On this concept: Formula

Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 9-12

Binomial Distribution

The binomial distribution gives the probability of exactly k successes in n independent trials, each a success/failure with fixed probability p.

P(X=k)=(nk)pk(1−p)n−k
Venn diagram of Event A and Event BA two-event view of binomial distribution.ABA ∩ B
A: Event A
B: Event B
A two-event view of binomial distribution.

Quick Answer

The binomial distribution gives the probability of exactly k successes in n independent trials, each a success/failure with fixed probability p. Use it when you repeat the same yes/no experiment a set number of times, such as 10 free throws, 20 coin flips or 50 seeds planted, and want the chance of a particular count. The formula (nk)pk(1−p)n−k multiplies the probability of one arrangement of successes by the number of arrangements. The distribution's mean is np.

Why This Matters

Many real questions are counts of yes/no events: how many of 100 patients respond to a drug, how many of 20 parts are defective, how many of 5 shots go in. The binomial formula answers "how likely is exactly this count?" and "is this count surprising?" Using it when draws are made without replacement, or forgetting the (nk) factor, gives probabilities that look precise and are wrong.

Intuitive Explanation

A shooter makes each free throw with probability 0.6 and takes 3 shots. What is the chance she makes exactly 2? One way is make, make, miss: 0.6×0.6×0.4=0.144. But the miss could come first, second or third, giving three orders (MMX, MXM, XMM). Each order has the same probability, 0.144, and there are (32)=3 of them, so P(X=2)=3×0.144=0.432.

That is the whole formula, P(X=k)=(nk)pk(1−p)n−k: the powers give one arrangement's probability, with k successes and n−k failures, and (nk) counts the arrangements. Doing this for every k from 0 to n gives the full distribution. For the 3 shots: P(0)=0.064, P(1)=0.288, P(2)=0.432, P(3)=0.216, which add to 1. The peak sits near the mean np=3×0.6=1.8.

Four conditions make a count binomial: a fixed number of trials, two outcomes per trial, the same p each time, and independent trials. Drawing marbles without replacement breaks the last two, because each red drawn leaves fewer reds behind and changes p. Counting flips until the first head breaks the first, because the number of trials isn't fixed; that count follows a geometric distribution.

Think of it like this. Picture a tray of n coins, each weighted to land heads with probability p. Shake the tray once and count the heads: that count is binomial. Each coin ignores the others (independence) and all share the same weighting (constant p). If the coins could nudge one another, or some were weighted differently, the picture, and the binomial model, would stop fitting.

Core idea. The binomial distribution gives the probability of getting exactly k successes in n independent yes/no trials, each with the same chance p.


When to Use

Use the binomial distribution when there is a fixed number of trials n, each a success or failure, with the same probability p and independent of the others, and you want the chance of a count of successes. If you count trials until the first success, use the geometric distribution. If you draw from a small group without replacement, p changes, so use the hypergeometric distribution. For large n, a normal approximation can replace long sums of binomial terms.

Tip. Ask: Are there a fixed number of independent trials, each success/failure with the same probability, and I want a count of successes?

How to Recognize It

Before writing the formula, check four facts about the trials: how many there are, how many outcomes each has, whether the chance of success stays the same, and whether one trial affects the next.

  1. Is the number of trials fixed before you start, like 10 free throws?

    Yes fits the binomial model. If you keep going until the first success, such as rolling until a 6 appears, the number of trials is what you are counting, and that is the geometric distribution.

  2. Are you drawing from a small group without putting items back?

    Yes means the chance changes after each draw, so the binomial model is off; 5 cards from one deck call for the hypergeometric distribution. With replacement, or when the group is huge compared with the sample, the binomial model works.

  3. Does the question want a probability, not just the number of ways the successes could be arranged?

    Yes means multiply (nk) by pk(1−p)n−k. If it asks only how many orders the successes could come in, (nk) by itself is the answer.

  4. Is n so large that adding up binomial terms would take ages, with np and n(1−p) both at least 10?

    Yes means a normal curve with mean np and standard deviation np(1−p) gives a good approximation. For small n, or p close to 0 or 1, compute the binomial terms exactly.

Binomial Distribution compared with Binomial coefficient, Normal distribution, and Geometric / hypergeometric

All four rows deal with repeated yes-or-no outcomes. Ask what the answer is: a count of arrangements (coefficient), the chance of k successes in n fixed trials (binomial), a wait for the first success or odds that change (geometric or hypergeometric), or a large-n approximation (normal).

Binomial Distribution

Meaning
Use this when you have a fixed number of independent yes/no trials with the same success probability and want the chance of a given count of successes.
Key test
Are there a fixed number of independent trials, each success/failure with the same probability, and I want a count of successes?
Formula
P(X=k)=(nk)pk(1−p)n−k
Example
A free-throw shooter makes each shot with probability p=0.6, independently. What's the chance of making exactly 3 of 5?

Binomial coefficient

Meaning
Just the count (nk), with no probabilities attached.
Key test
Use when you only need the number of ways, not a probability.
Formula
(nk)
Example
Ways to pick 3 of 5

Normal distribution

Meaning
A continuous bell curve; the binomial's approximation when n is large.
Key test
Use for continuous data or as the large-n limit, not for counting discrete successes.
Formula
N(μ,σ)
Example
Heights of adults

Geometric / hypergeometric

Meaning
Trials until first success, or sampling WITHOUT replacement (changing p).
Key test
Use when you wait for the first success, or draws change the probabilities.
Example
Cards drawn without replacement

Formula & Notation

P(X=k)=(nk)pk(1−p)n−k
P(X=k)=(nk)pk(1−p)n−k for k=0,1,…,n; E(X)=np, Var(X)=np(1−p)

How to read it: X∼B(n,p) reads 'X follows a binomial distribution with n trials and success probability p'

Worked Examples

Example 1 — Exactly 3 of 5

Easy

A free-throw shooter makes each shot with probability p=0.6, independently. What's the chance of making exactly 3 of 5?
Show the steps and the answer
  1. Check the conditions: n=5 shots, each a make or a miss, p=0.6 every time, shots independent.

    All four hold, so X, the number of shots made, is binomial with n=5 and p=0.6.

  2. Count the arrangements of 3 makes in 5 shots: (53)=10.

    The 3 makes could be any 3 of the 5 shots, and each arrangement is a separate way to get X=3.

  3. Find one arrangement's probability: 0.63×0.42=0.216×0.16=0.03456.

    Three makes at 0.6 and two misses at 1−0.6=0.4, multiplied because the shots are independent.

  4. Multiply: 10×0.03456=0.3456≈0.35.

    All 10 arrangements are equally likely, so adding them is the same as multiplying by 10.

Answer: P(X=3)≈0.35

Takeaway: Binomial probability = (number of arrangements) × (probability of one arrangement).

Example 2 — Probability changes each draw

Standard

A bag has 3 red and 2 blue marbles. You draw 3 without replacement and want exactly 2 red. Is this binomial?
Show the steps and the answer
  1. Track P(red) draw by draw: 35 on the first draw; if that was red, 24 on the second.

    Without replacement the bag changes, so p is not constant and the draws are not independent.

  2. Count sets instead: 2 of the 3 reds and 1 of the 2 blues gives (32)(21)=3×2=6 sets, out of (53)=10 possible sets of 3.

    Every set of 3 marbles is equally likely, so the probability is favorable sets over all sets.

  3. Probability: 610=0.6.

    This is the hypergeometric answer.

  4. Compare with the binomial formula used by mistake: (32)(0.6)2(0.4)=0.432.

    Treating p=35 as fixed is off by almost 0.17, because it ignores that each red drawn makes the next red less likely.

Answer: Not binomial — it's hypergeometric

Takeaway: Drawing without replacement from a small group changes p on every draw; count sets (hypergeometric) instead of using the binomial formula.

Example 3 — Find the mistake: the missing arrangements

Application

A quiz has 4 multiple-choice questions, each with 4 options, and a student guesses every answer. Find the probability of exactly 2 correct. Tara writes: "P=(0.25)2(0.75)2≈0.035." Find the mistake and correct it.
Show the steps and the answer
  1. Check the setup: n=4 guesses, p=14=0.25, independent, so binomial.

    Tara's model and her values of p and 1−p are right.

  2. Read what (0.25)2(0.75)2 stands for: one specific order, such as right, right, wrong, wrong.

    It is the probability of a single arrangement, not of every way to get 2 right.

  3. Count the arrangements: (42)=6 (RRWW, RWRW, RWWR, WRRW, WRWR, WWRR).

    Any 2 of the 4 questions could be the ones answered correctly.

  4. Multiply: 6×0.03515625≈0.211.

    There are six equally likely arrangements, so the true probability is six times Tara's.

Answer: P(X=2)=(42)(0.25)2(0.75)2≈0.211

Takeaway: pk(1−p)n−k covers one ordering; multiply by (nk) to count them all.

Where You'll See It

  • Seed packets and germination rates

    A packet promises 80% germination. Plant 10 seeds and the number that sprout is binomial with n=10, p=0.8. The chance that all 10 sprout is 0.810≈0.107, so gardeners sow extra seeds instead of expecting a perfect row.

  • Airline overbooking

    If each ticketed passenger shows up with probability 0.9, independently, then on a 100-seat flight with 105 tickets sold the number who show up is binomial with n=105, p=0.9 and mean 105×0.9=94.5. Airlines use the full distribution to estimate how often more than 100 people will arrive.

  • Spotting a faulty machine

    If 2% of light bulbs are defective, a box of 20 has no defects with probability 0.9820≈0.67, and 3 or more defects with probability only about 0.007. An inspector who finds 3 bad bulbs in one box therefore suspects the machine, not bad luck.

Common Mistakes

  1. The slip: Applying it when trials aren't independent or p changes

    Instead: drawing without replacement is hypergeometric, not binomial.

  2. The slip: Dropping the (nk) factor

    Instead: you must count the orderings of the k successes, not just multiply pk(1−p)n−k.

  3. The slip: Mismatching the exponents

    Instead: the success power is k and the failure power is n−k; swapping them inverts the answer.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. A fair coin is flipped 4 times. What is the probability of exactly 2 heads?

    Hint: Count the arrangements of 2 heads, then multiply by the chance of each one.

  2. A fair die is rolled 3 times. What is the probability of exactly one six?

    Hint: Use n=3, k=1, p=16.

  3. A fair coin is flipped 100 times. Find the mean and standard deviation of the number of heads.

    Hint: Use μ=np and σ=np(1−p).

  4. You guess on 5 true-false questions. What is the probability of getting at least 4 right?

    Hint: Add the probabilities of exactly 4 and exactly 5.

  5. You roll a die until a six appears and count how many rolls it takes. Is that count binomial?

    Hint: Check whether the number of trials is fixed in advance.

  6. For exactly 2 heads in 3 flips, Max writes: "(0.5)2(0.5)1=0.125." Correct his work.

    Hint: How many different orders give exactly 2 heads?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Is the most likely number of successes always exactly np?

Not always, because npnp need not be a whole number. With 10 trials at p=0.35p=0.35, np=3.5np=3.5, but you cannot get 3.5 successes. The mean is the long-run average count, and the most likely single count sits next to it: when (n+1)p(n+1)p is not a whole number, it is ⌊(n+1)p⌋\lfloor (n+1)p\rfloor, here ⌊3.85⌋=3\lfloor 3.85\rfloor=3.

How spread out is a binomial count, and when is it most spread out?

Its standard deviation is np(1−p)\sqrt{np(1-p)}, and p(1−p)p(1-p) is largest at p=0.5p=0.5, where it equals 0.25. A fair coin is the hardest to predict, since each flip could easily go either way. When pp is near 0 or 1, almost every trial gives the same result, so the count barely changes from run to run. For 100 trials the SD is 5 at p=0.5p=0.5 but only 3 at p=0.1p=0.1.

How do airlines use the binomial distribution?

To decide how far to overbook. If each ticket holder independently has a 95% chance of showing up, the number who show up out of 180 tickets is binomial with n=180n=180 and p=0.95p=0.95, with mean 171. An airline with 175 seats can find the chance that more than 175 arrive and judge whether selling 180 is worth it. Families travelling together break independence, so real models adjust for groups.

What is the quickest way to find the chance of at least one success?

Use the complement instead of adding many terms: P(X≥1)=1−P(X=0)=1−(1−p)nP(X\ge1)=1-P(X=0)=1-(1-p)^n. For 6 rolls of a die, the chance of at least one 6 is 1−(56)6≈1−0.335=0.6651-\left(\frac{5}{6}\right)^6\approx1-0.335=0.665. Adding P(X=1)P(X=1) through P(X=6)P(X=6) gives the same answer with six times the work.

Learning Path

Binomial Distribution

You are here

You need the Binomial Coefficient (counting arrangements), Probability, and Independent Events (multiplying chances of separate trials) first. The binomial distribution combines them into one formula for the chance of k successes in n trials, with mean np. Next, the Normal Distribution approximates the binomial when n is large, so a question like "at least 60 heads in 100 flips" can be answered quickly.

See Also