Standard Deviation of a Random Variable

Two bets can have the same expected value and still feel nothing alike. Imagine one bet that always pays exactly $10, and another that pays $0 half the time and $20 half the time. Both average $10. But the second one swings. Expected value tells you the center of a random variable's long-run behavior; it says nothing about how wild the ride is to get there. Standard deviation fills that gap by measuring typical distance from the mean.

The mechanism mirrors the standard deviation you already know for data sets, but every deviation gets weighted by its probability instead of counted equally. First find the mean, mu, by multiplying each value by its probability and adding. Then for each possible value, compute (x - mu)², multiply by that value's probability, and add everything up — this gives the variance. The standard deviation is the square root of that variance. Squaring the deviations before averaging keeps positive and negative distances from canceling out, and the probabilities make sure outcomes that almost never happen barely count, while common outcomes dominate the spread.

This sits right next to mean of a random variable, and the two always travel together: you cannot find standard deviation without finding the mean first. Looking ahead, this same idea of weighted spread reappears when random variables get combined — adding or subtracting independent random variables adds their variances, never their standard deviations directly. It also resurfaces later as the engine behind sampling distributions, where the standard deviation of a sample statistic (the standard error) tells you how much that statistic bounces around from sample to sample.