Expectation and Variance of the Geometric Distribution
What you'll learn: How to calculate and interpret the average wait time and variability when waiting for the first success in a geometric distribution.
The Core Idea
Remember that the geometric distribution counts how many trials it takes to get your first success (like flipping a coin until you get heads). Two natural questions arise: On average, how long will I wait? and How much does that waiting time vary?
The formulas are surprisingly elegant:
- Expected value (mean): E(X) = 1/p
- Variance: Var(X) = (1-p)/p²
where p is the probability of success on each trial.
Understanding the Expectation
If your success probability is p = 0.5 (like a fair coin), the expected number of flips is 1/0.5 = 2 flips. This makes intuitive sense—if you succeed half the time, you'd expect to wait about 2 tries.
If p = 0.1 (harder task, like making a difficult basketball shot), you'd expect 1/0.1 = 10 attempts. Lower success probability means longer average waits.
Understanding the Variance
The variance tells us how much the waiting time fluctuates. Using p = 0.1: Var(X) = (1-0.1)/0.1² = 0.9/0.01 = 90. The standard deviation is √90 ≈ 9.5 trials, showing substantial variability.
For the easier case (p = 0.5): Var(X) = 0.5/0.25 = 2, with standard deviation ≈ 1.4 trials—much more predictable.
Analogy: Think of waiting for a bus. If buses come frequently (high p), you don't wait long and the wait is consistent. If buses are rare (low p), not only do you wait longer, but the wait time is much less predictable.
Key Takeaway: The geometric distribution's mean (1/p) gives the expected waiting time for first success, while its variance ((1-p)/p²) quantifies how unpredictable that wait is—both increasing as success becomes rarer.