Visual Tools
Calculators
Tables
Mathematical Keyboard
Converters
Other Tools

How to Find an Exponential Probability

An exponential variable is the waiting time until the next event when events occur at a constant rate λ\lambda. Its CDF has a closed form, so every probability is one or two values of e−λxe^{-\lambda x}.

Recognize the task

A waiting time or lifetime is described by an average or a rate: "customers arrive on average every 1010 minutes", "a bulb lasts 10001000 hours on average", "0.20.2 failures per year". Test: the time is continuous, starts at 00, and the chance of the event in the next moment does not depend on how long you have waited. Then X∼Exp⁡(λ)X \sim \operatorname{Exp}(\lambda) with λ=1/mean\lambda = 1/\text{mean}.
The model is defined in the exponential distribution lesson.

Two ways to do it

Use the CDF and the tailstandard
P(X≤x)=1−e−λx,P(X>x)=e−λxP(X \le x) = 1 - e^{-\lambda x}, \qquad P(X > x) = e^{-\lambda x}
The exponential CDF answers any interval with two values.
Integrate the densityfrom scratch
P(c<X<d)=∫cdλe−λx dx=e−λc−e−λdP(c < X < d) = \int_c^d \lambda e^{-\lambda x}\, dx = e^{-\lambda c} - e^{-\lambda d}
The same result straight from the exponential PDF; use it when the CDF is not given.

Steps

StepOn arrivals every 1010 minutes on average, P(wait>15 min)P(\text{wait} > 15 \text{ min})
1Find the rate λ\lambda. The mean is 1/λ1/\lambda (exponential mean), so λ=1/mean\lambda = 1/\text{mean}. Keep the time unit.λ=110\lambda = \frac{1}{10} per minute
2Write the event with the CDF or the tail. "Within xx" uses 1−e−λx1 - e^{-\lambda x}, "longer than xx" uses e−λxe^{-\lambda x}, an interval uses e−λc−e−λde^{-\lambda c} - e^{-\lambda d}.P(X>15)=e−λ⋅15P(X > 15) = e^{-\lambda \cdot 15}
3Substitute in the same units. Rate and time must use the same unit, so λx\lambda x has no unit.e−0.1⋅15=e−1.5e^{-0.1 \cdot 15} = e^{-1.5}
4Evaluate and round. For a conditional "already waited" question, the memoryless property removes the time already spent.P(X>15)≈0.2231P(X > 15) \approx 0.2231

Worked examples

Example 1 · within a time

CDF
mean 1010 min, P(X≤5)P(X \le 5)
λ=0.1\lambda = 0.1
1−e−0.1⋅5=1−e−0.51 - e^{-0.1 \cdot 5} = 1 - e^{-0.5}
P(X≤5)≈0.3935P(X \le 5) \approx 0.3935

Example 2 · between two times

interval
mean 1010 min, P(5<X<15)P(5 < X < 15)
e−0.5−e−1.5e^{-0.5} - e^{-1.5}
tail at 55 minus tail at 1515
≈0.60653−0.22313=0.38340\approx 0.60653 - 0.22313 = 0.38340
P(5<X<15)≈0.3834P(5 < X < 15) \approx 0.3834

Example 3 · after waiting

memoryless
mean 1010 min, waited 1010; P(X>15∣X>10)P(X > 15 \mid X > 10)
e−1.5e−1=e−0.5\frac{e^{-1.5}}{e^{-1}} = e^{-0.5}
X>15X > 15 implies X>10X > 10
=P(X>5)≈0.6065= P(X > 5) \approx 0.6065
only the extra 55 min matter
Check your answer in the continuous distributions calculator.

See it happen

The survival curve e−λxe^{-\lambda x} loses the same fraction of what is left in every unit of time, which is why waiting gives no credit: after any delay the remaining curve has the same shape as at the start. The area under the density between two times is the drop of that curve between them. Move λ\lambda in the exponential distribution visualizer.

Common mistakes

Using the mean as the rate

✗ mean 1010 min, so P(X>15)=e−10⋅15P(X > 15) = e^{-10 \cdot 15}
✓ λ=110\lambda = \frac{1}{10}: P(X>15)=e−1.5≈0.2231P(X > 15) = e^{-1.5} \approx 0.2231

Swapping the CDF and the tail

✗ P(X>15)=1−e−1.5≈0.7769P(X > 15) = 1 - e^{-1.5} \approx 0.7769
✓ 1−e−λx1 - e^{-\lambda x} is "within xx"; "longer than xx" is e−λxe^{-\lambda x}

Mixed time units

✗ λ=6\lambda = 6 per hour with x=15x = 15 minutes: e−90e^{-90}
✓ convert first: 1515 min =0.25= 0.25 h, e−6⋅0.25=e−1.5e^{-6 \cdot 0.25} = e^{-1.5}

Subtracting the time already waited from the wrong place

✗ P(X>15∣X>10)=P(X>15)P(X > 15 \mid X > 10) = P(X > 15)
✓ P(X>15∣X>10)=P(X>5)P(X > 15 \mid X > 10) = P(X > 5) (memoryless property)