Free meals, hotel rooms and show tickets can make it look as though a casino is giving something away. It is not. Complimentary rewards, usually called comps, are calculated from a number the casino estimates for every rated player, and that number is not what the player won or lost. It is what the player was expected to lose.
That estimate is called theoretical loss, or theo. Understanding it explains how player ratings work, why a lucky night does not change the rating, and why comps are best seen as a partial refund of a cost that was built into the game from the start.
The theoretical loss formula
For table games, theoretical loss is usually estimated with four inputs multiplied together:
Theoretical loss = average bet × hands per hour × hours played × house edge
The first three inputs describe how much money went into action: the total amount wagered over the session. The house edge converts that total into an expected cost. It is the same idea our article on expected value describes from the player's side, and the same structure our article on hands per hour uses to explain exposure.
Nothing in the formula looks at the outcome. Two players who sit side by side for the same time at the same bet will generate the same theo, even if one walks away up and the other walks away down.
What a casino actually observes
A casino cannot see every decision a player makes, so the rating is built from things a floor supervisor or a tracking system can record or estimate:
- Average bet. Usually an observed estimate of typical wager size across the session, not a record of every hand.
- Time played. When the player sat down and when they left.
- Game speed. Casinos commonly use a figure around seventy hands per hour for blackjack, although real pace varies with the number of players, the dealer and the shuffle method.
- The game and its edge. The house edge the casino assigns to the game for rating purposes.
That last input deserves attention. The edge a casino uses for rating is its own figure, and properties do not generally publish it. It may reflect how a typical player performs rather than perfect basic strategy, since real players make mistakes that raise the effective edge. Under liberal multi-deck rules, correct basic strategy brings the edge to roughly 0.5 percent, while rules such as a 6:5 payout push it considerably higher. What any given casino assumes is not something outsiders can verify.
21 TrainerFree on the App Store. See how rules change the house edge with the built-in calculator, using virtual chips only.
Get the app
A worked example
The numbers below are illustrative, chosen to show how the formula behaves. They are not a quote of any casino's rating method.
Imagine a session with a $25 average bet, 70 hands per hour, and 4 hours of play. That is 280 hands and $7,000 wagered. Now apply two different edges from the commonly published figures:
| Illustrative game | Total wagered | House edge | Theoretical loss |
|---|---|---|---|
| Liberal 3:2 rules, correct basic strategy | $7,000 | About 0.5% | About $35 |
| Same rules but 6:5 payout | $7,000 | About 1.9% | About $133 |
The second row uses the commonly published estimate that a 6:5 payout adds roughly 1.4 percentage points to the edge. Same bet, same time, same pace, and nearly four times the expected cost. Rules move theo just as surely as bet size does, which is one reason our article on table minimums and rules looks at the rule card before the minimum.
Notice also how sensitive the result is to time. Double the session to 8 hours and every theo figure doubles. The same is true of average bet. Each input scales the whole product.
Where comps come from
Once a casino has a theo figure, it returns some portion of it to the player in comps. How large that portion is varies by property, by rewards program and by the kind of comp, and casinos do not publish a standard rate. Any specific percentage quoted as a universal rule should be treated with skepticism.
What is consistent is the structure. A comp is funded out of theoretical loss and is a fraction of it. Comp programs are generally structured so that, on average, rated players are expected to lose more than the comps they receive. The meal or the room is real, but it was priced into the play that earned it.
This is the reason a comp should never be treated as a reason to play. It is not value created by the session. It is a partial rebate on an expected cost that the session already carried.
Why the rating and the real cost can differ
Theoretical loss is an estimate on the casino's side, and a player's true expected cost is a separate number that depends on details the rating may not capture.
The biggest gap is strategy. The expected cost of a session depends on how well the hands are actually played. A player who deviates from basic strategy on stiff hands, soft doubles or pair splits faces a higher effective edge than the rules alone imply, and our article on what common mistakes cost shows how quickly small errors add up. A rating system that assigns one edge to every player cannot see that difference.
Pace is the second gap. A full table plays far fewer hands per hour than a nearly empty one, so a player's real exposure in an hour can sit well above or below whatever pace the rating assumes. Average bet is the third: it is typically an observed estimate, and it can drift from the true figure over a long session.
None of this changes the conclusion. The rating is an approximation of expected loss, and comps are a fraction of the approximation. Whatever the exact numbers, the structure still points the same way.
Why chasing comps runs backward
Because theo scales directly with bet size and time, the most obvious way to earn more comps is to bet bigger or play longer. The arithmetic of that choice is unfavorable by construction.
- Each extra unit of comp costs more than a unit of expected loss. If comps are returned as a fraction of theo, raising theo by some amount raises the comp by a smaller amount. The gap between the two is the casino's margin.
- Longer sessions increase exposure. More hours means more total wagered and a larger range of possible outcomes, not just a larger expected cost. Our article on session length explains what time does and does not change.
- Bigger bets increase swings. Variance scales with bet size, so a session played for a rating can swing far beyond its theo in either direction, as our guide to variance and losing streaks describes.
The honest summary is that comps make a cost feel smaller; they do not make it smaller. If a budget for entertainment exists, the useful question is its expected cost, and comps do not change the answer.
Theo versus actual results
Players are often surprised that a large win does not reduce their rating, or that a painful loss does not raise it. Table ratings are generally built on theoretical loss rather than actual results. Individual properties may apply their own policies, but the underlying principle is that the rating measures play, not luck.
This mirrors the way the game itself works. Over a single session, outcomes are dominated by variance. Over a very long run, results tend toward the expected cost. The casino rates the long-run number because it can predict that one; it cannot predict the short-run one, and neither can the player. Our article on what half a percent really means explains how a small edge becomes a reliable number only across very large numbers of hands.
Seeing that relationship directly is easier in a simulator than at a table. 21 Trainer's Pro bankroll simulator plays sessions with virtual chips and includes an ROI review and strategy-adherence grading, and the house edge calculator shows how each rule shifts the edge. None of it involves real money, and none of it promises a result: the point is to understand exposure, not to chase it.
Understand exposureHouse edge calculator, bankroll simulator with virtual chips, and full rule customization. No account, no ads.
Download
Frequently asked questions
What is theoretical loss in a casino?
Theoretical loss, often shortened to theo, is the amount a player is expected to lose based on how they play rather than what actually happened. For table games it is estimated as average bet multiplied by hands per hour, hours played, and the house edge of the game. It is a long-run average, so a single session's real result can be far above or below it.
How do casinos calculate comps for blackjack?
Casinos generally rate table play on an estimate of theoretical loss, built from your average bet, the time you play, the speed of the game and the house edge they assign to it. Comps are then returned as some fraction of that theoretical figure. The fraction varies by property and program and is not published, so there is no reliable universal percentage.
Do casinos give comps based on wins or losses?
Table ratings are generally built on theoretical loss rather than actual results. A player who has a winning session can still be rated on the expected loss of their play, and a player who loses heavily is rated on the same estimate. Individual properties may apply their own policies, but the underlying idea is that the rating measures play, not luck.
Is it worth playing more to earn casino comps?
Not as a matter of arithmetic. Comps are funded out of theoretical loss and returned as a fraction of it, so every increase in bet size or hours that earns more comps raises expected loss by more than the comp is worth. Comps are best understood as a partial rebate on a cost already priced into the game, not as value created by playing.
How many hands per hour do casinos assume for blackjack?
Casinos commonly use a figure around seventy hands per hour when rating blackjack play, although real pace varies widely with the number of players, the dealer and the shuffle method. Some properties may track or adjust pace differently. Our article on hands per hour covers the commonly published ranges for full and short-handed tables.