A “cold” craps table is a description of what players have just experienced: short hands, few made points, repeated seven-outs, and shrinking chip stacks. It is not a recognized dice condition and does not change the probability of the next valid throw.
Leaving a table because the atmosphere is frustrating can be sensible. Leaving because the dice have entered a predictive cold state is not supported by the game mathematics.
Why craps creates a stronger temperature story
Craps differs from games where every player receives a separate hand. One roll can settle many wagers around the table. When a 7 appears with a point established, Pass Line, Come, Place, and several other right-side bets can lose together. The collective reaction makes the event feel larger and more informative than it is.
A shooter also continues across several rolls. Players organize the sequence into a “hand,” which gives the beginning and ending a narrative shape. A quick point-seven sequence feels like a failed performance; a long roll feels like momentum.
Those social and structural features explain why the label is persuasive. They do not alter the dice.
A 7 is common before it is meaningful
With two fair six-sided dice, there are 36 equally likely ordered combinations. Six total 7:
P(7 on a roll) = 6/36 = 1/6 ≈ 16.67%
The probability of at least one 7 in ten independent valid rolls is:
1 − (5/6)^10 ≈ 83.85%
That does not mean every ten-roll sequence contains a seven, or that every 7 is a seven-out. On a come-out roll, 7 wins the Pass Line. With a point on, 7 ends the contract against Pass Line and Come bets.
The table state changes what the roll settles. It does not change the six combinations that total 7.
Short hands are built into point mathematics
Once a point is established, only the point number and 7 decide the Pass Line contract. The chance that 7 arrives before the point depends on the number of combinations for each:
| Point | Ways to roll point | Ways to roll 7 | Probability 7 arrives first |
|---|---|---|---|
| 4 or 10 | 3 | 6 | 6/(6+3) = 66.67% |
| 5 or 9 | 4 | 6 | 6/(6+4) = 60% |
| 6 or 8 | 5 | 6 | 6/(6+5) = 54.55% |
Even the easiest points, 6 and 8, lose to 7 more often than they repeat. A cluster of short hands is therefore not evidence of damaged dice, an unlucky shooter, or a table that has become hostile. It is one ordinary expression of the contract probabilities.
The point mix also changes how a session feels. A table that establishes several 4s and 10s can produce quick losses more often than a sequence dominated by 6s and 8s, even though neither mix predicts the next throw. Players often remember the seven-outs and forget the come-out winners, pushes, resolved proposition bets, and hands where no large personal wager was exposed.
For the matching true-odds payouts, see Craps True Odds vs Casino Payouts.
What would count as evidence of a real problem
A cold-looking result sequence is not enough, but casinos should investigate observable control issues such as:
- chipped, marked, or visibly damaged dice;
- dice that leave the table and are returned without inspection;
- repeated invalid throws or failure to hit the required back wall;
- foreign objects or interference on the layout;
- inconsistent calls, incorrect wager settlement, or unresolved disputes;
- a statistically unusual pattern supported by a large, properly collected sample and physical inspection.
The first five items concern procedure and game protection, not whether players have recently won. The final item requires far more evidence than a few short shooters. A fair process is tested through dice controls, roll validity, surveillance, records, and long-run analysis—not table mood.
Cold does not make Don’t Pass predictive
Players sometimes switch to Don’t Pass after several seven-outs because they believe the table has revealed a condition. The Don’t Pass bet can have a low house edge under standard rules, but its value comes from those rules—not from recent results.
Switching sides changes which outcomes you want. It does not make the previous sequence informative. If the next roll is a fair come-out roll, the same combinations remain available whether the last five shooters lost quickly or made several points.
The same applies to hopping 7, Any Seven, or other high-edge bets. A cold-table story can push players toward expensive wagers just as easily as it can push them to the Don’t side.
The belief can increase cost without changing odds
The financial effect is usually behavioral:
Expected loss = total action × house edge
Suppose frustration raises a player’s average total action from $25 to $60 per roll for the next 30 rolls. The added action is:
30 × ($60 − $25) = $1,050
If the added wager mix averages a 5% house edge:
$1,050 × 0.05 = $52.50 additional expected loss
The table did not become mathematically colder. The player’s reaction created more exposure at a higher average price.
What the official rules actually recognize
The Massachusetts Gaming Commission’s official craps rules define come-out rolls, point numbers, valid throws, Pass and Don’t contracts, Come and Don’t Come, and other wagers. They do not define hot or cold table states that modify future probabilities.
Operationally, a casino can investigate dice damage, invalid throws, interference, or procedural irregularities. Those are evidence-based game-protection questions. “The last shooters were short” is not equivalent evidence.
Field research on the gambler’s fallacy and hot-hand beliefs in casino play shows how players use recent sequences to infer reversal or continuation. The two beliefs contradict each other, yet both can feel convincing after the fact.
A cold table can still be a valid reason to leave
The prediction is weak, but the personal decision can be sound. Leave or reduce action when:
- the noise and crowd are causing impulsive bets;
- losses have reached the pre-set limit;
- frustration is changing the wager mix;
- the session has exceeded the planned time;
- you are betting against your preferred side only to react to the table;
- you no longer understand which wagers are working;
- the social pressure is making it hard to stop.
Those are present facts about the player and session. They do not require a claim about what the dice will do next.
Separate three questions
When a table feels cold, ask:
- What happened? Several shooters had short hands or sevened out.
- What does the math say next? The next valid roll uses the same dice combinations, subject to the current table state.
- What should I do? Follow the money, time, and wager limits chosen before the sequence began.
Mixing the questions turns a descriptive label into a betting signal.
How to test the myth without risking money
A variance simulator can generate many craps sequences. Record in advance what counts as “cold”: for example, three consecutive shooters making no point, or a specified number of seven-outs within a set number of hands. Then measure how often the condition appears and what happens afterward.
Do not redefine cold after seeing the output. Do not keep only the memorable sequences. Compare the next-roll and next-hand results with the ordinary baseline.
The exercise usually shows two things at once: cold-looking stretches occur naturally, and their appearance does not create a reliable forecast.
The table has a state, not a temperature
Craps does have meaningful states: come-out, point-on, working, off, established Come points, and unresolved bets. Those states determine settlement. “Hot” and “cold” are stories placed over them.
For the underlying spread of results, continue to Craps Variance and Why Seven Matters. Compare the opposite narrative in Hot Shooter Myth and the social belief around Don’t Bet Bad Luck. Use the craps odds calculator for wager math rather than table temperature.
A cold label can describe the last few hands. It cannot make the next fair roll owe the table a change.