Standard Deviation in Gambling: Session Swings in Roulette, Blackjack and Poker

Standard deviation measures how far results typically stray from their average. In gambling it explains a puzzle every player meets sooner or later: the house edge on a single-zero roulette wheel is only 2.7%, yet a short session can easily end 20% up or 20% down. Over one bet, the swing dwarfs the edge. Over thousands of bets, the edge grows steadily while the swing grows far more slowly, and that difference decides how sessions turn out in the long run.

Average versus spread

Every bet has two numbers worth knowing:

• The expected result, the average outcome per bet, which is the house edge with a minus sign in front.
• The standard deviation, the typical distance between what actually happens and that average.

Variance is the same idea before a square root is taken. Standard deviation is the more useful of the two because it comes back in the same units as the stake, so it can be read directly in chips.

One spin, two very different bets

On a single-zero wheel, a bet on red wins 18 times out of 37 and pays even money. Its expected result is −1/37 of the stake, about −2.7%, and its standard deviation is almost exactly one stake, since each spin either wins a unit or loses one.

A bet on a single number has the same expected result, because it pays 35 to 1 on a 1 in 37 chance. But it loses most spins and now and then pays 35 units, so its standard deviation is about 5.8 stakes. Same price, nearly six times the turbulence.

How the swing grows over a session

Here is the central rule. Over many independent bets of the same size:

• the expected result grows in direct proportion to the number of bets;
• the standard deviation grows only with the square root of the number of bets.

For one-unit bets on red, that looks like this:

After 100 spins, finishing ahead is entirely ordinary. After 10,000, even a fortunate typical session is still behind. The crossover comes at about 1,369 spins, where the expected loss of 37 units equals one standard deviation. These ranges use the normal approximation, which suits even-money bets well and long-shot bets less well, because long shots pay out in lumpier steps.

Standard deviation in blackjack

Blackjack is mostly an even-money game, but doubles, splits and naturals paid at 3 to 2 add some larger outcomes, so the standard deviation per hand sits a little above one betting unit. Blackjack sessions therefore swing much like even-money roulette bets, while the edge itself depends on the table rules and on how each hand is played. The chances behind many of those hands are set out in our article on blackjack probability.

Variance in poker

In poker the opponents are other players rather than the house, and the operator takes a fee from pots or tournament entries. Results are usually tracked in big blinds per hundred hands, and the spread around any player's average is wide: long stretches above or below it are normal even across many thousands of hands. Short-term results therefore reveal very little about how well anyone is playing, which is worth remembering after a good or bad week. The rules side of one popular variant is covered in our Omaha poker rules guide.

Slots: the widest swings of all

Slot payouts range from nothing to hundreds or thousands of times the stake, so their standard deviation per spin is usually far higher than that of any table bet. The volatility label in a game's help screen is an informal description of this spread. Our look at PAR sheets calculates the figure for a simple invented machine, and the RTP guide explains the average those results swing around.

What standard deviation is good for

• Setting expectations. A session budget should be able to absorb normal swings without touching money needed elsewhere.
• Choosing a pace. Smaller stakes shrink the swing in money terms; our ideas for enjoying slots on a small budget build on that.
• Reading results honestly. A winning month on a negative-edge game is variance, not a method.

What it cannot do is change the expected result. No betting pattern turns a −2.7% bet into a positive one; it only rearranges how the swings arrive.

Gambling is for adults only and works best as a paid pastime with a fixed spend, never as a way to make money. Rules on online play differ widely, so check where you live before you start.