errors and selective observation. Players are very good at remembering the improbable sequence that hurt them and forgetting the ordinary rolls around it.

What a player should take from the math

If the game is being run normally, treat each new throw as a fresh random event governed by the two-dice distribution. Do not assume a number is “due” because it has been absent, and do not assume a shooter has acquired a permanent advantage because the last hand was long.

The useful questions are simpler:

1. What is the probability of the bet?
2. What does it pay?
3. What is the resulting house edge?
4. How much total action will the session create?

The dice provide the randomness. The paytable prices it. Understanding both is more useful than trying to read intention into a streak of sevens, hard eights, or repeated points.