Key takeaways
- Volatility describes the shape of a payout distribution, not its average. Two machines can be built to return the same fraction of what is staked and still deliver that return in completely different chunks.
- Gambling regulators treat volatility as a calculable quantity rather than a marketing adjective: Pennsylvania's technical standard derives it from aggregate variance and a standard deviation, then multiplies by 1.96 for a 95% confidence interval.
- Because uncertainty falls with the square root of the number of plays, a machine with four times the standard deviation needs sixteen times as many spins before its running average is equally settled.
- Hit frequency and payout size are the two levers behind the feeling. A machine paying twice the stake on two spins in five and one paying twenty times the stake on one spin in twenty-five can share an identical long-run return.
- In a virtual-coin game the only thing volatility costs is coin balance and session length, which is why choosing a machine there is a question about the session you want rather than about money.
Volatility is the shape of a machine's payout distribution. It is not the average, it is not the size of the top prize on its own, and it is not difficulty. It describes how widely a short run of play is likely to scatter around the long-run figure the machine was designed to produce — and it is the single reason two machines can be mathematically equivalent on paper and feel like different games within thirty seconds.
That distinction gets lost because the average has a name people recognise and the spread usually does not. So start with the average, then put it aside.
The average is one number and it hides the interesting one
Pennsylvania's slot machine regulations define theoretical payout percentage as the aggregate awards expected to be paid out over one cycle of the game divided by the total number of combinations in that cycle. It is a property of the paytable, computed over every possible outcome, and it says nothing whatever about the order those outcomes arrive in.
Two machines can be designed around exactly the same theoretical payout percentage and deliver it in completely different chunks. Here is a worked pair, using round numbers so the arithmetic can be checked by hand.
Machine A pays back twice the stake on two spins out of five and nothing on the other three. Its expected return per spin is 0.4 × 2 = 0.8, or 80% of the stake.
Machine B pays back twenty times the stake on one spin in twenty-five and nothing on the other twenty-four. Its expected return per spin is 0.04 × 20 = 0.8. Identical.
Now the spread. For A, the expected value of the squared outcome is 0.4 × 4 = 1.6, so the variance is 1.6 − 0.8² = 0.96 and the standard deviation is about 0.98 stakes. For B, the expected squared outcome is 0.04 × 400 = 16, the variance is 16 − 0.64 = 15.36, and the standard deviation is about 3.92 stakes — exactly four times A's.
The two machines share an average and differ fourfold in spread. Everything a player notices lives in that second number.
What that fourfold difference does to a session
Uncertainty in a running average shrinks with the square root of the number of plays, so the gap widens in an interesting way rather than staying constant.
After one hundred spins, the standard deviation of A's average return per spin is 0.98 ÷ √100 ≈ 0.098, so a 95% interval around the designed 0.8 is roughly ±0.19. For B the same calculation gives 3.92 ÷ 10 ≈ 0.392, and an interval of roughly ±0.77 — wide enough to contain outcomes from "paid almost nothing" to "paid far more than designed". To get B's running average as settled as A's is after a hundred spins, you need sixteen hundred spins, because the required count scales with the square of the ratio of standard deviations.
The dry spells diverge too. A pays on 40% of spins, so ten consecutive blanks happen with probability 0.6¹⁰ ≈ 0.6% — about once in every hundred and sixty-five blocks of ten. B pays on 4% of spins, so ten consecutive blanks happen with probability 0.96¹⁰ ≈ 66%, and fifty consecutive blanks still happen about 13% of the time. Same designed return. Utterly different experience of a five-minute session.
Regulators treat this as a calculable quantity
None of this is folklore. Pennsylvania's slot machine minimum design standard specifies that machine volatility is calculated on a 95% confidence interval, and sets out the steps: aggregate variance is the sum of the probability of every winning combination multiplied by the square of the corresponding payout; a standard deviation is derived from that aggregate variance and the lowest payout percentage; and the volatility index is 1.96 multiplied by that standard deviation.
The standard then does something that makes the practical meaning obvious. It defines an approach percentage — the gap between the machine's lowest payout percentage and the regulatory floor — and specifies the number of plays needed to equal or exceed that floor as the square of the volatility index divided by the approach percentage. That figure may not exceed ten million plays. In other words, the regulator's own test is: how much play is required before this machine's actual return is expected to settle near its designed return? A high-volatility machine answers "a great deal", and there is an enforced limit on how much.
Researchers who obtained manufacturers' design documents through freedom-of-information requests describe the volatility index in the same terms: an indication of how much a game's payback percentage will vary for a given number of games played, with high-index games showing larger variance per session than low-index games. The same work makes a second point worth carrying: the same game can be approved in multiple versions whose main difference is the payback percentage. Two cabinets that look identical need not be identical, and that is a variable separate from volatility.
The two levers a designer actually pulls
Behind every volatility figure sit two adjustable quantities: how often something pays, and how much it pays when it does. Push hit frequency up and payout size down and you get a machine that hums along with small, frequent results. Push the other way and you get long silences punctuated by events large enough to reset a balance.
Reel layout is the mechanism. More reels, more symbol positions and more ways to combine them widen the range of possible outcomes; a paytable weighted towards a small number of high awards concentrates the return. Feature rounds add a third shape, because a machine can hold back a share of its designed return specifically to deliver it inside a bonus, which makes the base game feel drier than the overall figure suggests. If those terms are unfamiliar, the mechanics themselves are worth reading separately in our glossary of reel mechanics, because volatility is the consequence of choices made there.
Probability behaving counter-intuitively over short runs is not unique to slots, incidentally. The same "the average is right and the sample is nowhere near it" problem shows up whenever a simulator resolves outcomes, which is why we wrote about reading probability in sports simulations in almost the same terms.
What volatility means when there is no money involved
Vegas Casino Slots JACKPOT WIN is our own app, built by Reign Creative, and it is a social casino game: reel-based slot machines played with virtual coins and virtual jackpots, with no real-money gambling, no cash payouts or real-world prizes, and coins that have no real-world monetary value and cannot be redeemed for cash.
That is the reason this article exists. In a coin-only game, volatility is stripped of its financial consequence and left with its design consequence. It determines how often the screen does something, how quickly a coin balance moves, and whether a five-minute session is likely to contain any notable event at all. Choosing between machines becomes a question about the session you want — steady activity, or long quiet stretches with the possibility of something large — rather than a question about money. The game's listing does not publish a volatility figure for any machine, so the two levers above — how often something lands and how large it is when it does — are the way to read one; the app page covers what the listing does say.
It is worth being precise about that boundary, because the vocabulary is borrowed wholesale from an industry where the stakes are real. We set out where the legal and platform lines actually fall in our comparison of social casino play and real-money gambling, and the rest of this category sits under anime and creative games.
Good to know
Frequently asked questions
What is the difference between volatility and return to player?
How do regulators actually calculate slot volatility?
Does high volatility mean a machine pays more overall?
Does volatility matter in a game that only uses virtual coins?
Sources
- 58 Pa. Code Chapter 461b, § 461b.1: Slot machine minimum design standards (Technical Standard) (opens in a new tab)Pennsylvania Gaming Control Board — accessed
- PAR Sheets, probabilities, and slot machine play: Implications for problem and non-problem gambling (opens in a new tab)Journal of Gambling Issues — accessed
- 58 Pa. Code § 461a.1: Definitions (opens in a new tab)Pennsylvania Code — accessed
About the author
Reign Creative LLC — The independent app studio that builds the apps covered on this site and publishes every article here. On Google Play it publishes as Reign Collective Apps.
We use AI tools to help research, draft and edit our articles. Each article lists the sources its factual claims rely on, and details about our apps come from their live Google Play listings. No article on this site has been reviewed by an independent expert unless the reviewer is named on it.
Spotted an error? Tell us — how corrections are handled is set out in our editorial policy.