The machine with the highest coin-in on the floor is not necessarily the best-performing machine.
It may sit directly opposite the entrance. It may have been available six hours longer than comparable units. A single high-stakes player may have inflated its result. Or it may only look impressive before jackpot funding, free play, participation fees and technical costs are considered.
Slot machine performance metrics must therefore measure more than total revenue. A reliable evaluation combines wagering volume, mathematical expectation, actual results, capacity, availability, direct cost and net contribution.
The thresholds and figures in this article are illustrative. Meter definitions, theoretical values, jackpot accounting and reporting methods must be validated against the casino management system, manufacturer documentation, licence conditions and the property’s approved accounting policies.
Start with machine-days
Comparing two machines by monthly GGR alone is often misleading because the machines may not have had the same opportunity to earn.
Active machine-days = Σ(daily number of playable machines)
If one machine was available for 27 days during a 30-day month, it produced 27 active machine-days. Showing its performance against a full 30-day capacity without separately reporting the three lost days hides an availability problem.
The basic normalized measure is:
GGR per machine-day = Machine GGR ÷ active machine-days
Example:
- Machine A: 2.7 million GGR over 30 active days = 90,000 per day
- Machine B: 2.4 million GGR over 24 active days = 100,000 per day
Machine A generated more in total. Machine B was more productive on the days it was playable. Management must then investigate why Machine B lost six days of availability.
Management error: Removing out-of-service days from the performance denominator without reporting the lost availability as a separate operational failure.
Coin-in measures activity, not profit
Coin-in, handle or total bet is the sum of all wagers made on a machine. High coin-in can indicate strong player demand, but it does not show the full economic value of that demand.
Coin-in alone does not reveal:
- Free-play cost
- Jackpot contribution
- Licence or participation fees
- Player comps
- Direct technical cost
- The effect of short-term luck
Coin-in is therefore a measure of movement. Net contribution is a measure of value. A broader slot analytics framework should connect the two.
Actual GGR and hold
Slot GGR = Coin-in − coin-out
A property’s system may treat jackpots, free play or accounting adjustments differently. The report dictionary must state exactly what is included in each meter and calculation.
Actual hold % = GGR ÷ coin-in × 100
Example:
- Coin-in: 80 million
- GGR: 5.2 million
- Actual hold: 6.5%
Actual hold is affected by player luck in the short term. It is not, by itself, proof of game quality or a successful configuration. It must be interpreted alongside theoretical hold, sufficient wagering volume, game volatility and an appropriate reporting period.
Theoretical hold and theoretical win
Theoretical hold is the casino’s expected long-run share of wagers under the approved game mathematics.
Theoretical win = Coin-in × volume-weighted theoretical hold
If a machine contains several games, denominations or paytables, a simple average should not be used. Each theoretical hold must be weighted by the volume played through the relevant configuration.
Weighted theoretical hold = Σ(game volume × game theoretical hold) ÷ total volume
Example:
- 50 million in volume at a 7.0% theoretical hold
- 30 million in volume at a 9.0% theoretical hold
- Weighted theoretical hold:
(50 × 7 + 30 × 9) ÷ 80 = 7.75% - Theoretical win:
80 million × 7.75% = 6.2 million
Actual-to-theoretical index
Actual-to-theoretical index = Actual GGR ÷ theoretical win × 100
Using the example above:
5.2 ÷ 6.2 = 83.9%
This does not automatically mean that the machine is underperforming. If the sample is too small, the difference may be ordinary short-term variance.
The index should be reviewed together with:
- Games played
- Total coin-in
- Game volatility
- Large jackpot outcomes
- Length of the reporting period
- Meter and configuration accuracy
Games played and average bet
To understand what caused a change in coin-in, separate the number of games from the average amount wagered:
Average bet = Coin-in ÷ games played
Example:
- Previous period: 400,000 games × 120 average bet = 48 million coin-in
- Current period: 360,000 games × 150 average bet = 54 million coin-in
Coin-in increased by 12.5%, but the number of games fell by 10% while average bet increased by 25%. The growth is concentrated across fewer games and may have come from a narrower player base. Management should investigate whether the result is sustainable and how the player mix changed.
Machine availability
Availability % = Playable time ÷ scheduled open time × 100
A machine that suffers 18 hours of unplanned downtime in a 720-hour month has availability of:
(720 − 18) ÷ 720 = 97.5%
If the floor average is 99.0%, the difference may look small for one unit. On a high-traffic position, however, it can represent a material loss of revenue opportunity.
Estimated revenue lost to downtime
Estimated gross theoretical loss = Downtime hours × comparable hourly coin-in × theoretical hold
- Downtime: 18 hours
- Comparable coin-in for the same days and hours: 95,000 per hour
- Theoretical hold: 8%
- Estimated gross theoretical loss:
18 × 95,000 × 8% = 136,800
This calculation does not account for players who moved to other machines. To estimate the true incremental loss, management must also examine total floor performance and the effect on nearby units.
Net contribution after jackpot and game costs
Net machine contribution = Adjusted GGR − separately reported jackpot funding − actual free-play cost − licence/participation fees − direct technical cost
If jackpot or free-play cost has already been deducted in the GGR calculation, it must not be deducted a second time.
Example:
- Adjusted GGR: 5.2 million
- Separately reported jackpot cost: 430,000
- Actual free-play cost: 190,000
- Participation fee: 520,000
- Direct technical cost: 110,000
- Net contribution: 3.95 million
The machine ranked first by GGR may rank third by net contribution.
Build the right comparison group
Comparing every machine with the overall floor average is not always fair. A useful peer group should contain machines with broadly comparable operating conditions. This principle is central to effective slot operations management.
The comparison group should be as similar as possible in:
- Location and traffic class
- Denomination
- Cabinet type
- Game age
- Volatility and hit frequency
- Jackpot structure
- Player segment
- Scheduled operating hours
Performance index = Net contribution per machine-day ÷ peer-group median net contribution per machine-day × 100
An index of 118 means that the machine produced 18% more net contribution per active day than the median of its comparable group. Management should still confirm that the difference persists across several periods and is supported by a credible sample.
The core slot machine performance metrics
| Dimension | Primary metric | Management question |
|---|---|---|
| Demand | Coin-in per machine-day | Are players choosing the machine? |
| Usage | Games played and average bet | How is the wagering volume being created? |
| Mathematics | Weighted theoretical hold | What should the volume be expected to produce? |
| Result | GGR and actual-to-theoretical index | What happened in the reporting period? |
| Availability | Uptime and downtime | Did the machine have the opportunity to earn? |
| Economics | Net contribution per machine-day | What real value did the machine retain? |
| Risk | Concentration, faults and jackpots | Is the result repeatable and sustainable? |
Use decision thresholds
Do not remove or convert a machine because of one weak month. A disciplined process uses three decision stages:
- Monitor: Data quality or sample size is not yet sufficient.
- Intervene: An availability, placement or configuration problem has been demonstrated.
- Replace: After correcting controllable conditions and allowing enough time, the machine’s incremental net contribution remains below that of a credible alternative.
Every intervention should state the expected economic effect, the responsible person and the date on which the result will be reviewed.
Conclusion: measure the strongest unit, not the largest number
Slot performance is not coin-in alone. It is not GGR alone either.
A proper evaluation combines demand per machine-day, theoretical expectation, actual results, availability and real cost. It then answers one practical question:
Does this machine convert its location and available capacity into more incremental net contribution than the best realistic alternative?
A report that cannot answer this question is not a performance report. It is a meter summary.