Visual Tools
Calculators
Tables
Mathematical Keyboard
Converters
Other Tools

How to Find a Poisson Probability

A Poisson variable counts events that occur independently at a steady average rate over an interval of time or space. Scale the rate to the interval in the question, then each count has probability e−λλk/k!e^{-\lambda}\lambda^k / k!.

Recognize the task

An average rate is given and the question asks about a count in some interval, with no fixed number of trials: "33 calls per hour on average", "22 typos per page", "0.50.5 accidents per week". Test: events happen one at a time, independently, at a constant average rate, and there is no upper limit on the count. Then X∼Poisson⁡(λ)X \sim \operatorname{Poisson}(\lambda) with λ\lambda = expected count in the interval.
The model is defined in the Poisson distribution lesson.

Two ways to do it

Add PMF termsexact or at most
P(X=k)=e−λλkk!P(X = k) = \frac{e^{-\lambda}\lambda^k}{k!}
Evaluate the Poisson PMF at each count; factor out e−λe^{-\lambda} when you add several.
Use the complementat least
P(X≥k)=1−∑j=0k−1e−λλjj!P(X \ge k) = 1 - \sum_{j=0}^{k-1} \frac{e^{-\lambda}\lambda^j}{j!}
The counts go on forever, so "at least" always needs the complement rule.

Steps

StepOn 33 calls per hour, P(exactly 2 calls in an hour)P(\text{exactly } 2 \text{ calls in an hour})
1Scale the rate to the interval. λ\lambda is the expected count in the interval the question asks about: rate times length.λ=3⋅1=3\lambda = 3 \cdot 1 = 3
2Write the event as counts. "Exactly kk" is one term, "at most kk" is 0,1,…,k0, 1, \ldots, k, "at least kk" is the complement of 0,…,k−10, \ldots, k - 1.X=2X = 2
3Evaluate the PMF at each count. Use e−λλk/k!e^{-\lambda}\lambda^k / k! and remember 0!=10! = 1.e−3⋅322!=4.5 e−3e^{-3} \cdot \frac{3^2}{2!} = 4.5\, e^{-3}
4Add, or subtract from 11, and round. Round only the final value.P(X=2)≈0.2240P(X = 2) \approx 0.2240

Worked examples

Example 1 · no events at all

k = 0
22 typos per page on average; P(a page with no typo)P(\text{a page with no typo})
λ=2\lambda = 2
P(X=0)=e−2⋅200!=e−2P(X = 0) = \frac{e^{-2} \cdot 2^0}{0!} = e^{-2}
20=0!=12^0 = 0! = 1
P(X=0)≈0.1353P(X = 0) \approx 0.1353

Example 2 · a different interval

rescale
33 calls per hour; P(at least one call in 30 min)P(\text{at least one call in } 30 \text{ min})
λ=3⋅12=1.5\lambda = 3 \cdot \frac{1}{2} = 1.5
half an hour
P(X≥1)=1−e−1.5≈0.7769P(X \ge 1) = 1 - e^{-1.5} \approx 0.7769
complement of X=0X = 0

Example 3 · at most k events

short sum
λ=4\lambda = 4 arrivals per hour; P(X≤2)P(X \le 2)
counts 0,1,20, 1, 2
e−4(1+4+422)=13 e−4e^{-4}\left(1 + 4 + \frac{4^2}{2}\right) = 13\, e^{-4}
factor out e−4e^{-4}
P(X≤2)≈0.2381P(X \le 2) \approx 0.2381
Check your answer in the discrete distributions calculator.

See it happen

Split the interval into many tiny pieces, each with a tiny chance of holding one event. The count is then a binomial with huge nn and tiny pp, and in the limit that binomial becomes the Poisson formula with λ=np\lambda = np. Doubling the interval doubles the expected count, which is why λ\lambda must be rescaled. Watch the shape change with λ\lambda in the Poisson distribution visualizer.

Common mistakes

Not rescaling the rate

✗ 33 per hour, 3030 min: P(X≥1)=1−e−3≈0.9502P(X \ge 1) = 1 - e^{-3} \approx 0.9502
✓ λ=1.5\lambda = 1.5 for half an hour: 1−e−1.5≈0.77691 - e^{-1.5} \approx 0.7769

Wrong complement of "at least one"

✗ P(X≥1)=1−P(X=1)P(X \ge 1) = 1 - P(X = 1)
✓ P(X≥1)=1−P(X=0)P(X \ge 1) = 1 - P(X = 0)

Forgetting 0!0! or e−λe^{-\lambda} in a sum

✗ P(X≤2)=e−4+4+8P(X \le 2) = e^{-4} + 4 + 8
✓ e−4e^{-4} multiplies every term: e−4(1+4+8)e^{-4}(1 + 4 + 8)

Poisson with a fixed number of trials

✗ 1010 tosses, count the heads with λ=5\lambda = 5
✓ a fixed nn and a large pp make it binomial
Poisson stands in for a binomial only when nn is large and pp small; see Poisson approximation.