A craps rating is the casino’s attempt to estimate the value of a player’s action over time. It is not a perfect accounting of every chip and it is not a record of whether the player happened to win or lose today. The practical model combines rated average wager, time, pace, and an assumed house advantage, then uses that theoretical loss as the basis for offers and comps.
What the pit is trying to estimate
The casino does not need to know what will happen in one session. It wants an estimate of what a pattern of play is worth over many similar sessions.
A simplified model is:
Theoretical Loss ≈ Rated Action × Decisions × House Advantageor, when time is used explicitly:
Theoretical Loss ≈ Average Rated Bet × Decisions per Hour × Hours × House AdvantageReal player-tracking systems can be more sophisticated, and properties do not all use the same assumptions. The model is useful because it explains why four details matter so much: what was rated, for how long, at what pace, and at what theoretical price.
The craps expected-loss page explains the same logic from the player’s side.
Craps is unusually difficult to rate cleanly
Blackjack often gives the floor one prominent main wager to observe. Craps spreads action across the layout.
A single player may have:
- a $15 Pass Line bet;
- $30 or $75 in odds;
- $18 each on the 6 and 8;
- a Come bet moving to a number;
- occasional Field action;
- a hardway that stays up;
- presses and regressions after wins.
The “average bet” can therefore change from roll to roll.
That is why craps ratings are estimates. A supervisor has to watch the game, protect the table, manage fills and credits, resolve disputes, handle player requests, and maintain ratings at the same time. At a busy table, recording every $6 press as though it were a stock-market trade is not realistic.
Odds create a rating paradox
The true-odds portion behind Pass/Come or Don’t bets has 0% house edge. The Wizard of Odds craps guide separates that fair odds portion from the house-edge flat wager.
That creates an important comp paradox.
Player A might have $15 Pass plus $75 odds: $90 physically at risk, but only the $15 flat wager contains the casino advantage.
Player B might have $90 spread across place and proposition bets. The same visible dollar amount can carry much more theoretical value to the casino.
Some properties exclude odds from the rated average, some discount them, and some tracking systems handle them in their own way. A player should therefore not assume that “I had $500 on the table” means the casino rated $500 of comp-generating action.
Two players can lose the same amount and earn different offers
Imagine two players each finish a trip down $300.
Player A played $25 Pass with odds for two hours, used few side bets, and maintained relatively low theoretical action.
Player B played similar time but spread larger house-edge action across place bets and center bets.
Their actual loss is identical: $300. Their theoretical profiles are not.
A casino using theoretical loss can value Player B more highly even if Player A happened to lose faster in this particular session. That is why comps are usually not a rebate on today’s cash loss.
This distinction also explains why arguing “I lost $1,000, where is my free room?” can miss the way the system is designed. Actual loss matters operationally and for host judgment, but the repeatable baseline is theoretical value.
A comp is a partial rebate, not free money
Suppose a simplified rating estimates:
- average rated action: $75;
- 80 relevant decisions per hour;
- 2 hours;
- assumed blended house advantage: 2.0%.
Then:
Theo = $75 × 80 × 2 × 0.02 = $240If the casino returns 20% of theo in comp value:
Comp Value = $240 × 0.20 = $48The $48 meal or room credit is valuable. But it came from a play profile expected to cost about $240 under the assumptions. It did not turn the gambling into a profitable transaction.
This is the central warning behind craps comp value: do not buy a small benefit by deliberately increasing expensive action.
Rating error can come from ordinary table conditions
A craps rating can be off without anyone acting dishonestly.
Common sources of approximation include:
- joining or leaving mid-shooter;
- changing denomination or spread repeatedly;
- large odds that are not rated the same way as flat action;
- pressing and regressing after hits;
- turning place bets off for some rolls;
- moving between spots or tables;
- long pauses caused by disputes or dice changes;
- a supervisor updating the average periodically rather than continuously.
For this reason, two competent supervisors can produce slightly different estimates from the same messy session.
If comps matter to you, the practical response is not to game the supervisor. Present the player card when you buy in, keep your wagering instructions clear, and check your account afterward if something looks obviously wrong.
Chasing a rating can make the underlying deal worse
A common mistake is to add high-edge action because it looks good to the rating system.
Suppose a player adds frequent proposition bets hoping to generate more offers. The casino may indeed see more theoretical value. That does not mean the extra offer exceeds the extra expected loss.
The right comparison is:
Incremental Comp Value vs. Incremental Expected LossIf $100 of additional expected gambling cost produces $20 of additional benefits, the player has not found a bargain. The player has purchased $20 of value for an expected $100 price.
The same logic applies to status chasing. A tier level can be enjoyable or useful, but it should be evaluated as a benefit with a cost, not as proof that more play is economically justified.
Read the rating from the casino’s point of view
The floor is not asking, “How loyal does this player feel?” It is asking questions such as:
- How long has the player been active?
- What action is consistently working?
- How much of the visible stake carries house edge?
- Is the current average representative or just one temporary press?
- How quickly is the game producing decisions?
That perspective explains why a clear, stable pattern is easier to rate than constant layout changes.
It also explains why an electronic product can produce a different level of precision. Some hybrid systems can record wagers directly at the terminal rather than relying entirely on a human estimate. That improves measurement but does not change the underlying comp principle: theoretical value drives the offer.
Actual loss and theoretical loss can diverge sharply in one trip
A player can be rated at $200 of theoretical loss and leave $1,000 ahead, or be rated at the same $200 and leave $1,500 behind. That gap is not evidence that the rating was wrong. Theoretical loss is an average-cost estimate; actual win/loss is the result of one volatile sample. Craps can produce especially wide short-term differences because odds, place bets, and proposition bets create very different variance profiles.
For comp analysis, keep those ledgers separate. Use actual loss to understand what happened to your bankroll. Use theo to understand how the casino is pricing your future value. Mixing the two makes both numbers less useful and encourages players to chase benefits after a bad session simply because the cash loss feels larger than the rating.
Evaluate the offer after you understand the cost
Use Craps Roll Speed and Total Action to understand how pace changes exposure and Craps Expected Loss Per Hour to put the estimate in dollars. Then compare best craps bets with worst craps bets before deciding whether a richer rating is actually desirable.
The expected loss calculator and house edge calculator are better decision tools than the size of the buffet comp. A comp can reduce the effective cost of play; it cannot make an expensive wager cheap by changing its label.