In casino language, cycle has three separate meanings that are often confused:
- the average number of independent trials between occurrences of an event;
- the technical period of a pseudo-random number generator before its internal state can repeat;
- a repeated operational routine, such as a count, reporting, or maintenance cycle.
Only the first meaning is normally relevant when a player says a bonus or number is “due,” and even then the word describes an average, not a timetable.
Average cycle: 1 ÷ probability
For an independent event with probability p on each trial, the expected waiting time to the next occurrence is:
Expected cycle length = 1 ÷ p
If a simplified bonus has a probability of 0.01 per spin:
1 ÷ 0.01 = 100 spins
It is reasonable to say that the average cycle is 100 spins. It is not correct to say that the bonus must occur once in every block of 100 spins or that spin 101 becomes more likely after 100 misses.
This waiting-time model is the geometric distribution. Penn State's probability lesson on geometric and negative-binomial distributions treats the geometric variable as the number of trials required to obtain the first success under repeated independent trials with a constant success probability.
The assumptions matter:
- each trial must have the same event probability;
- trials must be independent for the simple model;
- the event must be defined consistently;
- a changed bet, game mode, paytable, or feature state can invalidate the calculation.
Why a 100-spin cycle often contains no hit
The probability of no occurrence in n independent trials is:
P(no event in n trials) = (1 − p)ⁿ
Using p = 0.01:
| Spins | Probability of no bonus | Probability of at least one bonus |
|---|---|---|
| 50 | 0.99⁵⁰ ≈ 60.5% |
39.5% |
| 100 | 0.99¹⁰⁰ ≈ 36.6% |
63.4% |
| 200 | 0.99²⁰⁰ ≈ 13.4% |
86.6% |
| 300 | 0.99³⁰⁰ ≈ 4.9% |
95.1% |
A 100-spin average cycle still leaves a 36.6% chance of no hit in a 100-spin sample. It also allows two hits close together. The average emerges from many waiting times, some short and some long.
The formula for at least one occurrence is:
P(at least one event) = 1 − (1 − p)ⁿ
These formulas measure probability before the sample is observed. They do not turn previous misses into a promise about the next trial.
The next trial does not remember the gap
If outcomes are independent and the event probability remains 1%, then after 300 misses:
P(event on the next spin) = 1%
It is still 1%, not 100%, 50%, or “overdue.” This memoryless feature is the mathematical reason that cycle language can feed the gambler's fallacy.
The same principle applies to independent roulette spins. A fair wheel does not increase a number's next-spin probability because it has been absent. In baccarat, a road pattern does not create a compulsory completion. In craps, a long sequence without a particular total does not alter the next roll's dice combinations.
Cards can be different because dealing without replacement changes composition. The probability of a card can shift as cards leave the shoe. That is a change in conditional probability, not a mystical cycle becoming due. See independent events for the distinction.
Cycle is not hit frequency
The terms are related but not identical.
- Event probability is the chance of a specified event on one trial.
- Average cycle is the reciprocal of that probability when the simple assumptions hold.
- Hit frequency is the share of plays that produce a defined winning return.
- Volatility describes the spread of return sizes, not merely how often any win occurs.
A slot may have a high overall hit frequency because it produces many small returns, while a particular bonus has an average cycle of hundreds of spins. A “win” below the wager can count as a hit in some game statistics even though the player's balance falls.
RNG period is a different technical cycle
A pseudo-random number generator has an internal state. Under a deterministic update rule, the state sequence can eventually repeat. The number of state transitions before repetition is the generator's period.
That technical period is not a player-facing jackpot schedule. Approved gaming systems are designed and tested so the generator, outcome mapping, game rules, and paytable work together according to the applicable standards. A player cannot infer the internal state or period from a visible run of results.
For the technical layer, read random number generator and RNG vs pseudo-random RNG. The word cycle in an RNG discussion should not be mixed with average bonus spacing.
Operational cycles have no probability meaning
Casinos also use cycle for recurring work:
| Operational phrase | Meaning |
|---|---|
| Drop and count cycle | Scheduled removal and counting of gaming funds |
| Reporting cycle | Daily, weekly, monthly, or accounting-period reporting |
| Maintenance cycle | Repeated inspection or service interval |
| Audit cycle | Planned rotation of control testing |
| Shift cycle | Repeating roster or handover pattern |
These are scheduled processes. They can be late, changed, or missed because people and systems control them. A random game outcome is not controlled by an operational calendar.
A useful way to challenge “due” claims
When someone says an event is due because its cycle has passed, ask four questions:
- What exactly is the event?
- What is its verified probability?
- Are trials independent with a constant probability?
- Does the claim describe an average, or does it wrongly predict the next result?
Suppose a player says, “The feature comes every 80 spins, and it has been 120, so it must hit soon.” Even if the verified probability were 1 in 80, the conclusion would be false under an independent model. The next-spin chance would remain 1/80 = 1.25%.
The missing 40 spins do not accumulate into a debt owed by the machine.
The definition to keep
A casino cycle can be an average recurrence, a technical generator period, or a scheduled operating routine. State which meaning is intended. When the subject is random play, never convert an average interval into a prediction that the next hand, spin, or roll is due.
Continue with probability, sample size, long run, and short-term variance to see why small samples rarely follow an average neatly.