The hypergeometric distribution gives the probability of getting exactly k successes when you draw n items from a finite population without replacement. In Intro to Probability, it fits problems where each draw changes the next one.
The hypergeometric distribution is the model you use in Intro to Probability when you sample from a finite population without replacement and want the probability of a certain number of successes. Think of it as the “counting successes” version of a draw where the pool changes every time you pick something.
It is built for situations with three pieces of information: the population size N, the number of successes in that population K, and the sample size n. If you want the probability of exactly k successes in your sample, the formula counts two things at once: the ways to choose k successes from the K available, and the ways to choose the remaining n - k failures from the rest of the population. That gives the familiar expression P(X = k) = [C(K, k) C(N - K, n - k)] / C(N, n).
The big idea is that the probabilities change after each draw. If you draw one success from a deck, there is one fewer success left in the deck, so the next draw is not independent. That is what separates this distribution from the binomial distribution, where the probability of success stays constant from trial to trial.
A quick example helps. Suppose a box has 20 items, 5 are defective, and you inspect 4 items without replacement. The hypergeometric distribution gives the probability that exactly 1 of the 4 is defective. You are not guessing with a fixed p every time, because the chance of drawing a defective item shifts after each pick.
A common mistake is using a binomial model just because the problem asks for a number of successes. If the problem says “without replacement,” or if it gives a finite pile, deck, batch, or lot, pause and check for hypergeometric structure. That one phrase usually tells you the sampling rule.
Hypergeometric distribution shows up whenever a probability problem is really about a finite group and a sample pulled from it. In Intro to Probability, that makes it a direct bridge between counting methods and random variables, because you are turning a combinatorics setup into a probability model.
It also sharpens your intuition about independence. Many early probability problems use coin tosses or other repeated trials where the chance stays the same each time. Hypergeometric problems force you to notice when that assumption breaks, which matters for card draws, quality control checks, selecting students from a class list, or picking parts from a production batch.
This distribution is also a good place to practice reading problem wording carefully. Phrases like “without replacement,” “random sample from a finite population,” or “exactly k defective items” are clues that the binomial model is too simple. Once you spot that, you can set up the correct sample space and avoid overcounting or using the wrong p.
It connects directly to later topics too. Once you can recognize hypergeometric structure, it becomes easier to compare it with binomial and to reason about expected values, conditional probability, and sampling methods. It is one of those concepts that makes the rest of probability feel more precise, because you stop treating every “success count” problem the same way.
Keep studying Intro to Probability Unit 8
Visual cheatsheet
view galleryBinomial Distribution
Binomial distribution is the close comparison point, but it assumes independent trials and a constant probability of success. Hypergeometric distribution is the version you use when the sample comes from a finite population without replacement, so the probability changes after each draw. If a problem says the pool stays the same, think binomial; if the pool shrinks, think hypergeometric.
Finite Population
A finite population is the setting that makes hypergeometric probability possible. You need a countable group with a known number of successes and failures, like a deck of cards or a shipment of parts. The model depends on knowing how many items are in the whole population, not just the success rate.
Sampling Without Replacement
Sampling without replacement is the mechanism behind the distribution. Each item you draw is removed from the pool, so the next draw has different odds. That changing probability is the whole reason the hypergeometric formula is different from binomial formulas.
p (probability of success)
The success probability p is easy to talk about in binomial settings, but hypergeometric problems are usually written with counts instead of a fixed p. You can sometimes compute an initial success rate from K/N, but that rate does not stay constant through the draws. The changing denominator is what makes this a hypergeometric setup.
A problem set or quiz item will usually give you a finite group, a sample size, and a label for what counts as a success, then ask for the probability of exactly, at least, or at most a certain number of successes. Your first move is to check whether the sample is taken without replacement. If it is, you set N, K, n, and k, then plug into the hypergeometric formula or use a calculator command if your class allows it.
You may also be asked to explain why the binomial distribution does not fit. A strong answer points out that each draw changes the remaining pool, so the trials are not independent. When a problem gives a deck, batch, or committee selection, that is usually the cue to identify hypergeometric structure before doing any arithmetic.
These two are confused all the time because both count successes in repeated draws. The difference is the sampling rule: binomial uses independent trials with constant success probability, while hypergeometric uses a finite population without replacement. If the probability changes after each draw, you are not in binomial land.
Hypergeometric distribution gives the probability of exactly k successes in n draws from a finite population without replacement.
Use it when the pool changes after each draw, because the draws are not independent.
The setup always starts with N for the population size, K for the number of successes in the population, and n for the sample size.
Its formula counts favorable samples over all possible samples, which is why combinations appear in both the numerator and denominator.
If a word problem says “without replacement,” that is your strongest clue that hypergeometric, not binomial, is the right model.
It is the probability model for counting how many successes you get when you sample from a finite population without replacement. Each draw changes the composition of the remaining population, so the chance of success is not constant across trials. That makes it different from the binomial distribution.
Look for the phrase “without replacement” or any situation where the sampled items are removed from a finite group. If each trial is independent and the probability stays the same, that points to binomial. If the pool shrinks, hypergeometric is usually the right choice.
The probability of exactly k successes is P(X = k) = [C(K, k) C(N - K, n - k)] / C(N, n). The numerator counts the favorable samples, choosing successes and failures separately, and the denominator counts all possible samples of size n. The combinatorics are doing the counting for you.
If a box has 12 lightbulbs and 3 are defective, and you choose 4 without replacement, hypergeometric distribution can find the probability that exactly 1 chosen bulb is defective. The key detail is that once you remove a bulb, the next draw comes from a smaller pool. That changing pool is what makes the setup hypergeometric.