What Standard Deviation Is
Standard deviation measures the spread of outcomes around the expected value. In roulette, the expected value of betting red on a European wheel is negative (you lose 2.7 percent of your wager on average). But the actual outcome on any single spin is not your expected value. It's red or black.
Think of it this way: if you play roulette for one hour, you'll probably lose 2.7 percent of your starting bankroll. But you might lose 30 percent in swings while waiting for variance to settle. Or you might win 15 percent while the house edge slowly grinds you down over the session.
Standard deviation is the size of those swings.
The Math
On a European roulette wheel, betting on even money bets (red or black, odd or even, high or low) has a standard deviation of approximately 1 unit per bet. If you bet $100 per spin, the standard deviation is roughly $100 per spin.
What does that mean? In a 100-spin session, you'd expect your results to deviate from your expected loss by 10 units of standard deviation. That means your actual result could be anywhere from negative $1,000 to positive $700, roughly, within one standard deviation.
Two standard deviations pushes the range wider. Three standard deviations wider still. The probability of being outside three standard deviations approaches zero, but approaches is the operative word.
Where This Hurts
People see variance and think it gives them a chance. It does, mathematically. You can play perfectly correct roulette and leave a winner 48 percent of the time (slightly less than 50 percent because of the house edge). But the variance is small because roulette is simple. Each spin is independent. There's no skill correction.
Behaviorally, standard deviation is where loss aversion becomes dangerous. You're down $500 after 50 spins. Your expected loss was $67, but variance pushed you down faster. You feel like you're playing badly. You feel like you need to recoup. You raise your bet size.
Raising your bet size doesn't change your expected value. It just increases your standard deviation. You're buying more variance in the hope that variance goes in your direction. This is called chasing, and it's the mechanism that turns a bad game into financial ruin.
The Recovery Trap
Here's the behavioral economics part: your brain models the situation incorrectly. You think you're below expected value and need to catch up. In reality, you're within normal variance and need to wait longer or accept the loss.
The prospect theory work by Tversky and Kahneman showed that people overweight losses relative to gains. A $500 loss feels much worse than a $500 gain feels good. So when variance goes against you, you feel it acutely. When you're up $500, the feeling is pleasant but not proportionally pleasant.
This asymmetry is what creates the recovery trap. You're down, you feel terrible, you increase your bet to try to get back to even quickly, and the increased bet size compounds the loss when variance continues to be unkind.
The Right Response
If you understand standard deviation, you understand that a $500 loss after 50 roulette spins is not a signal that you should play differently. It's a signal that you're playing something with a house edge and variance is doing what variance does.
The correct response is either to increase your sample size (play longer to let variance work toward expected value) or to accept the loss and stop playing. Not to increase your bet size hoping to recoup faster.
Standard deviation is why bankroll management matters. If your bankroll is too small relative to your bet size, standard deviation can wipe you out before expected value even gets a chance to grind you down. The swings are large enough to hit your total loss limit.
