Casinos calculate player value by estimating what a player’s tracked action is expected to be worth, then combining that estimate with the property’s wider customer relationship. The core gaming number is usually theoretical loss, or theo. It is built from wager volume and game advantage, not from how painful or lucky one visit felt.
That distinction explains a common casino-floor puzzle: a player can lose heavily and still have modest recorded value, while another player can win and still be treated as valuable. The first may have played briefly or produced little action. The second may have generated many hours of accurately tracked play at a meaningful average bet.
The basic player-value model starts with action, not cash brought to the table
A buy-in tells the casino how much money entered the table at that moment. It does not tell the casino how much was wagered.
A player who buys in for $500 and makes ten $25 blackjack bets has produced $250 of main-bet action. Another player can buy in for the same $500, recycle chips for three hours, and produce thousands of dollars of action. Their buy-ins match; their gaming value does not.
For a simplified table-game rating:
Theoretical loss = average bet × decisions per hour × hours played × house edge
For electronically tracked machine play:
Theoretical loss = coin-in × theoretical hold percentage
Those formulas are models. A casino may use different game-speed assumptions, side-bet treatment, rule sets, player segments, or proprietary factors. The important idea is stable: the value estimate comes from expected economics of the action, not the ending cash result alone.
Read theoretical loss for the core term and average bet for the table-rating input that creates many disputes.
A worked table-game example shows why two similar-looking players can rate differently
Assume two blackjack players both sit down with $1,000.
Player A
- average bet: $50
- rated pace: 60 hands per hour
- time: 2 hours
- assumed edge for the rating model: 1.0%
Estimated action:
$50 × 60 × 2 = $6,000
Estimated theoretical loss:
$6,000 × 0.01 = $60
Player B
- average bet: $150
- rated pace: 60 hands per hour
- time: 4 hours
- same 1.0% assumed edge
Estimated action:
$150 × 60 × 4 = $36,000
Estimated theoretical loss:
$36,000 × 0.01 = $360
Player B’s recorded gaming value is six times Player A’s in this example even if Player A happens to lose $800 and Player B happens to win $2,000. One result is actual variance. The other number is the casino’s estimate of repeatable expected value.
Table-game value is an estimate because the inputs are observed, not perfectly metered
Slots can record coin-in automatically when a loyalty account is active. Traditional table games are less exact. A floor supervisor or electronic rating system may record start time, stop time, average wager, game, buy-in, and sometimes meaningful changes in betting pattern. But the system does not necessarily record every hand and every chip placed.
That creates several sources of measurement error:
- a player starts at $25, later bets $100, and expects the whole session to be rated at $100;
- a rating remains open during a long break;
- a player moves tables and one segment is not attached correctly;
- a crowded game runs slowly while the rating model assumes a standard pace;
- high-edge side bets are treated differently from the main wager;
- the floor uses a conservative average because the bet moved up and down repeatedly.
This is why player rating matters. The casino is not calculating a person’s moral worth. It is estimating the expected gaming economics attached to an imperfect record.
Game choice changes value because not every dollar wagered is priced the same
A dollar of action on one wager can carry a different expected cost from a dollar on another.
Suppose two players each generate $10,000 in total action:
| Player | Illustrative game edge | Estimated theo |
|---|---|---|
| Low-edge main wager | 0.6% | $60 |
| Higher-edge wager | 4.0% | $400 |
The second player produces more theoretical value from the same action because the wager is more expensive mathematically.
This is also why side bets can materially change a rating. A player may think, “I only added $5,” but a $5 side bet made 100 times creates $500 of additional action. If that wager carries a much higher edge than the main game, its expected cost can be significant.
A casino’s exact rating formula is property-specific, so players should not assume a universal blackjack percentage, baccarat percentage, or fixed decisions-per-hour number.
Actual loss can matter without replacing theoretical value
Actual loss answers: what happened?
Theo answers: what was the play expected to produce?
A good player-development or host review may consider both. A very large actual loss can matter for service, discretionary attention, trip review, or relationship management. Long-term actual results can also reveal whether a theoretical model is consistently misaligned with the customer’s real mix of games and wagers.
But using actual loss as the only value measure would make reinvestment unstable. A casino could over-reward one unlucky short session and under-reward a valuable player who happened to win.
That is why how casinos decide comps is a different question from how they calculate player value. Player value is the estimate; the comp decision is what the property chooses to reinvest from that estimate.
Reinvestment is not the same thing as player value
Once the casino estimates gaming value, marketing and player-development teams decide how much to return through offers. That may include rooms, food, free play, event access, transportation, points, or discretionary host benefits.
A simplified model is:
Comp budget = theoretical loss × reinvestment rate
If theo is $400 and a particular campaign uses a 20% reinvestment target, the illustrative budget is $80. That does not mean the player is entitled to exactly $80 of retail-priced benefits. A room, meal, show seat, or free-play offer has its own internal cost, availability, restrictions, and marketing purpose.
The property may also consider visit frequency, recency, market demand, historical response to offers, hotel or dining spend, credit relationship, and whether the customer is likely to return. Those are business layers after the basic gaming-value calculation.
Current public loyalty programs show the underlying inputs without publishing a universal theo formula. Caesars states that table-game Tier Credits depend on the type of table game, average bet, and length of play, while MGM Rewards says table-game earning varies with time played, average bet, and game type. Those public descriptions support the practical floor model: wager size, duration, and game are core rating inputs. See Caesars Rewards earning information and MGM Rewards earning information.
Why a player’s own estimate and the casino’s number can disagree
A player may calculate four hours at $100 and assume the casino did the same. The casino may have recorded 3.4 hours at $75 because of breaks and bet variation. Or the player may remember the largest bets rather than the average.
The clean way to think about a disagreement is to separate the variables:
- Was all play attached to the account?
- Was the time reasonably accurate?
- Was the average bet representative?
- Was the correct game or wager mix recorded?
- Was the player comparing theo with actual loss, which are different numbers?
A host or rewards desk may be able to review missing play, but they may not disclose every proprietary factor used in the valuation model.
Player value should never become a reason to gamble more than planned
The most expensive comp mistake is trying to “improve” a rating by increasing action that would not otherwise have been played.
Suppose an extra hour of play creates $5,000 more action at a 2% edge:
$5,000 × 0.02 = $100 expected loss
Even if that extra play increases the value of an offer by $25, the player has accepted roughly $100 of additional expected gambling cost to pursue $25 of promotional value. Variance can make the immediate result much better or worse, but the economic trade-off is still poor.
Use why casinos track players to understand the data side and how comps are calculated for the casino-management view. If rewards are changing the amount of time or money you planned to gamble, the Responsible Gambling section is the more important next stop.
The practical answer from the casino side
Casinos calculate player value by turning tracked gambling activity into expected economic value. At tables, that normally means an estimated average bet, time, game speed, and game edge. On machines, coin-in and game math can be captured much more directly. Actual results, trip history, non-gaming spend, and relationship factors may influence decisions around that number, but they do not erase the distinction between actual loss and theoretical value.
So when two players receive different treatment, the useful question is not “Who lost more tonight?” It is “What action did the casino record, what was that action expected to be worth, and how much of that value did the property choose to reinvest?”