The flat Come bet has a house edge of about 1.41%, the same long-run price as the Pass Line. The reason is structural: a Come bet is essentially a new Pass Line cycle that begins after the table’s original come-out roll.
That simple statement hides the part that confuses many players. A 7 can be an immediate winner for a newly placed Come bet and, on the very same roll, wipe out older Come bets that have already traveled to numbers. The bet changes state after its first roll, so the house-edge calculation has to follow the entire cycle rather than looking at only one dice result.
A Come bet has its own first-roll phase even while another point already exists
A Come bet is normally placed after a table point has been established. On the roll immediately after the wager is made:
- 7 or 11 wins the flat Come bet;
- 2, 3, or 12 loses it;
- 4, 5, 6, 8, 9, or 10 establishes a Come point for that individual wager.
If a Come point is established, the bet is moved to that number. From then on, the bet wins if its number repeats before 7 and loses if 7 arrives first.
Massachusetts’ regulated craps rules describe the same two-stage structure: a Come wager can be placed after the come-out roll, wins immediately on 7 or 11, loses immediately on 2, 3, or 12, and otherwise becomes attached to the number rolled. The rule also recognizes supplemental odds behind an established Come point. Exact table limits and working/off procedures can still vary by property.
For the physical table procedure, see Come Bet Explained. This page stays focused on why the price is 1.41% and what happens to that price when odds are added.
The 1.41% edge comes from the complete decision, not the first roll alone
There are 36 equally likely combinations of two fair six-sided dice.
On the Come bet’s first roll:
| Result | Ways | Immediate effect |
|---|---|---|
| 7 | 6 | Win |
| 11 | 2 | Win |
| 2 | 1 | Lose |
| 3 | 2 | Lose |
| 12 | 1 | Lose |
| 4, 5, 6, 8, 9, 10 | 24 | Establish a Come point |
So eight combinations win immediately and four lose immediately. The remaining 24 combinations do not settle the wager; they send it into a point race.
Once a Come point exists, only that point and 7 matter to the eventual decision. The conditional chance of making each point before 7 is:
| Come point | Ways to roll point | Ways to roll 7 | Chance point wins before 7 |
|---|---|---|---|
| 4 or 10 | 3 | 6 | 3/9 = 33.33% |
| 5 or 9 | 4 | 6 | 4/10 = 40.00% |
| 6 or 8 | 5 | 6 | 5/11 ≈ 45.45% |
When the immediate results and all six possible point cycles are combined, the flat Come bet wins 244 out of 495 equivalent resolved cases and loses 251 out of 495.
Player EV
= (244 / 495 × 1 unit) - (251 / 495 × 1 unit)
= -7 / 495
≈ -0.014141
Therefore:
House edge ≈ 1.414%
Theoretical return ≈ 98.586%
That is the same calculation as the Pass Line because the Come bet uses the same win/loss logic, only at a different point in the shooter’s hand.
A first-roll 7 can help one Come bet and hurt several older bets
This is the feature that makes the wager look contradictory when several bets are active.
Assume the table point is 8. You already have:
- a $10 Come bet that traveled to 5;
- another $10 Come bet that traveled to 6;
- a new $10 wager sitting in the Come box.
The shooter rolls 7.
The new Come wager wins because 7 is an immediate winner on its first roll. But the Come 5 and Come 6 lose because, for those older contracts, 7 arrived before their individual points repeated. The original Pass Line also loses because the table point was 8.
Nothing inconsistent happened. Four wagers were in different states even though they were resolved by the same dice roll.
This is why the right mental model is not “Come bets like 7” or “Come bets hate 7.” It is:
A Come bet likes 7 before it has a point and loses to 7 after it has a point.
Odds do not erase the flat bet’s disadvantage
After a Come bet travels to 4, 5, 6, 8, 9, or 10, a casino may allow the player to put an additional odds wager behind it. The odds portion is paid at the true mathematical odds of making that point before 7:
| Come point | Typical true-odds payout on Come odds |
|---|---|
| 4 or 10 | 2 to 1 |
| 5 or 9 | 3 to 2 |
| 6 or 8 | 6 to 5 |
A true-odds wager has 0% house edge by itself. That does not make the original Come bet free. The flat bet still carries its approximately 1.41% expected cost.
Suppose the player makes a $10 Come bet and, after it travels, adds $20 in odds. The expected loss still comes from the $10 flat wager. The additional $20 has zero theoretical house advantage if it is paid at true odds.
A useful way to express the blended percentage is:
Blended edge on total money committed
= Expected loss on flat Come bet
/ Average total flat + odds money committed
The numerator does not become zero. The denominator gets larger because more fair-odds money is attached.
For a simplified illustration, if every $10 flat Come decision carried $20 of odds for the relevant point phase, the expected loss attributable to the flat bet remains about:
$10 × 1.414% ≈ $0.14 per resolved flat Come decision
The total capital at risk can be much larger even though the percentage measured against combined action looks smaller.
More odds means lower blended percentage but larger swings
This is where players often confuse price with risk.
Adding fair odds is mathematically efficient because those extra dollars do not carry a house edge. But more odds also means more dollars can be won or lost when the Come point resolves.
A player with three $10 Come points and substantial odds behind each can have a large amount on the layout. A single 7 can remove all established Come flats and all of their odds at once.
That does not make the bets mathematically worse than before. It makes the session variance and bankroll exposure larger.
The distinction is:
- house edge tells you the long-run expected cost of the priced wager;
- odds multiple changes how much zero-edge money is attached;
- total exposure tells you how much can disappear on one resolving roll;
- variance tells you how unevenly the results can arrive.
A low house-edge strategy can still be too aggressive for a small bankroll if it builds many positions at once.
Come odds can have working/off rules that affect the next come-out roll
When the shooter sevens out, the next roll begins a new table come-out cycle. Existing Come bets do not normally survive a seven-out, but there are situations in craps where a player may have Come-point odds active while a new table come-out is occurring—for example after the shooter makes the Pass point rather than sevening out.
House procedure matters here. Odds behind Come bets are often treated as off on a come-out roll unless the player calls them working, but this is not a universal rule to assume blindly. Electronic craps can also present the choice differently from a staffed table.
The important discipline is to ask before the roll:
- Is the flat Come bet working?
- Are the odds behind it working or off?
- If working, what happens if its number rolls during the table come-out?
- Does the casino require a verbal call or marker to change the default status?
A mathematically correct strategy can still produce a dispute if player and dealer are operating under different assumptions about whether an odds wager was live.
Come, Place, and Pass wagers can occupy the same number but are not interchangeable
A Come 6 and a Place 6 may both appear around the 6 box, yet they have different contracts.
A Come 6 got there by first surviving the Come box and establishing 6 as its point. The flat bet pays even money if 6 repeats before 7, and eligible odds behind it pay true odds.
A Place 6 is directly placed on 6 and is normally paid according to the posted Place-bet payout. It did not pass through a Come phase and can have different rules for removal, working status, and payout.
Likewise, the Come bet is mathematically related to Pass Line but operationally separate from the table’s original Pass Line point. The shooter can be trying to make 8 for the Pass Line while a player simultaneously owns Come points on 5 and 9.
For pricing comparisons, use Craps House Edge rather than assuming that two bets on the same box number have the same edge.
A low edge can still produce high hourly expected loss if action expands
Expected loss depends on both price and volume.
For one $25 flat Come bet:
Expected loss per resolved decision
≈ $25 × 1.414%
≈ $0.35
That number is small because it prices only one flat decision. A real session may contain repeated Come bets, multiple simultaneous points, and odds. The fair odds do not add expected loss, but repeated flat bets do.
If a player resolves 60 separate $25 flat Come bets over a long enough sample, a simple expectation estimate is:
60 × $25 × 1.414% ≈ $21.21 expected loss
Actual session results can be hundreds of dollars above or below that expectation because dice variance is large. Expected loss is not a prediction of tonight’s result; it is the long-run average cost of the priced action.
Dealers track ownership and state, not just chip location
Come betting can create one of the busier layouts in craps. The base dealer may need to manage several players with Come bets on the same box number, each with different flat amounts and odds.
The control questions are concrete:
- Was the wager in the Come box before the dice were out?
- Which player owns the traveled bet?
- Which number is that individual Come point?
- How much flat action and how much odds are attached?
- Are those odds working on the current roll?
- Was a change to odds made before betting closed?
- Is the payout being made as a Come contract rather than as a Place bet?
This is why a fast Come bettor can increase dealer workload even though the underlying math is clean. The game has to preserve the history of each wager as it moves around the layout.
The useful way to read the 1.41% number
The Come bet’s 1.41% house edge applies to the flat Come wager over its complete decision cycle. It does not mean 1.41% is lost on every roll, and it does not mean odds carry the same edge.
The complete picture is:
- the flat Come bet repeats Pass Line mathematics;
- the wager changes from a first-roll contract into a point contract if a box number appears;
- true odds add no house edge of their own;
- more odds reduce the blended percentage measured against total money committed but increase dollar swings;
- multiple Come bets can build substantial exposure even while every individual flat bet remains a relatively low-edge wager.
That is the distinction worth keeping: good pricing does not automatically mean small risk.
Continue with Come Bet Explained for table mechanics, Pass Line House Edge for the matching probability structure, and Odds Bet House Edge for the zero-edge supplemental wager.