Variance is the statistical measure of how far a game's real results typically scatter around its average outcome. It is not RTP, and it is not the same as house edge, those tell you what a game pays back on average. Variance tells you how wild the ride is on the way to that average: whether results stay close to it almost every session, or swing far above and below it before finally settling down over the long run.
What does variance measure?
Variance measures the spread of results around the average, not the average itself. Picture two friends who both average $20 in tips a shift. One earns somewhere between $15 and $25 almost every night. The other sometimes walks out with $0 and sometimes with $80, and it still averages out to $20 over a month. Same average, completely different night-to-night experience. That difference in spread is variance. In a casino game, the "average" is the RTP, and variance is the number that describes how far any single session is likely to land from it.
The more useful number in practice is standard deviation, which is just the square root of variance. Variance itself is measured in squared units (squared dollars, if you are working in dollars), which is not something anyone can picture. Standard deviation converts that back into ordinary units, like dollars or multiples of your bet, so "a standard deviation of 6x your bet" means something you can compare directly across games.
Why can two games with the same RTP feel completely different?
Because RTP only tells you the long-run average, not how the game gets there, and two paytables can reach the same average through very different routes. Here is a hypothetical pair of paytables that make the point concretely (these are simplified examples built to illustrate the math, not any real released game).
Game A, low variance. On every $1 spin, you lose your bet 54% of the time and win $2 back 46% of the time. Multiply it out: 0.46 times $2 equals $0.92, so this game returns 92% of every dollar wagered, a 92% RTP. Because wins land on nearly half of all spins, results stay close to that 92% average almost every session. Its standard deviation per spin works out to about 1.0x your bet.
Game B, high variance. On every $1 spin, you lose your bet 97.7% of the time, but the 2.3% of spins that hit pay out $40. Multiply it out: 0.023 times $40 also equals $0.92, the exact same 92% RTP as Game A. But because the win is so much rarer and so much bigger, results swing far more before settling near that average. Its standard deviation per spin comes out to about 6.0x your bet, roughly six times wider than Game A's.
Same 92% RTP for both. One game feels steady, the other feels like a lottery ticket. That gap is variance, and no RTP number on its own will ever tell you it exists. This is also why you cannot judge a game's variance from a couple of sessions: with Game B, going 200 spins without hitting the big win is completely normal, not a sign anything is broken.
How is variance calculated?
In principle, variance is the average squared distance of every possible outcome from the expected value, no heavy notation needed to follow the logic. Walk through the steps in plain terms:
- List every possible outcome of a bet and how often each one happens (its probability).
- For each outcome, work out how far it sits from the average result (the expected value, which is what RTP is built from).
- Square that distance. Squaring does two things: it makes every distance positive, so a miss above average and a miss below average both count, and it punishes big misses far more than small ones.
- Multiply each squared distance by its probability, then add all of those up. That sum is the variance.
- Take the square root of that number to get the standard deviation, back in ordinary units.
Wizard of Odds, a widely cited casino-math reference site, states the same idea as a formula: variance equals the expected value of the squared result, minus the square of the expected result. That is exactly steps one through four above, just written in symbols instead of sentences. You do not need the formula to understand what it is doing: it is measuring how far results typically stray from the average, and punishing the big strays more than the small ones.
How does the plausible range of results change as you play more?
It narrows, and it narrows in a very specific, predictable way: the plausible range around the average shrinks relative to how much you have wagered as spins pile up, even though the variance level itself never changes. That happens because standard deviation across a whole session only grows with the square root of the number of spins, while your expected result grows in direct proportion to spins. Play 100 times as many spins and your expected result multiplies by 100, but the spread around it only multiplies by 10. Relative to the money on the table, the swings shrink. That is the same law of large numbers behind what RTP measures: given enough spins, real results gravitate toward the true average.
Try it below. Pick a variance level and a number of spins, and see how wide the plausible range of outcomes is around the same fixed average.
See the plausible range narrow
Assumes flat $1 spins and a fixed 96% RTP (a 4% house edge), a typical figure for an online slot. Move the variance slider to change how wild each single spin is, and the spins slider to change how long the session runs.
Total wagered
$100
Expected result
-$4
Plausible range (~95%)
-$104 to $96
Range width vs. money wagered
200%
At 100 spins, the plausible range dwarfs the money wagered. A session this short tells you almost nothing about the true average yet.
This is a simplified illustrative model, not a prediction for any real game or session. Planning illustration only, not gambling advice.
Notice what stays fixed and what moves. The variance level never changes when you drag the spins slider, but the range width, shown as a percentage of the money wagered, keeps shrinking. That is the whole idea of the law of large numbers in one number: more play does not lower variance, it just gives the average more room to overwhelm it.
Is variance the same thing as volatility?
They describe the same idea, just from two different rooms. Variance is the term used in statistics, game mathematics, and regulatory testing, the formal, calculated version of the concept. Volatility is the word casinos and slot studios reach for in player-facing marketing and game guides, the everyday label for the exact same spread. Neither term changes what a game pays back over time, that is still RTP's job entirely. If you want the plain-language, player-facing version of this same concept, with the low, medium, and high labels you see on lobby pages, see what volatility means in iGaming.
The academic side of this is not just semantics. A 2008 study published in Cornell Hospitality Quarterly, by researchers Anthony Lucas and A.K. Singh, examined a slot's coefficient of variation, a variance-based measure, standard deviation divided by the average bet, against how long players stayed at the machine. The finding cut against the common assumption that house edge alone drives play time: the study found the coefficient of variation was the stronger predictor. That is a real, peer-reviewed example of variance doing work in gaming research that house edge alone could not do on its own.
What should you take away from this?
Variance is the number that tells you how bumpy a game's ride is on the way to its long-run average, and it is a separate question entirely from whether that average is any good. Two games can post the same RTP and still feel nothing alike, because one pays small and often while the other pays huge and rarely. The math behind it is simple in principle: measure how far every possible outcome sits from the average, square that distance, and average the squares. And the practical upshot is one every player and every operator should hold onto: the more you play, the closer results drift toward the true average, but variance itself never goes anywhere, it is baked into the paytable from the start.
If you are picking games to run on your own platform, variance is a real design lever, not a footnote. Whitelabels.com's casino games library publishes RTP on every title so you can pair it with the variance profile that fits your player base. Pair this with what RTP means for the average side of the equation, and how to manage a casino bankroll for how the two numbers together should shape real betting decisions.
Do you really understand variance?
Answer 5 questions to see how solid your grasp of variance really is. Nothing is sent anywhere, this runs in your browser.
Key takeaways
- Variance is a statistical measure of how far individual results spread out from a game's average, found by averaging the squared distance of every outcome from the expected value.
- Standard deviation, the square root of variance, is easier to read because it comes back in the same units as the payout itself, like dollars or bet multiples.
- Two games can post the exact same RTP and still have very different variance, since RTP measures the average and variance measures the spread around that average.
- A session's standard deviation grows with the square root of the number of spins, while the expected result grows in direct proportion to spins, so the relative spread narrows the longer you play.
- Variance is the term used in formal statistical analysis, academic research, and regulatory testing; volatility is the same idea in casino-marketing language.
- A 2008 study in Cornell Hospitality Quarterly found that a slot's coefficient of variation, a variance-based measure, predicted player time on device better than house edge alone.
