A slot report can show a machine holding 12% this month even though its designed hold is 7.2%. That does not automatically mean the game was secretly tightened. It may mean the machine experienced ordinary short-term variance, a different play mix, a jackpot timing effect, incomplete data, or a genuine configuration or accounting issue that needs investigation.
From the casino side, hold and RTP are paired views of the same wagering stream:
Actual Hold = Casino Win / Coin-In
Actual RTP = Coin-Out / Coin-In
Actual Hold + Actual RTP = 100%
The words “actual” and “theoretical” must remain visible. A designed percentage and a period result are not interchangeable.
Read the machine report in four passes
Pass 1: establish the wagering base
Coin-in is the total amount wagered, including credits that players win and replay. It is not the same as cash inserted.
A guest may load $100, cycle winnings repeatedly, and create $800 of coin-in before cashing out or losing the balance. Casino win is measured against that wagering volume, not only the original cash.
Pass 2: calculate the accounting result
For a simplified machine-period example:
| Metric | Amount |
|---|---|
| Coin-in | $420,000 |
| Coin-out and credited prizes | $386,400 |
| Actual casino win | $33,600 |
Actual Win = Coin-In - Coin-Out = $420,000 - $386,400 = $33,600Actual Hold = $33,600 / $420,000 = 8.00%
Actual RTP = $386,400 / $420,000 = 92.00%
This is the observed result for that period.
Pass 3: compare with the approved theoretical design
Assume the approved game configuration has a theoretical RTP of 92.8%, equivalent to a theoretical hold of 7.2%.
Theoretical Win = Coin-In × Theoretical Hold = $420,000 × 7.2% = $30,240
Win Variance = Actual Win - Theoretical Win = $33,600 - $30,240 = +$3,360
The machine won $3,360 more than theoretical expectation during the measured period. That is a performance difference, not proof of wrongdoing or a permanent new hold.
The UK Gambling Commission’s monitoring guidance distinguishes theoretical RTP from actual RTP and explains that volatility and volume determine the tolerance around observed results. See live RTP performance monitoring.
Pass 4: ask whether the result is operationally meaningful
A good manager does not stop at the percentage. The next questions are:
- How many games were played?
- Is the variance reasonable for this game’s volatility?
- Was a jackpot or bonus liability recognized in the same period?
- Was the machine available for play?
- Did denomination, bet mix, or occupancy change?
- Did a software, meter, communications, or configuration event occur?
- Is the result consistent across machine, system, and accounting records?
Theoretical hold is not a daily target
A 7.2% theoretical hold does not instruct the machine to keep exactly $7.20 from each $100 or exactly 7.2% every day. It is the long-run expected relationship produced by the approved probability and prize structure.
Actual results can sit above or below theory for substantial periods, especially when:
- the sample is small;
- the game is highly volatile;
- large prizes are rare;
- the machine has a progressive component;
- play is concentrated in unusual bet or denomination patterns.
The correct response to one unusual period is not “change the game.” It is “validate the data, quantify the tolerance, and look for repeated or unexplained deviations.”
Hold alone does not rank the floor
Consider two machines over the same month:
| Metric | Machine A | Machine B |
|---|---|---|
| Coin-in | $900,000 | $280,000 |
| Actual hold | 6.5% | 11.0% |
| Actual win | $58,500 | $30,800 |
| Theoretical hold | 6.8% | 9.5% |
| Uptime | 99.3% | 91.0% |
Machine B has the higher hold percentage. Machine A generates much more casino win and has stronger availability. Declaring B “better” from hold alone would ignore volume, designed math, and lost operating time.
A useful floor review separates at least five dimensions:
- volume — coin-in, games played, active hours;
- yield — actual and theoretical win;
- variance — difference between actual and expected performance;
- availability — uptime and service interruptions;
- contribution — revenue after direct costs and promotional effects.
Add the costs that hold does not show
Gross machine win is not final contribution. A cabinet may carry lease fees, participation payments, wide-area progressive fees, content charges, maintenance costs, and promotion expense.
A simplified contribution view is:
Direct Contribution
= Actual Win
- Participation or Lease Cost
- Progressive/Network Fees
- Direct Promotion Cost
- Attributable Service Cost
Assume the example machine’s $33,600 actual win also carries:
| Direct item | Amount |
|---|---|
| Participation fee | $5,040 |
| Progressive/network fee | $1,200 |
| Direct free-play cost | $2,100 |
| Attributable service cost | $660 |
Direct Contribution
= $33,600 - $5,040 - $1,200 - $2,100 - $660
= $24,600
This does not make hold irrelevant. It places hold inside the economics that management actually controls.
The denominator must remain stable
Casino teams sometimes compare percentages built from different bases:
- slot hold against coin-in;
- property gaming win against cash drop;
- promotion cost against theoretical win;
- cabinet contribution against occupied hours.
A percentage without its denominator can create a false comparison. Reports should name both the numerator and denominator and preserve the same definition across periods.
Configuration and access are control issues
A property should be able to prove which approved game and paytable were active, when changes became effective, who authorized them, and how system records reconcile with the device. Nevada’s slot MICS require documented controls over player-tracking and promotional parameters and independent review of key system changes. See the Nevada Slots MICS.
That control principle applies beyond player points: critical configurations should not depend on undocumented memory or unrestricted access.
Three management mistakes to avoid
Chasing last month’s luck
A temporary high hold does not prove a superior game, and a temporary low hold does not prove a defective one.
Comparing unlike games without volatility context
Two games with the same theoretical RTP can produce very different short-run paths.
Treating player experience and yield as opposing goals
A floor that produces revenue through confusing rules, poor service, or inaccessible information may damage retention and increase complaints. Sustainable performance includes comprehension, uptime, and trust.
The casino-side verdict
RTP describes the designed return. Hold describes the casino share. Actual performance tells what happened in a defined period. Theoretical performance tells what the game was expected to produce over sufficient volume. Operations explains why the difference may exist.
The strongest review does not ask only, “What was the hold?” It asks:
Was the game configured correctly, available to play, measured on a reliable denominator, performing within a defensible tolerance, and contributing after its direct costs?
Continue with Slot Performance Metrics, Slot Meter Readings, Slot Volatility Explained, RTP, and Coin-In.