Craps can be cheap or expensive per hour without changing a single posted payout. The missing variable is how much money actually cycles through the layout while you are playing. A $300 buy-in does not describe that exposure. Neither does the largest stack you remember having in front of you. What matters is the sequence of working bets, how often they resolve, and how quickly the table creates the next decision.
One hour at a craps table is not one unit of action
A live craps table does not have a simple equivalent of “100 hands per hour.” Different wagers live on different clocks. A Field bet resolves on the next roll. A hardway can sit through many rolls. A Pass Line bet may win or lose on the come-out, or survive into a point cycle. Odds can add a large amount of stake without adding house edge.
That makes hourly cost a two-part problem:
- What is the mathematical price of each wager?
- How many dollars pass through that price during the session?
The first question is covered by the craps house-edge guide and craps odds. The second is where roll speed and total action become important.
The Wizard of Odds craps rules and probability tables show why the edge differs by wager. Once those edges are known, hourly exposure is an accounting problem rather than a mystery.
Count exposure, not the buy-in
Consider two players who each buy in for $300.
| Player | Typical action | What the buy-in hides |
|---|---|---|
| A | $15 Pass Line, $30 odds | Much of the stake may be zero-edge odds |
| B | $15 Pass, $18 Place 6, $18 Place 8, frequent $10 Field | Several wagers can be working or resolving on the same roll |
| C | Same as B plus $5–$10 center action | Small proposition bets can resolve repeatedly and carry much higher edges |
All three may cash out with similar-looking chip stacks after an hour, yet they can have produced very different total wagering volume.
This is why “I only bought in for $300” is not a useful measure of gambling cost. The same $300 can be recycled many times. A player can wager thousands of dollars in cumulative action without ever having thousands of dollars on the rail at once.
Three clocks are running at the same time
Craps exposure is easier to understand when you separate three clocks.
The roll clock counts physical dice outcomes. A clean crew, few disputes, and decisive players can increase rolls per hour. A crowded table, late bets, payout corrections, or a slow shooter can reduce them.
The resolution clock counts how often a particular wager reaches a win/loss decision. A one-roll proposition has a resolution clock of one roll. A Place 6 may stay working until either 6 or 7 appears. A Pass Line cycle can span several rolls.
The betting clock counts how often the player adds, presses, regresses, replaces, or turns wagers on and off. Two players can see the same dice at the same speed while creating very different action because one keeps adding new bets.
That is why “rolls per hour” is useful but incomplete. The casino edge applies to wagers, not to the sound of the dice hitting the wall.
A worked hour shows how action mix changes the answer
Suppose a player produces the following approximate action in one hour:
| Bet group | Hourly action | Illustrative edge | Expected loss contribution |
|---|---|---|---|
| Line bets | $600 | 1.41% | $8.46 |
| Odds | $900 | 0.00% | $0.00 |
| Place 6/8 | $700 | 1.52% | $10.64 |
| Higher-edge props | $250 | 11.11% | $27.78 |
The player has put $2,450 through the layout, but the expected loss in this simplified example is about $46.88 because the action is not all priced the same.
Two lessons follow. First, odds can make the total dollars at risk look very large while contributing no expected loss of their own. Second, a relatively small amount of expensive proposition action can contribute a disproportionate share of the theoretical cost.
The useful equation is therefore not “buy-in × house edge.” It is:
Total Expected Loss = Σ(Action in Bet Group × House Edge of Bet Group)For a blended shortcut:
Expected Loss per Hour = Total Hourly Action × Blended House EdgeThe shortcut only works if the blended edge reflects the actual mix.
Fast tables can be good operations and expensive sessions
From the casino side, pace is not just “throw the dice faster.” Good pace comes from clean procedure: bets booked before the cutoff, clear calls, correct payouts, few disputes, and a crew that can read the layout without repeated corrections.
That matters because a well-run table can create more legitimate decisions without feeling rushed. A messy table may look energetic while actually producing fewer resolved wagers because the crew is constantly repairing mistakes.
Electronic and hybrid formats change this balance. Automated interfaces can eliminate some chip movement and calculation time, but they can also make it easier for a player to keep many wagers active. The format itself does not make the bets cheaper. If anything, a quicker interface can raise hourly exposure if the player keeps the same bet mix and simply gets more decisions.
For a hybrid example, compare Roll-to-Win Craps with bubble craps. The interface and staffing model change; the expected value still comes from the posted wager rules.
One-roll action changes the texture of the hour
Craps has a special exposure problem because some bets are persistent and others are consumed immediately.
A $5 hardway might remain on the layout for several rolls. A $5 Any Seven or Horn component is decided immediately. If the player replaces that one-roll wager again and again, the same nominal $5 can produce substantial hourly action.
This is one reason players often underestimate proposition cost. They remember the small denomination, not the number of repetitions.
The reverse mistake also occurs with odds. A player sees a large odds stack and assumes it must be expensive because it is large. The odds portion can be volatile—money can disappear quickly on a seven—but true odds do not contain a house advantage. Cost and variance are different measurements. The craps variance guide explains the swing side of that distinction.
Build an action map before judging a session
If you want a realistic estimate, break the session into wager families rather than guessing one average bet.
Track:
- time at the table;
- approximate rolls or decision pace;
- flat line action;
- odds action separately;
- place/buy/lay action;
- Field and other one-roll action;
- center proposition action;
- how often wagers were pressed, replaced, or regressed.
You do not need perfect surveillance-grade records. The point is to stop treating the buy-in as the amount wagered.
A practical session note might say: “90 minutes; $15 Pass; usually $30–$60 odds; 6 and 8 working most point cycles; Field about one roll in four; $5 hardway occasionally.” That description is far more useful than “bought in $500, lost $120.”
What slower play actually changes
If bet sizes and wager mix stay constant, fewer decisions usually reduce expected loss per hour because less action reaches the house edge. That does not mean a slow table changes the edge of any individual wager. It changes the amount of exposure accumulated in a unit of time.
Likewise, choosing low-edge bets is useful but not magical. A 1.5% blended edge on $4,000 of hourly action can have a larger expected dollar cost than a 4% edge on $1,000 of action.
This is the practical meaning behind why low house edge still costs money: price matters, but volume matters too.
Use the right next page for the question you actually have
For wager pricing, continue with craps house edge and craps odds. For a dollar-per-hour model, use Craps Expected Loss Per Hour and the expected loss calculator. If your concern is how widely results can swing around that expectation, use the variance simulator.