“Place bet house edge” is not one number. Place 6/8, Place 5/9, and Place 4/10 are three distinct prices hidden inside one bet family. Their decisive probabilities differ, their common payouts differ, and the gap between fair odds and casino payout grows as you move from 6/8 toward 4/10.
One layout label hides three very different prices
A Place wager wins if your selected box number appears before seven. Other totals leave the wager unresolved. Because seven always has six combinations, the cost of the bet depends on two things: how many combinations make your number and how much the casino pays when it hits.
The common schedule is:
| Place number | Ways to hit | Ways to seven | Common payout | House edge per resolved decision |
|---|---|---|---|---|
| 6 or 8 | 5 | 6 | 7:6 | about 1.52% |
| 5 or 9 | 4 | 6 | 7:5 | 4.00% |
| 4 or 10 | 3 | 6 | 9:5 | about 6.67% |
That table is the core of the page. Everything else explains why those percentages differ and how to interpret them correctly.
For how to book the wager, working/off status, ownership and presses, use Place Bets Explained. This page is about price.
Place 6 and 8: the casino gives up only a small amount from true odds
Six and eight each have five combinations. Seven has six. Conditional on the bet resolving, the number wins 5 times out of 11 and seven wins 6 times out of 11.
True odds would pay 6:5. The common Place payout is 7:6.
For a correctly sized $6 Place 6:
Win probability on a decision = 5/11
Loss probability on a decision = 6/11
Win = $7
Loss = $6
EV = (5/11 × $7) - (6/11 × $6)
EV = 35/11 - 36/11
EV = -$1/11 ≈ -$0.0909
House edge = $0.0909 / $6 ≈ 1.52%The payout gap is small: fair 6:5 on $6 would be $7.20; the casino pays $7. That is why 6 and 8 are the cheapest standard Place numbers.
This also explains the importance of $6 units. If you choose awkward stakes and the table cannot pay exact fractional amounts, rounding or required unit conventions can change the practical result.
Place 5 and 9: a larger payout shortfall raises the edge to 4%
Five and nine each have four combinations. Against six ways to make seven, the decisive race is 4 versus 6.
True odds are 3:2. The common Place payout is 7:5.
Using a $5 stake:
Win probability on a decision = 4/10
Loss probability on a decision = 6/10
Win = $7
Loss = $5
EV = (4/10 × $7) - (6/10 × $5)
EV = $2.80 - $3.00
EV = -$0.20
House edge = $0.20 / $5 = 4.00%The bet is not worse because five or nine are “cold” numbers. It is worse because the payout is further below the fair 3:2 price.
That distinction matters when comparing Place 5 and 9 with a Buy version. If a casino offers win-only vig on Buy 5/9, the commission-based price may be cheaper than the Place payout.
Place 4 and 10: the payout gap is large enough to invite a Buy comparison
Four and ten each have only three combinations. Seven has six, so seven is twice as likely to appear first once the decision is reduced to those two totals.
True odds are therefore 2:1. A common Place 4/10 pays only 9:5.
For a $5 wager:
Win probability on a decision = 3/9
Loss probability on a decision = 6/9
Win = $9
Loss = $5
EV = (3/9 × $9) - (6/9 × $5)
EV = $3.00 - $3.33...
EV ≈ -$0.333
House edge ≈ $0.333 / $5 = 6.67%This is the most expensive standard Place-number family. It is also why players should compare Buy Bet House Edge rather than automatically placing 4 or 10. A true-odds Buy with favorable commission can be materially cheaper.
Per-decision edge and per-roll cost answer different questions
Craps creates a terminology trap because many bets survive neutral rolls. The standard Place-bet house-edge figures above are measured per resolved decision relative to the amount staked. Wizard of Odds also publishes a per-roll/bet-made basis:
| Place number | Edge per resolved decision | Approx. expected loss per roll as % of stake |
|---|---|---|
| 6 or 8 | 1.52% | 0.46% |
| 5 or 9 | 4.00% | 1.11% |
| 4 or 10 | 6.67% | 1.67% |
Why is the per-roll number smaller? Because many rolls do nothing to a Place wager. If you place 6, a roll of 3, 4, 5, 8, 9, 10, 11, or 12 leaves the bet sitting there. The wager has not been resolved, so spreading its expected loss over every roll reduces the per-roll percentage.
Neither denominator is “fake.” They answer different questions. The decision edge prices one completed wager. The per-roll figure helps when estimating hourly exposure on a bet that can remain working across many rolls.
Correct betting units protect the published payout
Place bets are commonly booked in units that let the dealer pay the stated ratio cleanly:
- 6 and 8: multiples of $6 for a 7:6 payout.
- 5 and 9: multiples of $5 for a 7:5 payout.
- 4 and 10: multiples of $5 for a 9:5 payout.
On a $12 Place 6, a hit pays $14. On a $15 Place 5, a hit pays $21. On a $10 Place 4, a hit pays $18.
If a table allows nonstandard amounts, ask how fractional payouts are handled rather than assuming the theoretical percentage survives unchanged. The clean mathematical edge assumes the advertised ratio is implemented without hidden rounding loss.
This is an operational reason dealers care about units, not just a tradition. Correct units make payouts faster, disputes rarer, and player ownership easier to track.
Place, Buy, and line-with-odds are different economic products
A player comparing only the name of the number can miss the important distinction: how the casino prices the race against seven.
- A Place bet pays a fixed short payout.
- A Buy bet pays true odds and charges commission.
- A Pass/Come bet first has line-bet rules and may then support a separate true-odds wager.
For 6/8, the standard Place price is already efficient enough that buying the number is usually unnecessary. For 4/10, the Place price is expensive enough that a favorable Buy often makes sense. A line bet with odds is a different contract entirely; it cannot be reduced to “betting the 6” because the come-out or Come-roll phase is part of the wager.
The percentage comparison belongs on Craps House Edge. The procedural comparison belongs on the specific bet pages.
Covering more numbers changes total dollars at risk, not the edge on each chip
Suppose you place $12 on 6 and $12 on 8. Each individual wager still has roughly a 1.52% edge per resolved decision. Covering both numbers does not magically average the house edge upward or downward.
What changes is exposure. A seven can remove both wagers at once. More chips are working, more payouts can occur before the seven, and the dollar amount cycling through the game rises.
That is why “low edge” and “low expected loss” are not synonyms. Expected loss is driven by both price and volume:
Expected loss ≈ amount of relevant action × house-edge basisIf you double the amount of action at the same edge, you roughly double expected dollar loss over a comparable sample. If you add expensive Place 4/10 action to cheap Place 6/8 action, you also change the weighted average price of the portfolio.
For pace and dollar consequences, use Expected Loss Per Hour rather than relying on the headline edge alone.
A practical way to rank Place action
If your only goal is to minimize the casino price within the Place-bet family, the order is straightforward:
- Place 6 or 8 — approximately 1.52% per decision.
- Place 5 or 9 — 4.00% per decision.
- Place 4 or 10 — approximately 6.67% per decision.
But that ranking does not tell you how much to bet, whether to work the wager on the come-out, whether to press, or whether another bet family better fits your purpose. Those are separate decisions.
Use Place 6 and 8, Place 5 and 9, and Place 4 and 10 for number-specific play. Use the house edge calculator when you want to compare percentages, and craps variance when the real concern is session swing rather than price.