Mathematics
Discrete vs Continuous Distributions
Quick fact
If you roll a fair six-sided die, the probability of getting exactly 3.5 is zero—that's because the distribution is discrete. For a continuous variable like temperature, the probability of any exact value is also zero; probabilities only make sense for intervals.
Why this is interesting
Why are some random outcomes, like the number of cars passing by, always whole numbers, while others, like your exact height, can be any value? This distinction defines the two fundamental types of probability distributions.
Read the full explanation
Understanding Discrete vs Continuous Distributions
Imagine you have a bag of marbles—each marble is a distinct object. Picking one gives a discrete outcome: you can list all possibilities (red, blue, green). Now think of a continuous stream of water—you can measure its volume precisely, but there’s no smallest 'drop' that you can count. A discrete distribution deals with outcomes you can count and list individually, like the number of heads in 10 coin flips. A continuous distribution deals with outcomes that can take any value within a range, like the exact time a runner finishes a race. In discrete distributions, we use a probability mass function (PMF) that assigns a probability to each specific outcome. In continuous distributions, we use a probability density function (PDF); probabilities are found by calculating the area under the PDF over an interval, not by evaluating at a single point.
A deeper explanation
The underlying principle is the nature of the sample space: discrete distributions have a countably infinite or finite set of possible values, while continuous distributions have an uncountably infinite set. This difference drives the mathematics: for discrete, probabilities sum to 1 (sum of PMF values); for continuous, the integral of the PDF over the entire range equals 1. More importantly, for continuous variables, the probability of any exact value is zero because there are infinitely many points—only intervals carry non-zero probability. Understanding this distinction matters because it determines how we model real-world phenomena. For example, counts like 'number of emails received per day' are discrete, often modeled with a Poisson distribution. Measurements like 'time until a radioactive atom decays' are continuous, modeled with an exponential distribution. Mistaking a continuous situation for discrete (or vice versa) leads to incorrect calculations and misinterpretations of data.