Expected hold is the casino win projected before the actual result is known. It combines an expected volume of wagering with the mathematical cost of the game. The phrase can refer to a dollar amount or a percentage, but it is incomplete unless the reporting base is stated.
That last point is the one that prevents bad analysis. Slot coin-in, table-game handle, and table-game drop are not interchangeable denominators.
Start with the base, not the percentage
Three calculations are commonly placed under the label expected hold:
| Setting | Appropriate base | Core calculation | What the result represents |
|---|---|---|---|
| Slot or electronic game | Coin-in | Coin-in × theoretical hold percentage | Expected machine win |
| Table game mathematics | Estimated handle | Handle × blended house edge | Expected table win from wagers |
| Table game operating report | Drop | Expected win ÷ drop | Expected win-to-drop hold percentage |
The first two begin with wagering volume. The third expresses expected win against money or value entering the table’s drop system. It is a useful operating ratio, but it is not the same thing as the house edge.
A game may have a 1% house edge and a 15% table hold in the same month without contradiction. Chips can be wagered repeatedly after entering the table, so handle can be several times larger than drop.
Slot expected hold is the complement of RTP
For a slot paytable with a theoretical return to player of 92.5%, theoretical hold is:
Theoretical hold percentage = 100% − RTP
100% − 92.5% = 7.5%
If projected coin-in is $500,000:
Expected slot win = Coin-in × theoretical hold percentage
$500,000 × 7.5% = $37,500
The machine may win far more or less during a short period. The $37,500 figure is an expectation based on the approved game math and forecast wagering volume. It is not a guaranteed budget receipt.
When a multi-game or multi-denomination machine contains several paytables, the operator needs the mix of wagering by paytable. Applying one hold percentage to all coin-in can misstate expected win if customers concentrate on a different game or denomination than forecast.
Table-game expected win begins with handle
For table games, the mathematically clean starting point is total estimated wagers:
Expected table win = Estimated handle × blended house edge
The blended house edge reflects the mixture of wagers actually made. A roulette table dominated by even-money bets and a table dominated by high-edge special wagers do not have the same weighted expectation, even if their drop is identical.
Suppose a blackjack pit is expected to generate $320,000 in handle. Based on rules, player mix, and observed decisions, management uses a blended edge of 1.2%:
$320,000 × 1.2% = $3,840 expected win
If the same pit is expected to receive $80,000 in drop, the expected win-to-drop hold is:
Expected hold on drop = Expected win ÷ expected drop × 100%
$3,840 ÷ $80,000 × 100% = 4.8% expected hold on drop
The 1.2% and 4.8% figures measure different things. One is the expected share of wagers. The other is the expected share of drop. The relationship depends on how many times the dropped bankroll turns over through repeated bets.
Turnover is the hidden bridge
A useful operating estimate is the turnover multiple:
Turnover multiple = Estimated handle ÷ drop
In the blackjack example:
$320,000 ÷ $80,000 = 4.0 times turnover
Expected hold on drop can then be written as:
Expected hold on drop = Turnover multiple × blended house edge
4.0 × 1.2% = 4.8%
This relationship explains why two properties offering similar rules can report different table hold percentages. Average session length, buy-in behavior, game speed, credit, player mix, chip recycling, and the timing of cash-outs all affect turnover relative to drop.
It also explains why “increase the house edge” is not the only route to more revenue. A lower-edge game that attracts more play, retains customers longer, and produces greater handle can generate more expected win than a high-edge game with little action.
Expected, actual, and realized hold
Assume the blackjack pit above finishes with $9,600 actual win on $80,000 statistical drop.
Actual hold = Actual win ÷ drop × 100%
$9,600 ÷ $80,000 × 100% = 12.0% actual hold
Compared with the $3,840 expectation, the favorable deviation is:
$9,600 − $3,840 = $5,760 above expected win
That does not prove the pit performed better operationally. A few large player losses may have produced the result. Likewise, a negative actual win does not prove the rules or staff failed. Short-term table results can swing sharply around expectation.
The useful comparison is:
| Metric | Meaning |
|---|---|
| Expected hold | What the forecast and game math projected |
| Actual or realized hold | What the records show happened |
| Hold variance | Difference between actual and expected result |
| Operational explanation | Volume, player mix, large play, game changes, errors, promotions, credit, or ordinary variance |
Realized Hold and Actual Win cover the reported outcome. Variance explains why a sound expectation can still miss badly in a short sample.
Where an expected-hold model goes wrong
The formula is simple. The assumptions are not.
Forecast volume is unrealistic
An expected hold model based on last year’s holiday weekend may overstate a normal week. Closed tables, construction, weather, events, competition, and staffing can change available play.
Average bet is not properly weighted
Averages can be distorted by one high-limit player or by taking snapshots at unrepresentative moments. Player-rating accuracy matters when handle is estimated from average bet and decisions.
The wrong edge is assigned
The posted game rules do not reveal the exact player mix. Blackjack decisions, baccarat side bets, roulette wager distribution, and carnival-game options change the blended edge.
Game speed is assumed rather than measured
Decisions per hour vary with staffing, player count, fills, disputes, shuffles, side bets, and manual procedures. A fast theoretical model applied to a slow table inflates expected handle.
Drop is treated as handle
This is the most damaging conceptual error. Drop records value entering the table system. Handle records wagers. One dollar of drop can support several dollars of handle before the player leaves.
Promotions and free play are mixed inconsistently
Match play, non-negotiable chips, rebates, free play, and other offers can affect reported drop, win, theoretical value, and marketing cost differently. The model and actual report must use compatible treatment.
How casinos use the benchmark
Expected hold supports decisions about staffing, table availability, game mix, marketing reinvestment, slot placement, paytable mix, credit exposure, and budget variance. It is also a starting point for investigating results that move far outside established patterns.
Public revenue reports show actual outcomes rather than internal forecasts. The Nevada Gaming Control Board’s Gaming Revenue Information publishes monthly, three-month, and twelve-month data, including reported win percentages. Those reports are useful benchmarks, but a property still needs its own game mix, operating conditions, and historical base levels to build an expected-hold model.
An unusual result should lead to questions, not an automatic accusation. Review may identify:
- a high-value player whose result dominated the period;
- a change in rules, limits, or side-bet mix;
- unusual credit or promotional activity;
- a recording or rating error;
- a procedural exception;
- a large jackpot or bonus event;
- ordinary statistical fluctuation.
Expected hold and player reinvestment
Player offers are usually safer when based on theoretical win rather than one lucky or unlucky result. A simple planning relationship is:
Reinvestment budget = Theoretical win × approved reinvestment rate
If tracked play produces $1,000 in theoretical win and the approved reinvestment rate is 20%, the planning budget is $200. Whether the player actually won $5,000 or lost $5,000 during the visit does not change the underlying value of the same rated action.
The reinvestment rate is a business policy, not a universal casino constant. It must account for offer cost, redemption, taxes, hotel or food margins, player segment, and the property’s commercial strategy.
The clean definition
Expected hold is best understood as a forecasted retention result tied to a named base:
- for slots, expected win from coin-in and theoretical hold;
- for table-game math, expected win from handle and blended house edge;
- for table operating reports, expected win expressed as a percentage of drop.
Without the base, “we expected 8% hold” is ambiguous. With the base, volume assumptions, time period, and game mix stated, expected hold becomes a useful management benchmark rather than a number that appears precise while hiding incompatible calculations.