Before the numbers
Slots are games of chance with a built-in house edge. No calculation, system or timing changes the odds, and most players lose money over time. Never gamble money you cannot afford to lose. Regulation and legality vary by country, and gambling is age-restricted or illegal in many places. Everything below is a hypothetical illustration of how a game is built, not advice on how to play.
The core idea: a feature is a share of the total
A slot’s average return to player is the long-run proportion of all stakes that it pays back. Designers set that overall figure first, then decide how to divide it between the ordinary spins and the bonus round. The division is the whole job of bonus round math.
The contribution of a bonus round to the total comes from two numbers:
- How often it triggers, on average (for example, once in so many spins).
- How much it pays, on average, each time it triggers, measured in multiples of the stake.
Multiply the average payout by the trigger probability, and you get the feature’s contribution per unit staked. This is an application of expected value: the average result of an uncertain event, weighted by how likely each result is.
A hypothetical example
Suppose a made-up game has these properties:
- The bonus triggers, on average, once every 150 spins. That is a probability of 1/150 per spin.
- When it triggers, it pays an average of 30 times the stake that started the spin.
The contribution per spin is:
30 x (1/150) = 30/150 = 0.20
So on average, the bonus round returns 0.20 of each unit staked, or 20%.
Now suppose the game’s designers want an overall return of 95%. The ordinary spins must supply the rest:
95% - 20% = 75%
The base game, with its line wins and small payouts, would then be tuned to return 75% on average. The complete picture looks like this:
| Component | Contribution |
|---|---|
| Bonus round (30 x 1/150) | 20% |
| Ordinary line wins | 75% |
| Total return | 95% |
| House edge | 5% |
All of these are invented figures chosen for easy arithmetic. Real games differ and publish different amounts of detail, and rules about disclosure vary.
Changing one number moves another
The useful lesson comes from nudging the inputs. The total has to stay at 95% for the game to remain the same product, so changing the feature forces a change somewhere else.
Make the feature rarer, keep the payout. If it triggers once in 200 spins instead of 150, with the same average payout of 30 times the stake:
30 x (1/200) = 0.15, or 15%
The feature now supplies 15%, so the base game must supply 80% to keep the total at 95%.
Make it rarer and bigger. If it triggers once in 200 spins but averages 40 times the stake:
40 x (1/200) = 0.20, or 20%
That gives the same 20% contribution as the first example, even though the experience is different: fewer triggers, but larger ones.
Make it more frequent and smaller. If it triggers once in 100 spins and averages 20 times the stake:
20 x (1/100) = 0.20, or 20%
Again 20%. Three quite different features can produce the same contribution. This is the reason rarer features generally have to pay more, and why a game with a big, infrequent feature feels so different from one with a small, regular feature despite an identical overall return.
What the average hides
An average payout of 30 times the stake does not mean every trigger pays 30. A realistic bonus round has a spread of results: many triggers might pay less than the average, and a few might pay far more. The big ones pull the average up.
That means a typical trigger can pay below the stated average, and some triggers can pay less than the total staked during the spins that led to them. The average is a property of a very large number of rounds. Over a short session, the spread of results, which statisticians describe with standard deviation, can dominate everything. In practice, designers model the full distribution of outcomes, not just the mean, though the mean is the figure that fixes the return.
Why none of this helps a player
It is tempting to read a calculation like this and conclude that knowing the numbers gives an edge. It does not. Each spin is decided by random number generators, and spins are independent events. The probability of a trigger on the next spin is the same regardless of how many spins have passed since the last one. A game that has gone 300 spins without a bonus is no more likely to trigger on spin 301.
The calculation also describes the long run. Over a few hundred spins you might see no triggers, or several, and neither tells you anything about the underlying figures. Short runs are poor evidence about a game.
Using the math for perspective
The best use of this arithmetic is realism. If a feature carries 20% of the return and triggers once in 150 spins on average, then a large part of what the game pays is concentrated in an event that many sessions will not include. Knowing that can help set expectations about what a session is likely to feel like, and it can help you recognize that the excitement of a feature round is not a measure of value.
Treat play as paid entertainment, set a limit before you begin, and stop when you reach it. If you notice that you are spending more than planned or feel unable to stop, a national problem-gambling helpline or local health service can provide confidential help.