Recovering PMF and PDF from the CDF
What you'll learn: How to reconstruct the PMF or PDF of a random variable when you only have its cumulative distribution function.
The Core Idea
The CDF contains all the probabilistic information about a random variable, so we can always work backward to recover the original PMF or PDF. The method depends on whether the random variable is discrete or continuous.
From CDF to PMF (Discrete Case)
For a discrete random variable, the CDF has a staircase shape with jumps at each value in the range. The PMF is hiding in those jumps:
The jump size equals the probability at that point.
Mathematically: P(X = x) = F(x) - F(x⁻)
where F(x⁻) means the limit of the CDF just before reaching x.
Example: If the CDF jumps from 0.3 to 0.7 at x = 2, then P(X = 2) = 0.7 - 0.3 = 0.4.
Think of it like water filling a glass in steps—each jump tells you how much water was added at that level.
From CDF to PDF (Continuous Case)
For a continuous random variable, the CDF is smooth (no jumps). The PDF is the derivative of the CDF:
f(x) = dF(x)/dx
wherever the derivative exists.
The intuition: the PDF measures how quickly probability is "accumulating" at each point. A steep slope in the CDF means high density; a flat region means low density.
Example: If F(x) = x² for 0 ≤ x ≤ 1, then f(x) = 2x.
Key Takeaway: To recover the PMF from a discrete CDF, find the jump sizes at each point. To recover the PDF from a continuous CDF, take the derivative. The CDF is a complete summary—you can always go back.