The Don’t Pass bet reverses the main Pass Line objective, but not perfectly. On the come-out roll, 2 or 3 wins, 7 or 11 loses, and 12 usually pushes under a Bar 12 rule. If 4, 5, 6, 8, 9, or 10 becomes the point, Don’t Pass wins when 7 appears before that point repeats.
The push is the casino’s mathematical adjustment. Without it, the wrong-side bet would be too favorable to the player.
The complete contract
| Stage | Roll | Don’t Pass result |
|---|---|---|
| Come-out | 2 or 3 | Wins even money |
| Come-out | 7 or 11 | Loses |
| Come-out | 12 under Bar 12 | Pushes; stake remains yours |
| Come-out | 4, 5, 6, 8, 9, 10 | Number becomes the point |
| Point phase | 7 before point | Wins even money |
| Point phase | Point repeats before 7 | Loses |
| Point phase | Any other total | No decision |
Some layouts use Bar 2 instead: 12 wins and 2 pushes. Because 2 and 12 each have one dice combination, the standard house edge is unchanged. Read the printed layout rather than assuming every table bars 12.
The Pass Line bet has the opposite objective after a point, but its come-out treatment is not a perfect mirror. That one-combination push is why Don’t Pass is only slightly cheaper, not dramatically better.
Exact probability and house edge
Under Bar 12, the complete-cycle probabilities are:
- win: $949/1980\approx47.93%$;
- lose: $976/1980\approx49.29%$;
- push: $55/1980=1/36\approx2.78%$.
Expected value on a one-unit wager is:
$$ EV=\frac{949}{1980}(+1)+\frac{976}{1980}(-1)+\frac{55}{1980}(0) $$
$$ EV=-\frac{27}{1980}=-\frac{3}{220} $$
Therefore:
$$ \text{House edge}=\frac{3}{220}=1.3636% $$
A $10 Don’t Pass wager has expected loss:
$$ $10\times1.3636%\approx$0.14 $$
The actual resolved result is normally a $10 win or $10 loss; 12 returns the unchanged stake.
Building the win probability from the point branches
The wager wins immediately on 2 or 3 with three of 36 combinations. If a point is established, 7 has six combinations and the point has three, four, or five:
| Point | Chance 7 arrives first | Player position after point |
|---|---|---|
| 4 or 10 | 6/(6+3) = 2/3 | Strongest |
| 5 or 9 | 6/(6+4) = 3/5 | Favorable |
| 6 or 8 | 6/(6+5) = 6/11 | Slightly favorable |
The full win probability is:
$$ P(\text{win})=\frac{3}{36} +2\left(\frac{3}{36}\times\frac{6}{9}\right) +2\left(\frac{4}{36}\times\frac{6}{10}\right) +2\left(\frac{5}{36}\times\frac{6}{11}\right) $$
This simplifies to $949/1980$.
The dice-combinations guide shows why seven becomes the favorite against every individual point.
What the Bar 12 push actually does
A push is neither a win nor a loss. The dealer leaves or returns the flat wager according to procedure, and the next come-out roll begins a new attempt for that same money if the player keeps it in action.
Because 12 appears in one of 36 combinations, the probability that a come-out attempt produces a decision or point rather than another push is $35/36$. The expected number of come-out attempts needed to move past the Bar 12 push is:
$$ E(A)=\frac{1}{35/36}=\frac{36}{35}\approx1.0286 $$
Most bets move forward immediately. Repeated 12s merely delay settlement.
The push is not a player bonus. It removes one natural winning combination from the wrong side. If 12 paid and every other standard rule stayed unchanged, Don’t Pass would have a player advantage. Bar 12 restores the casino edge without changing the later point race.
The point phase is favorable, but the whole bet is not
Once a point exists, the flat Don’t Pass wager has positive conditional expectation for the player. If the point is 4, an even-money $10 bet has:
$$ EV=\frac{6}{9}(+$10)+\frac{3}{9}(-$10)=+$3.33 $$
That does not mean Don’t Pass is a player-advantage bet from the start. The unfavorable come-out outcomes—especially 7 and 11—have already been survived to reach this branch. House edge must be measured across the complete original contract, not from a favorable state selected after the fact.
This distinction matters when a player considers removing the wager after a point is set. Taking down a Don’t Pass bet at that moment gives up a mathematically favorable position.
Laying odds behind Don’t Pass
After the point is established, players can usually make a separate lay-odds wager that 7 will appear first. Because 7 is more likely, the player risks more than the amount won.
| Point | Lay-odds payoff | Common example |
|---|---|---|
| 4 or 10 | 1 to 2 | Risk $20 to win $10 |
| 5 or 9 | 2 to 3 | Risk $15 to win $10 |
| 6 or 8 | 5 to 6 | Risk $12 to win $10 |
These are true odds. The odds portion carries no house edge before rounding or nonstandard limits.
Suppose a $10 Don’t Pass bet establishes point 4 and the player lays $20 odds:
- if 7 arrives first, the flat bet wins $10 and the odds win $10, for $20 profit;
- if 4 arrives first, the player loses the $10 flat bet and $20 odds, for a $30 loss.
Conditional expected value after point 4 is:
$$ EV=\frac{6}{9}(+$20)+\frac{3}{9}(-$30)=+$3.33 $$
The odds add volatility but no expectation; the positive $3.33 comes from the already-established flat-bet state. Across all original Don’t Pass cycles, expected loss remains about 13.6 cents per $10 flat wager.
The odds-bet guide explains why laying odds requires larger chip amounts and how table multiples affect exact payouts.
One-times lay odds cut the blended percentage in half
A useful way to compare flat Don’t Pass with odds is to define “one-times odds” on the wrong side as an odds amount that wins the same $10 as the flat wager:
- lay $20 behind 4 or 10 to win $10;
- lay $15 behind 5 or 9 to win $10;
- lay $12 behind 6 or 8 to win $10.
The expected odds amount placed per original $10 Don’t Pass cycle is:
$$ 2\left(\frac{3}{36}\times$20\right) +2\left(\frac{4}{36}\times$15\right) +2\left(\frac{5}{36}\times$12\right)=$10 $$
Adding the $10 flat wager gives average total committed action of $20 per cycle. The expected loss still comes only from the flat bet:
$$ $10\times\frac{3}{220}=$0.13636 $$
Measured against average total committed action:
$$ \text{Blended edge}=\frac{$0.13636}{$20}=0.6818% $$
The percentage is lower because fair odds have been added to the denominator. The expected dollar loss has not changed, and the worst individual loss is larger. A player who lays $20 behind point 4 can lose $30 on one resolution instead of $10.
This is why an odds multiple should be selected from bankroll tolerance, not merely because a table permits it.
Removal and reduction are unusual advantages of the wrong side
Unlike Pass Line after a point, Don’t Pass may be removable or reducible under many rule sets because the player is surrendering a favorable position rather than escaping a disadvantage. Increasing or replacing the wager after a reduction is commonly prohibited.
The published New Jersey wager-removal rule, for example, permits Don’t Pass and Don’t Come wagers to be removed or reduced but not replaced or increased after the reduction. Other jurisdictions and house procedures may differ.
A player should still ask before touching anything. The dealer controls chip movement on contract areas, and odds may follow separate rules from the flat wager.
Mathematically, removing Don’t Pass after a point is normally poor value. Operationally, a player may still reduce exposure for bankroll or stop-loss reasons. Risk control and expected value are related but not identical decisions.
A worked table sequence
A player places $15 Don’t Pass.
- Come-out roll is 12: Bar 12 produces a push. The $15 remains in place.
- Next come-out roll is 5: point 5 is established.
- The player lays $30 odds to win $20.
- Shooter rolls 8: no decision.
- Shooter rolls 7: the flat wager wins $15 and the odds win $20.
Total profit is $35, and the $45 combined stake is returned.
Now change only the last roll to 5. The flat $15 and $30 odds both lose, for a $45 loss. The larger amount lost on a point result is the price of receiving true odds on the more likely seven.
Don’t Pass versus Don’t Come
The Don’t Come bet uses the same wrong-side logic after the shooter already has a point. Its first roll is personal to that wager, just as Come mirrors Pass Line.
The practical difference is timing:
- Don’t Pass is placed before the table come-out roll;
- Don’t Come is placed while a table point is active;
- both can travel to a number and then win on 7 before that number;
- a new Don’t Come wager can win on 2 or 3 while existing traveled wrong-side bets lose or win according to their own states.
Do not combine their chip positions mentally. Each contract and odds stack must be tracked separately.
The biggest practical errors
Several mistakes are more expensive than choosing Pass instead of Don’t Pass:
- reading Bar 12 as a win because 12 is craps;
- placing lay odds in an amount that creates awkward payout rounding;
- removing the flat bet after a point because seven “has not shown lately”;
- increasing exposure after a loss to recover on the next seven;
- confusing a working Don’t Come position with the table’s Don’t Pass line;
- celebrating a seven-out without checking that all wrong-side odds were actually booked.
The mathematical position can be favorable after the point while the bankroll decision is still too large. A correct bet at an unaffordable stake remains an unsafe session choice.
Etiquette without superstition
Betting Don’t Pass is permitted action, not an attack on the shooter. The casino accepts both sides because both are priced into the game.
The friction comes from behavior, not the wager itself. Loudly celebrating a seven-out while most of the table loses can be rude. Quietly placing the bet, confirming payouts, and avoiding comments about the shooter keeps the game professional.
Dealers should settle the wager exactly like any other approved bet. They are not supposed to discourage correct action because the table mood favors Pass Line players.
What the small edge difference means
Don’t Pass at 1.3636% is cheaper than Pass Line at 1.4141%. The difference is:
$$ 1.4141%-1.3636%=0.0505\text{ percentage points} $$
Across $1,000 of flat action, the expected-loss difference is only about:
$$ $1{,}000\times0.000505\approx$0.51 $$
That small pricing benefit does not protect a session from variance. Wager size, number of completed bets, odds exposure, and bankroll limits will dominate the short-run experience.
Use the expected-loss calculator to compare total action, and the bankroll-risk guide before treating a low percentage as permission to increase stakes.
Don’t Pass is a mathematically sound low-edge option when its social position does not bother the player. The correct reason to choose it is the contract and price—not a belief that a shooter is cold, that seven is due, or that betting against the crowd predicts the dice.