Start with the odds, not the swings
Casino games are games of chance with a built-in house edge. No strategy, system or choice of timing changes the odds in the player’s favor over time, and most players lose money over the long run. Variance does not change that. It only affects how bumpy the road is. Nobody should gamble money they cannot afford to lose, and anyone who is struggling with gambling can contact a local support service.
What variance means
Variance is a measure of how spread out a game’s results are around their average. A game with low variance produces results that cluster close to the average, so a player’s balance moves in small steps. A game with high variance produces results that are widely scattered: many small losses punctuated by occasional large wins, or the reverse.
The average itself is a separate quantity. The expected value of a bet is the probability-weighted average of all its outcomes. In a casino game it is negative for the player, and its size per unit wagered is the house edge. Variance describes how far individual results typically stray from that average. The square root of variance is the standard deviation, which is expressed in the same units as the bet and is easier to picture.
A worked comparison on one wheel
Take a single-zero roulette wheel with 37 pockets. These numbers are for illustration, using a bet of 1 unit.
A straight-up bet on one number. It wins with probability 1/37 and pays 35 units of profit; otherwise it loses 1 unit.
- Expected result per bet: (35 + 36 × (−1)) / 37 = −1/37, about −0.027 units.
- The average of the squared results is (35² + 36 × 1²) / 37 = 1,261 / 37, about 34.08.
- Variance is about 34.08 (the square of the average, roughly 0.0007, barely changes it). The standard deviation is about 5.8 units.
A bet on red. It wins 1 unit with probability 18/37 and loses 1 unit with probability 19/37.
- Expected result per bet: (18 − 19) / 37 = −1/37, again about −0.027 units.
- Every result is exactly 1 unit in size, so variance is just under 1, and the standard deviation is about 1 unit.
Both bets have the same expected loss per unit wagered. Yet the straight-up bet has a standard deviation nearly six times larger. That gap is the whole story of variance: the same average cost, delivered in very different patterns.
How variance scales over many bets
When bets are independent events, the variance of a total adds up. The standard deviation of the total therefore grows with the square root of the number of bets, while the expected loss grows in direct proportion to the number of bets.
Over 100 bets of 1 unit each:
- Expected loss on either bet: 100 / 37, about 2.7 units.
- Standard deviation on red: about 1 × √100 = 10 units.
- Standard deviation on the straight-up number: about 5.8 × 10 = 58 units.
So after 100 red bets a result several units above or below the average is routine, and after 100 straight-up bets the range of ordinary outcomes is far wider. Some players will finish well ahead and many will finish well behind, with the average sitting at a small loss. Being ahead after a short run is luck, not evidence that the odds were favorable, and the pull of the law of large numbers only becomes visible over far more bets than a typical visit includes.
What variance does to the experience
Variance shapes three things a player actually feels.
Frequency of wins. Low-variance games tend to return something often, which makes sessions feel steady. High-variance games can go a long time without a meaningful return, and a long dry spell is normal for them, not a sign that something is wrong. The relationship between how often a game pays and how much it pays is covered by hit frequency, and in slots it is a central design choice.
Size of swings. With high variance, a balance can move by many times the bet size in either direction. A player who sized a bankroll for a calm game may find that a volatile one drains it far faster than the house edge alone would suggest. This is the practical reason risk of ruin is discussed separately from the edge.
Memorable moments. A rare large win is vivid; a thousand small losses are not. High-variance games therefore tend to be remembered for their best moments, which can distort a player’s sense of how they actually did. This is one of several cognitive biases that make results feel different from the arithmetic.
Variance is not a way to win
Because variance creates the chance of a big short-term win, it is easy to mistake it for an opportunity. It is not. Taking a higher-variance bet raises the odds of a large swing in both directions while leaving the expected loss per unit wagered unchanged. A lower-variance bet does not make the game cheaper either; it just spreads the same cost more evenly. Choosing between them is a question of which kind of experience a person prefers, and of whether they have set a gambling budget made only of money they can truly spare.
Some extra practical points follow from this:
- A short session cannot reveal a game’s real odds, because variance dominates the result.
- Results that look like patterns, such as hot or cold streaks, are normal outputs of variance and do not predict what comes next.
- Playing longer does not smooth luck into a profit. It gives the house edge more bets to work on.
Understanding variance will not tilt a game toward the player, but it explains why a few good minutes at a table or machine say very little about how the game is built.