The worst craps bets are not simply the ones that lose most often. A rare bet can be fairly priced, and a frequent winner can still be expensive if the payout is too short. The useful comparison is house edge under the actual paytable, speed of resolution, and how much total action the wager encourages.
Put the price beside the bet name
Under common U.S. payouts, several familiar craps wagers are dramatically more expensive than Pass, Don’t Pass, or Place 6/8.
| Bet or bet family | Common rule/payout | Approx. house edge |
|---|---|---|
| Any Seven | 4:1 | 16.67% |
| 2 or 12 straight up | 30:1 | 13.89% |
| 3 or 11 straight up | 15:1 | 11.11% |
| Any Craps | 7:1 | 11.11% |
| Hard 4 / Hard 10 | 7:1 | 11.11% |
| Hard 6 / Hard 8 | 9:1 | 9.09% |
| Big 6 / Big 8 | even money | 9.09% |
| Field, both 2 and 12 pay 2:1 | standard otherwise | 5.56% |
The Wizard of Odds craps house-edge table shows how sharply the result changes when a paytable changes. A Field bet, for example, can improve from 5.56% to 2.78% if one extreme pays 3:1 instead of 2:1.
That is why a serious “worst bets” list must name the rule, not just the label.
Any Seven is the classic payout trap
Seven is the most common two-dice total: six combinations out of 36. That frequency makes an Any Seven proposition feel intuitive.
At 4:1, a $5 bet wins $20 when 7 appears and loses $5 on the other 30 combinations.
EV = (6/36 × $20) - (30/36 × $5)
= $3.333... - $4.166...
= -$0.833...On a $5 stake, losing about $0.833 in theoretical value is a 16.67% house edge.
The problem is not that 7 is rare. The problem is that a fair net payout for a one-in-six event would be 5:1, not 4:1. The missing payout unit is the casino edge.
Straight-up horn numbers sell a huge payoff at a steep discount
Totals 2 and 12 each have one combination out of 36. A fair one-roll net payout would be 35:1. A common 30:1 return creates a 13.89% house edge.
Totals 3 and 11 each have two combinations. A fair net payout would be 17:1. Paying 15:1 produces an 11.11% edge.
These bets are easy to market because a $1 chip can return a visually satisfying stack. But the correct comparison is not “I risked only one dollar and won thirty.” It is “How much value was removed from the fair payout each time I bought that event?”
The same logic applies to a Horn Bet, where the stake is split across 2, 3, 11, and 12. The blended price depends on the exact pay rules and how the losing units are handled in the quoted return.
Hardways charge for a story the dice do not care about
A hard 8 wins only when 8 appears as 4-4 before either an easy 8 or 7 ends the wager. The “hard way” condition feels special because the dice must match.
With common 9:1 pricing, Hard 6 and Hard 8 carry about a 9.09% house edge. Hard 4 and Hard 10 commonly pay 7:1 and carry about an 11.11% edge.
That is much more expensive than simply placing 6 or 8 at 7:6, where the common house edge is about 1.52%.
A hardway can survive many neutral rolls, which creates suspense and makes the player feel that the bet is lasting. Duration does not make the price better. It only spreads the resolution across more rolls.
Big 6 and Big 8 are bad because a better version sits on the same table
Big 6 and Big 8 are especially useful teaching examples. They win if 6 or 8 appears before 7 and commonly pay even money, producing about a 9.09% edge.
Place 6 and Place 8 use essentially the same target-versus-7 race but pay 7:6, lowering the common house edge to about 1.52%.
So the comparison is unusually clean:
Big 6: same basic 6-before-7 race, usually 1:1
Place 6: same basic 6-before-7 race, usually 7:6Choosing Big 6 is not paying for a different probability. It is accepting a worse price for convenience.
Side bets can be worse than the old proposition box
Modern craps layouts may add Repeater Bets, Fire Bet, All Tall Small, bonus craps, hops, or other specialty propositions. Some carry house edges above the traditional center action.
Repeater versions, for example, can sit around the high teens or above 20% depending on the selected total and paytable. A different branded product can use a different rule, so the number must be calculated rather than borrowed from another layout.
The lesson is broader than any one side bet: large posted returns and long suspense are not evidence of good value.
If a layout adds a wager you do not recognize, ask for the rules or open the electronic help screen before placing it.
Fast resolution makes an expensive edge more damaging per hour
House edge is a percentage of action. A one-roll proposition can be placed and resolved repeatedly, which means the bankroll can cycle through the edge very quickly.
Suppose a player makes a $5 Any Seven on 80 rolls:
Total action = $5 × 80 = $400
Expected loss ≈ $400 × 16.67% ≈ $66.68Now compare $400 of total action on Place 6 at roughly 1.52%:
Expected loss ≈ $400 × 1.52% ≈ $6.08The actual short-run outcome can be anything from a profit to a much larger loss. The expected-loss gap is still enormous.
This is why “I only make a $5 prop bet” is incomplete. The missing variable is how many times the $5 is put at risk.
The Field belongs in the middle, not automatically at the bottom
Players often hear “the Field is a bad bet” as though every layout were identical. The paytable matters.
Common examples:
- if both 2 and 12 pay 2:1, house edge is about 5.56%;
- if one pays 2:1 and the other 3:1, house edge is about 2.78%;
- unusually generous rules can improve it further.
A 2.78% Field is still more expensive than Pass/Don’t Pass or Place 6/8, but it is nowhere near Any Seven at 16.67%. Ranking it without reading the printed 2 and 12 payouts is lazy analysis.
The same discipline applies to Buy/Lay commission timing and promotional paytables.
A practical “do not buy blindly” ranking
Instead of memorizing a permanent list, use three bands.
Red flag — usually very expensive:
- Any Seven at 4:1;
- many straight-up one-roll propositions;
- many Repeater or jackpot-style side bets;
- hardways when compared with low-edge alternatives.
Orange flag — materially expensive but rule-sensitive:
- Big 6 / Big 8;
- weak Field schedules;
- some Buy/Lay bets when commission treatment is unfavorable;
- specialty side bets whose paytable is not posted clearly.
Lower-price core bets:
- Don’t Pass around 1.36%;
- Pass around 1.41%;
- Place 6 / 8 around 1.52%;
- true odds at 0% house edge on the odds portion.
Lower edge does not make any of these safe or profitable. It simply means less mathematical cost per dollar of action.
“Fun money” is still action
A player can knowingly choose an expensive bet for entertainment. The mistake is describing the wager as harmless because the unit is small.
If you reserve $20 for center action and are comfortable losing it, that is a bankroll choice. If the same $20 is replenished five times because each proposition “was only a dollar,” the real side-bet budget was $100.
For anyone tracking a session, separate:
Core-bet action
+ Odds action
+ Side-bet / proposition action
= Total actionThen estimate cost by the edge of each category instead of applying one blended percentage to everything.
The best defense is a comparison bet, not a slogan
When a wager looks appealing, ask what nearby bet buys similar action at a better price.
- Want 6 or 8? Compare Big 6/8 with Place 6/8.
- Want to back 7? Compare the urge for Any Seven with the fact that Don’t Pass uses 7 favorably after a point at a much lower overall edge.
- Want a long sweat? Compare a high-edge bonus bet with a low-edge line bet plus odds.
- Want one-roll excitement? At least read the exact payout before assuming the posted return is fair.
Use Best Craps Bets as the opposite comparison, craps odds for combination counts, and the house edge calculator for paytable testing. The worst craps bet is often the one a player repeats without ever asking what the payout removed from fair value.