On this concept: Formula

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

Math · Statistics & Probability · Grade 9-12

Geometric Distribution

The geometric distribution models the number of independent Bernoulli trials needed to get the FIRST success, each trial succeeding with probability p.

P(X=k)=(1−p)k−1⋅p,k=1,2,3,…
Venn diagram of Event A and Event BA two-event view of geometric distribution.ABA ∩ B
A: Event A
B: Event B
A two-event view of geometric distribution.

Quick Answer

The geometric distribution models the number of independent Bernoulli trials needed to get the FIRST success, each trial succeeding with probability p. A Bernoulli trial is a single try with two outcomes. Use the geometric distribution when you repeat the same independent try until something happens for the first time: rolling until a 6, calling customers until one says yes, shooting until the first basket. The chance the first success comes on try k is (1−p)k−1p, and on average you wait 1p tries.

Why This Matters

Waiting questions are everywhere: how many calls until a sale, how many packs until the rare card, how many attempts until a part fails. The geometric distribution turns them into exact probabilities and a mean wait of 1p, so a 1-in-20 chance means about 20 tries on average, with a long tail of unlucky runs. Using the binomial instead answers a different question and gives wrong numbers.

Intuitive Explanation

Roll a die until a 6 appears. The first 6 comes on roll 1 with probability 16. To come first on roll 2, roll 1 must fail and roll 2 succeed: 56⋅16=536. On roll 3 it is (56)216. The pattern is P(X=k)=(1−p)k−1p: k−1 failures in a row, then one success. Each extra roll multiplies by 56, so the probabilities shrink steadily but never reach zero.

The single most likely outcome is always k=1, yet the average wait is E(X)=1p=6 rolls. Both are true because the tail is long: waits of 10, 15 or 20 rolls are uncommon but possible, and they pull the average up. The chance of waiting more than k rolls has a neat form, (1−p)k, because it means the first k rolls all fail; for 10 rolls that is (56)10≈0.16.

The die has no memory. After 5 failures, the chance of a 6 on the next roll is still 16, and the expected remaining wait is still 6 rolls. Believing a 6 is "due" is the gambler's fallacy.

The geometric and binomial distributions both use p and 1−p but ask different questions. Binomial: in a fixed 10 rolls, how many 6s? Geometric: how many rolls until the first 6? A fixed number of trials means binomial; stopping at the first success means geometric.

Think of it like this. Waiting for a geometric success resembles waiting for a bus that arrives completely at random, each minute carrying the same chance. Having already waited ten minutes doesn't make the next minute any luckier, so the wait so far says nothing about the wait to come. Real buses run on timetables, so the comparison holds only for truly random arrivals.

Core idea. The geometric distribution gives the probability that the first success occurs on trial k, when independent trials each succeed with probability p.


When to Use

Use the geometric distribution when you repeat independent trials with a constant chance p and count the trials until the first success: "how many tries until…", "first success on the kth attempt", "expected number of attempts." The count can be 1, 2, 3 and so on with no upper limit. If the number of trials is fixed and you count successes, use the binomial. If the chance changes from try to try, as when drawing without replacement, neither model applies directly.

Tip. Ask: Am I counting the number of trials up to and including the first success (not the number of successes in a fixed set of trials)?

How to Recognize It

Before choosing the geometric model, check when the trials stop, whether every try has the same chance, and exactly which count the question wants.

  1. Is the number of tries decided in advance, such as exactly 10 free throws?

    If yes, the geometric model does not fit: count the successes in those tries with the binomial distribution. If the tries simply go on until the first success, the count is geometric and has no upper limit.

  2. Is the chance of success the same on every try, as with rolling a fair die or drawing with replacement?

    If yes, the geometric formula applies. If drawing without replacement changes the chance after each try, the trials are dependent, and you must multiply the changing probabilities step by step instead.

  3. Does the variable count the try on which the first success lands, rather than only the failures before it?

    If yes, use (1−p)k−1p and the mean 1p. Some books count only the failures, X−1, whose mean is 1−pp, so read how the count is defined.

  4. Does the question want only the average number of tries, rather than the chance of one particular count?

    If yes, use E(X)=1p: with p=0.2, expect 5 tries. If it asks for the chance of needing more than k tries, use (1−p)k, since the first k must all fail.

Geometric Distribution compared with Binomial distribution, Expected value, and Exponential distribution

All four rows describe repeated chance events, so they are easy to swap. Ask what is being counted: trials until the first success (geometric), successes in a fixed number of trials (binomial), only the long-run average (expected value), or continuous waiting time (exponential).

Geometric Distribution

Meaning
Use this when independent identical trials are repeated and you count how many until the first success occurs.
Key test
Am I counting the number of trials up to and including the first success (not the number of successes in a fixed set of trials)?
Formula
P(X=k)=(1−p)k−1⋅p,k=1,2,3,…
Example
You roll a fair die until you get a 6. What is the probability the first 6 appears on the 4th roll?

Binomial distribution

Meaning
Counts the number of successes in a FIXED number n of trials.
Key test
Use when n is set in advance and you count how many succeed.
Formula
P(X=k)=(nk)pk(1−p)n−k
Example
Number of heads in exactly 10 flips

Expected value

Meaning
A single summary number (here E(X)=1/p), not the full distribution of outcomes.
Key test
Use when you only need the average wait, not each probability.
Formula
E(X)=1p
Example
On average 6 rolls to get a 6

Exponential distribution

Meaning
The continuous waiting-time analogue, for time until an event rather than count of trials.
Key test
Use when time is continuous, not discrete trials.
Formula
f(t)=λe−λt
Example
Minutes until the next bus

Formula & Notation

P(X=k)=(1−p)k−1⋅p,k=1,2,3,…
P(X=k)=(1−p)k−1p for k=1,2,3,…; E(X)=1p, Var(X)=1−pp2

How to read it: X∼Geom(p). Mean: E(X)=1p. Standard deviation: σ=1−pp.

Worked Examples

Example 1 — First six on a die

Easy

You roll a fair die until you get a 6. What is the probability the first 6 appears on the 4th roll?
Show the steps and the answer
  1. Translate the event: rolls 1, 2 and 3 are not 6, and roll 4 is a 6.

    'First 6 on roll 4' means exactly three failures followed by one success.

  2. Multiply the chances: 56⋅56⋅56⋅16.

    The rolls are independent, so the probability of the whole sequence is the product.

  3. Simplify: (56)3⋅16=125216⋅16=1251296≈0.0965.

    This is (1−p)k−1p with p=16 and k=4.

Answer: 1251296≈0.0965

Takeaway: First success on try k means k−1 failures, then a success: (1−p)k−1p.

Example 2 — Sixes in ten rolls

Standard

Now you roll the die exactly 10 times and ask for the probability of getting exactly two 6s. Is that geometric?
Show the steps and the answer
  1. Notice that the number of rolls is fixed at 10 and the question counts 6s.

    A geometric count stops at the first success; here all 10 rolls happen no matter what.

  2. Set up the binomial: (102)(16)2(56)8.

    The two 6s can fall on any 2 of the 10 rolls, so the arrangements must be counted.

  3. Compute: 45×136×0.2326≈0.291.

    About 29% of 10-roll sessions contain exactly two 6s.

Answer: Not geometric — it's a binomial probability

Takeaway: A fixed number of trials with a count of successes is binomial; stopping at the first success is geometric.

Example 3 — Find the mistake: one failure too many

Application

A basketball player makes each free throw with probability 0.7, independently. Find the probability that her first make comes on her 3rd shot. Rosa writes: "P(X=3)=(0.3)3(0.7)=0.0189." Find the mistake and correct it.
Show the steps and the answer
  1. Write out the sequence: miss, miss, make.

    A first make on shot 3 needs exactly 2 misses before it, not 3.

  2. Read Rosa's formula: (0.3)3(0.7) is miss, miss, miss, make.

    That is the probability of a first make on shot 4.

  3. Fix the exponent: P(X=3)=(0.3)2(0.7)=0.09×0.7=0.063.

    The exponent is k−1=2, one less than the shot number.

  4. Check: P(X=1)+P(X=2)+P(X=3)=0.7+0.21+0.063=0.973, and 1−0.33=0.973.

    The chance of a make within 3 shots must equal 1 minus the chance of missing all three, and it does.

Answer: P(X=3)=(0.3)2(0.7)=0.063

Takeaway: The exponent on 1−p counts the failures before the success, which is k−1.

Where You'll See It

  • Sales calls until the first yes

    If 1 call in 8 leads to a sale, independently, the number of calls until the first sale is geometric with p=0.125. The salesperson should expect about 10.125=8 calls, but there is a 0.87515≈13% chance of 15 calls in a row without a sale, so a dry spell doesn't mean something is wrong.

  • Hunting for a rare trading card

    If a rare card turns up in 1 pack out of 50, the expected number of packs until it appears is 50. Yet the chance of still not having it after 100 packs is 0.98100≈0.13, so some collectors buy well over 100 packs.

  • Rolling doubles to leave jail

    In Monopoly, two dice show doubles with probability 16. A player in jail gets three tries, so the chance of rolling out is 1−(56)3≈0.42; without the three-try limit, the average wait would be 6 turns.

Common Mistakes

  1. The slip: Using the binomial because both involve p and 1−p

    Instead: the geometric has no fixed n; it counts trials until the first success.

  2. The slip: Writing the exponent as k instead of k−1

    Instead: the first k−1 trials are failures, so it's (1−p)k−1⋅p.

  3. The slip: Forgetting the trials must be independent with constant p

    Instead: if p changes or trials are dependent, the geometric model fails.


Mini Practice

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

  1. You flip a fair coin until you get heads. What is the probability that the first head appears on the 3rd flip?

    Hint: The first two flips must be tails.

  2. A carnival game is won 20% of the time, and each try is independent. What is the probability that your first win comes on your 4th try?

    Hint: Three losses come before the win.

  3. Each cereal box has a 0.25 chance of holding a prize. On average, how many boxes will you open to get your first prize?

    Hint: The mean of a geometric distribution is 1p.

  4. A carnival game is won with probability 0.2 on each independent try. What is the probability that you win at least once within your first 3 tries?

    Hint: Find the chance of losing all 3 times, then use the complement.

  5. A quiz has 8 multiple-choice questions, and a student guesses each with p=0.25. You want P(exactly 2 correct). Geometric or binomial?

    Hint: Is the number of trials fixed, or do you stop at the first success?

  6. For the probability that the first 6 on a fair die appears on the 3rd roll, Ana writes (56)3⋅16. Correct her work.

    Hint: How many rolls come before the 6?

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

If I have already failed 10 times in a row, isn't a success now due?

No. Each trial is independent, so a die has no memory of the last 10 rolls. The chance of a 6 on the next roll is still 16\frac{1}{6}, and the expected number of further rolls is still 6. This is the memoryless property, and among distributions on the counting numbers the geometric is the only one that has it. Believing a success is overdue is called the gambler's fallacy.

Why does the average wait equal 1p?

If an event happens on a fraction pp of trials, then over a long run successes appear about once every 1p\frac{1}{p} trials, so the gaps between them average 1p\frac{1}{p}. For a fair die, p=16p = \frac{1}{6}, and a 6 shows up about once every 6 rolls. The algebra agrees: summing the series E(X)=∑k=1∞k(1−p)k−1pE(X) = \sum_{k=1}^{\infty} k(1-p)^{k-1}p also gives exactly 1p\frac{1}{p}.

What is the most likely number of trials for the first success?

Always 1, whatever pp is. The probabilities p,(1−p)p,(1−p)2p,…p, (1-p)p, (1-p)^2p, \ldots shrink by a factor of 1−p1-p at every step, so the first one is the largest. Even when waiting for a 6, the single most likely outcome is success on roll 1, with probability about 0.167, although the average wait is 6 rolls. The distribution is strongly skewed right: most waits are short, and a few are very long.

Is the geometric distribution related to geometric sequences?

Yes, that is where its name comes from. The probabilities pp, p(1−p)p(1-p), p(1−p)2p(1-p)^2, and so on form a geometric sequence with common ratio 1−p1-p. Adding them is an infinite geometric series, p1−(1−p)=pp=1\frac{p}{1-(1-p)} = \frac{p}{p} = 1, which confirms that the probabilities of all possible waits add up to 1. The same ratio is why each bar in its graph is a fixed fraction of the bar before it.

Learning Path

Geometric Distribution

You are here

Next →

Nothing after this yet
You need the Binomial Distribution, Independent Events and Expected Value first. The geometric distribution keeps the yes/no trial with constant p but changes the question to how long until the first success, giving (1−p)k−1p and a mean wait of 1p. It prepares you for waiting-time problems in queues and reliability, the negative binomial (waiting for the rth success), and the exponential distribution for continuous waiting times.

See Also