Casino games are games of chance with a built-in house edge. No strategy, system or timing changes the odds in the player’s favor over time, and most players lose money over time. Nobody should gamble money they cannot afford to lose. Measuring swings explains how uncertain results can be; it does not make them controllable.
The average is only half the story
Expected value gives the average result per bet. Two games can share the same average and feel completely different. One returns small amounts steadily, the other pays rarely and sharply. The measure that captures this difference is the standard deviation: roughly, the typical distance of a result from the average.
A small standard deviation means results cluster tightly around the average. A large one means they are widely scattered.
How it is calculated
For a single bet:
- Find the average result (the expected value).
- For each outcome, subtract the average and square the difference.
- Weight each squared difference by the outcome’s probability and add them. That total is the variance.
- Take the square root. That is the standard deviation.
Squaring makes all deviations positive and emphasizes large ones; the square root puts the answer back into the original units, such as dollars.
Example: $1 on red, single-zero roulette
- Win $1 with probability 18/37, lose $1 with probability 19/37.
- Average = (18 − 19)/37 = −1/37, about −0.027.
- A shortcut: variance = (average of the squared results) − (average)². Each result squared is 1, so the average of squares is 1. Then variance = 1 − (1/37)² = 1 − 0.00073 = 0.99927.
- Standard deviation = √0.99927, about 0.9996, which is practically $1.
So a single even-money bet swings about $1 either side of a tiny negative average. The spread is far larger than the average cost, which is why the outcome of one bet is dominated by luck.
Example: $1 straight-up on one number
- Win $35 with probability 1/37, lose $1 with probability 36/37.
- Average = (35 − 36)/37 = −1/37, about −0.027. The same as before.
- Average of squares = (35² × 1 + 1² × 36)/37 = (1,225 + 36)/37 = 1,261/37, about 34.08.
- Variance = 34.08 − 0.00073, about 34.08.
- Standard deviation = √34.08, about 5.84.
Both bets cost about 2.7 cents per dollar on average. The straight-up bet has a standard deviation nearly six times larger, which means a much bumpier ride. This is the core of what volatility means, and the guide to high volatility covers the experience from the player’s side.
Adding up many bets
For independent bets, variances add. So after n identical bets, the standard deviation of the total is the single-bet standard deviation multiplied by √n, while the expected total is n times the average.
Bets of $1 on red, single-zero:
| Bets | Expected total | Standard deviation |
|---|---|---|
| 100 | about −$2.70 | about $10.0 |
| 10,000 | about −$270 | about $100 |
Bets of $1 straight-up, single-zero:
| Bets | Expected total | Standard deviation |
|---|---|---|
| 100 | about −$2.70 | about $58.4 |
| 10,000 | about −$270 | about $584 |
The expected loss grows in proportion to the number of bets, but the spread grows only with the square root. After 100 bets on red, the spread (about $10) is several times the expected loss (about $2.70), so luck dominates and many such sessions finish ahead. After 10,000 bets, the expected loss of $270 is more than twice the spread of $100, so finishing ahead becomes rare. That is the law of large numbers at work, and it is why the house, which takes huge numbers of bets, sees results close to the theoretical figure. The rule that spread rises with the square root of the number of bets also underlies the effect of session length.
Reading a standard deviation
For sums over many bets, results are often roughly bell-shaped. As a rough guide for such cases:
- About two thirds of outcomes land within one standard deviation of the average.
- About 95% land within two.
For the 100-bet red example, that would suggest most results fall roughly between −$13 and +$7 (the average of −$2.70 plus or minus $10), with about 95% between about −$23 and +$17. These are rough approximations. The bell shape is a poor fit for short sessions of very skewed bets, like single-number bets, where most results are small losses and rare results are large wins. For those games, the standard deviation understates how lopsided the picture is.
What it tells you and what it does not
- It describes a type of game, not a forecast. A large standard deviation means big swings are normal, in both directions.
- It does not remove the edge. A game can be volatile and still lose on average. Volatility changes the odds of leaving a session ahead, not the long-run cost.
- It cannot be used to pick a good time to play. Streaks of luck are not signals, and past swings do not predict future ones.
- Different games have different numbers. Slots with rare large prizes tend to have large standard deviations relative to the bet, which is part of why game variance shapes the way sessions feel.
Game designers sometimes summarize spread with a volatility rating, which is built on the same ideas. When you see two games with the same return but different volatility, the standard deviation is what separates the smooth one from the streaky one. Neither is cheaper to play per dollar wagered, and both are uncertain in any single session, which is a short-term fluctuation you should be prepared to absorb in full.