What this calculator does
The exponential distribution models the waiting time between independent events that occur at a constant average rate. It is the continuous counterpart of the Poisson process and appears in queueing, reliability, radioactive decay, and service times. The distribution is memoryless: if you have already waited ten minutes, the chance of waiting another five is the same as it was at the start. That property makes the math simple and makes the model a poor fit for wear-out failures or appointments that bunch together. This calculator takes the mean time between events and a time of interest, then reports the cumulative probability, the survival probability, the density, and the median and 90th percentile. It is a teaching and planning tool, not a prediction of a specific event.
How to use it
- Enter the mean time between events in the unit you are using.
- Enter the time of interest in the same unit.
- Read P(X <= x) as the chance of waiting at most x.
- Read P(X > x) as the chance of waiting longer than x.
- Use the median and 90th percentile to describe typical and long waits.
- Check that the constant-rate and independence assumptions fit your situation.
How it works
Example calculations
Example 1: A mean wait of 4 units
With a mean of 4.00, the rate is 0.2500 per unit. The chance of waiting at most 2.00 units is 39.35%, and the chance of waiting longer is 60.65%. The median wait is 2.77 units, and the 90th percentile is 9.21 units.
Example 2: A mean wait of 10 units
With a mean of 10.00, the rate is 0.1000. The chance of waiting at most 5.00 units is 39.35%. Because the mean is longer, the same time of interest is less likely to be enough. The median is 6.93 units.
Example 3: A fast process with a mean of 0.5
With a mean of 0.50, the rate is 2.0000 per unit. The chance of waiting at most 0.10 units is 18.13%. The median is 0.35 units, and the 90th percentile is 1.15 units. Fast processes have short typical waits but still have a long tail.
What the result means
The cumulative probability is the answer to "how likely is a wait of this length or less?" In the default case there is a 39.35% chance of waiting at most 2.00 units, which is below the median wait of 2.77. The survival probability is the complement, and it is often the more useful number for planning capacity: the chance that a customer, a machine, or a call waits longer than the service window. The long right tail is a feature of the distribution, not a data error. The mean is larger than the median because a few very long waits pull it up. If your data show a decreasing hazard or a wear-out pattern, the exponential model is the wrong tool.
Common mistakes
- Confusing the mean with the rate; lambda is 1 divided by the mean.
- Using the distribution for a process whose rate changes over time.
- Assuming the memoryless property applies to wear-out or appointment-based arrivals.
- Reading P(X <= x) as the probability of exactly x rather than at most x.
- Mixing time units between the mean and x.
- Treating the 90th percentile as a maximum rather than a threshold exceeded 10% of the time.
Assumptions and limits
- Events occur independently at a constant average rate.
- The mean time between events is known and stable.
- The time of interest is measured in the same unit as the mean.
- The waiting time is continuous and non-negative.
- The process is memoryless; past waiting does not change future probabilities.
- The result is a model, not a guarantee for a specific event.
Frequently asked questions
What is the difference between the mean and the rate?
The mean is the average time between events, such as 4 minutes. The rate is the number of events per unit time, such as 0.25 per minute. The rate is 1 divided by the mean. Mixing the two is the most common source of errors.
Why is the median smaller than the mean?
The exponential distribution is right-skewed. A few very long waits pull the mean above the median. For any exponential distribution, the median is ln(2) times the mean, about 69.3% of the mean, while the mean is the mean.
What does memoryless mean?
It means the probability of waiting an additional time does not depend on how long you have already waited. If the mean wait is 4 minutes, the chance of waiting another 2 minutes is the same after 10 minutes as at the start. That is a strong assumption and often unrealistic for real queues.
Can I use this for reliability or failure times?
Only if the failure rate is constant, which is rare for mechanical wear-out. Electronic components sometimes have a long constant-rate period, but many failures increase with age. Use a Weibull or another life distribution when the hazard changes over time.
How do I find the probability of exactly one value?
For a continuous distribution, the probability of any exact value is zero. The density f(x) describes the relative likelihood near x, not the probability of x. Use the cumulative or survival probability for a range.
What is the 90th percentile used for?
It is the wait that is exceeded only 10% of the time. In the default case it is ${number(exp.p90, 2)} units. It is useful for service-level planning, such as choosing a capacity that covers all but the longest 10% of waits.
Sources
2026 update
How to use the professional output
This page is not just a formula box. The result is designed to be checked, compared, and taken away.
- Professional worksheet. The result includes the inputs used, the method and scope, and a one-variable sensitivity check for finance calculators.
- A/B scenarios. Save two input sets in the browser and compare the primary, secondary, and breakdown figures side by side.
- Print / PDF and CSV. Print a branded worksheet, or export the year-by-year table when the calculator produces one.
The formulas and assumptions above still govern the number. The worksheet makes the estimate auditable; it does not make it a lender quote, tax return, or personalized advice.