Best Craps Bets — Lowest-Cost Choices by Goal

If “best” means lowest long-run mathematical cost, the strongest common craps choices are straightforward: Don’t Pass/Don’t Come, Pass/Come, true odds behind those bets, and Place 6 or 8. A properly priced buy bet on 4 or 10 can also be competitive when the commission rule is favorable.

None of these bets beats craps. They simply give up less to the house than hardways, horn bets, Any Seven, and many other proposition wagers.

The better way to choose is by goal rather than by one ranking.

Best choices at a glance

A player does not need all of them at once. In fact, one of the easiest ways to turn “good bets” into an expensive session is to stack too many of them together.

For a first live table: Pass Line plus modest odds

For a beginner who wants to follow the shooter and learn the game, the Pass Line is usually the cleanest starting point.

On the come-out roll:

• 7 or 11 wins;
• 2, 3 or 12 loses;
• 4, 5, 6, 8, 9 or 10 becomes the point.

After a point is established, the Pass bet wins if the point repeats before 7 and loses if 7 appears first. The standard house edge is about 1.41%.

If the table allows odds and the player understands the extra exposure, adding odds behind the Pass wager improves the price of the combined action because the odds portion is paid at true odds: commonly 2:1 on 4/10, 3:2 on 5/9 and 6:5 on 6/8.

Example:

$10 Pass Line
Point = 5
$20 odds behind the line

If 5 wins, the flat $10 normally pays $10 profit and the $20 odds pays $30 profit. The $20 odds portion carries no house edge at true odds; the original $10 Pass bet still does.

That distinction matters. Odds do not repair the flat bet. They add more money at a fair price.

For exact mechanics, see Pass Line Bet and Odds Bet Explained.

For the smallest common flat-bet edge: Don’t Pass

The Don’t Pass wager usually has a slightly lower house edge than Pass Line—about 1.36% versus about 1.41% under the familiar rule where 12 pushes on the come-out.

The difference is small. On $10,000 of resolved flat-bet action, a rough expected-loss comparison is:

Pass Line:  $10,000 × 1.414% ≈ $141.40
Don't Pass: $10,000 × 1.364% ≈ $136.40

The expected difference is only about $5 per $10,000 of flat action. That is mathematically real but not large enough to justify playing a wager you find confusing.

Don’t Pass also has a table-culture issue: it wins after a point when 7 appears before the shooter repeats the point. Some players dislike being on the opposite side of the shooter. That is social convention, not mathematics.

The Don’t Pass guide explains its push and removal rules.

Place 6 and 8 are the strongest simple box-number choices

A Place 6 or Place 8 bet normally pays 7:6. Five dice combinations make 6, five make 8, and six make 7. That payout produces a house edge of about 1.52%.

They are attractive to many beginners because the instruction is simple and the player can usually take the wager down or call it off before a roll, subject to house procedure.

Use amounts that fit the payout ratio. At a table accepting those stakes, $12 on 6 pays $14, while awkward amounts can force payout rounding or may not be accepted as requested.

The dedicated Place 6 and 8 guide covers the clean betting units and working status on come-out rolls.

Why not Place 4 and 10?

At the standard 9:5 payout, Place 4 and Place 10 carry about a 6.67% house edge. The underlying dice are not the problem; the payout is.

If the table offers buy bets at true odds with a commission collected only on winning wagers, buying 4 or 10 can be much cheaper than placing them. But commission timing and rounding matter, so this is not a universal “always buy” instruction.

For example, a $20 Buy 4 paid 2:1 with a $1 commission only when it wins has this conditional EV before 4 or 7 resolves the bet:

Probability 4 before 7 = 3 / (3 + 6) = 1/3
Probability 7 before 4 = 6 / 9 = 2/3

Net win when 4 appears = $40 - $1 vig = $39

EV = (1/3 × $39) - (2/3 × $20)
   = $13.00 - $13.33
   ≈ -$0.33

That is about a 1.67% expected loss on the $20 wager. If the commission is charged differently or rounded differently, the result changes. Read Buy Bets Explained before treating a generic percentage as your table’s price.

Come and Don’t Come are good bets, but they multiply action

A Come bet works like a new Pass Line wager made after the shooter’s point is established. Don’t Come follows the corresponding wrong-side structure. Mathematically they belong near Pass and Don’t Pass.

The trap is not their price. It is quantity.

A player can have a Pass Line bet with odds, then add Come bets on several rolls, then add odds to those Come points. Within a short sequence, the table can contain several simultaneous low-edge wagers. When 7 appears, many of them can resolve together.

Low percentage edge does not mean low dollar exposure.

Suppose the player has:

• $10 Pass plus $20 odds;
• $10 Come on 5 plus $20 odds;
• $10 Come on 6 plus $20 odds;
• $10 new Come bet waiting in the Come box.

That is $100 in active or immediately resolving action. A seven can settle several positions at once. The mathematics may be efficient, but the bankroll swing is much larger than a single $10 Pass wager.

“Best” can mean lowest edge, lowest expected dollars, or easiest control

These are not the same objective.

A player with a $100 entertainment budget might prefer one $10 Pass wager at a time over $10 Pass plus $50 odds, even though adding odds lowers the percentage edge of the combined money. The larger odds bet raises the amount at risk when the point-versus-seven decision resolves.

Expected loss is:

Expected Loss = Total Action × Weighted Average House Edge

But bankroll stress also depends on how much action can lose together and how volatile the sequence is.

That leads to three different definitions of “best”:

Best percentage price: use true odds where available alongside a low-edge base bet.

Best simple beginner procedure: make one Pass Line bet, understand the point cycle, and add only modest odds if desired.

Best control of exposure: keep the number of simultaneous bets small, even if each individual bet is mathematically reasonable.

The craps house-edge page gives the full pricing math. This page is about choosing among those prices.

Field is convenient, not a top mathematical choice

The Field is easy to place and resolves on one roll. Its price depends on what the layout pays on 2 and 12. Common versions can have about a 2.78% or 5.56% house edge.

A 2.78% Field is not catastrophic compared with many proposition bets, but it is still almost twice the edge of Place 6/8 and it resolves extremely quickly. Repeating it every roll can create substantial coin-in.

That is why a bet can look acceptable on a percentage chart yet become expensive through speed.

The center of the table is rarely where “best” lives

Hardways, horn bets, hop bets, Any Seven, and other one-roll propositions offer dramatic payouts because the winning events are less frequent. The advertised payoff is not the same as fair odds.

For example, Any Seven wins on six of 36 combinations. Fair profit odds would be 5:1, because five losing outcomes exist for every winning outcome. A common casino payout is 4:1, producing a 16.67% house edge.

That is a much larger pricing gap than the roughly 1.4% on a line bet.

The active Massachusetts Gaming Commission craps rules provide a useful real-world reference for permitted wagers, standard payout floors, true odds, wager timing, and working status. Your casino’s posted rules may differ, so verify the layout before applying any ranking mechanically.

For the high-cost side of the menu, use Worst Craps Bets and One-Roll Bets.

A disciplined $200 example

A player brings $200 and wants a straightforward session without learning the entire layout at once.

One reasonable structure might be:

1. $10 Pass Line.
2. If a point is established, add $10 or $20 odds depending on comfort.
3. Do not add a second wager until the player can explain exactly what the first one needs to win.
4. If more action is genuinely wanted, choose one $12 Place 6 or Place 8 rather than several center propositions.
5. Set a session stop before increasing stakes after losses.

This is not a winning system. It is a way to keep the rules visible and the average pricing low while controlling how many chips can disappear on one seven.

Compare that with a player who starts with $10 Pass, then adds $5 Any Seven, $5 hard 8, $5 horn, $10 Field and several place bets. The second player may have more excitement, but also more money resolving at higher average edges.

Do not optimize the percentage and ignore the dollars

A common advanced-sounding mistake is to obsess over squeezing the house edge from 1.41% to 1.36% while doubling or tripling bet size.

If one plan creates $1,000 of action at 1.41%, its expected loss is roughly $14.10. Another plan creating $3,000 at 1.36% has an expectation around $40.80. The “better percentage” costs more expected dollars because much more money was wagered.

The Expected Loss Calculator is the correct tool when comparing plans with different stakes or numbers of bets.

The shortest useful recommendation

If you want a concise approach rather than a system:

• Learn Pass Line first.
• Add odds only in an amount your bankroll can tolerate.
• Use Don’t Pass instead if you understand it and prefer its slightly lower flat-bet edge.
• For a direct number bet, Place 6 or 8 is normally the cleanest standard choice.
• Consider Buy 4/10 only after checking exactly when and how the commission is charged.
• Avoid turning a low-edge core into a high-action layout full of proposition bets.

There is no “best craps bet” that makes the game positive expectation under ordinary casino rules. The best choices simply pay less for the same entertainment and make the cost easier to understand.