Formula Used To Calculate Sum Of Poisson Random Variables

Use this interactive calculator to find the combined rate, mean, variance, and exact event probability when independent Poisson random variables are added together. The core rule is simple: if independent counts follow Poisson distributions, their sum is also Poisson with parameter equal to the sum of the individual rates.

Interactive Calculator

Formula Summary

If X1 ~ Poisson(λ1) and X2 ~ Poisson(λ2) are independent, then:

S = X1 + X2 ~ Poisson(λ1 + λ2)

More generally, if Xi ~ Poisson(λi) are independent, then:

S = ΣXi ~ Poisson(Σλi)

For identical rates, if Xi ~ Poisson(λ) for i = 1, …, n:

S ~ Poisson(nλ)

The exact point probability is:

P(S = k) = e^-λS λS^k / k!

Important assumption: this closure property requires independence. If the variables are dependent, the sum need not be Poisson.

Quick interpretation: the total rate, expected value, and variance all equal the combined λ for a Poisson sum.

What formula is used to calculate the sum of Poisson random variables?

The formula used to calculate the sum of Poisson random variables is one of the most elegant results in probability theory. If several random variables are independent and each one follows a Poisson distribution, then their sum also follows a Poisson distribution. The new parameter is simply the sum of the original rate parameters. In notation, if X₁, X₂, …, X_n are independent and X_i ~ Poisson(λ_i), then the total S = X₁ + X₂ + … + X_n satisfies S ~ Poisson(λ₁ + λ₂ + … + λ_n). This is often called the additivity or closure property of the Poisson distribution.

This result matters because Poisson models are widely used for count data such as incoming calls, website visits, machine defects, arrivals at a service desk, photons detected by an instrument, and biological event counts over time. In many real systems, counts come from multiple independent sources. Rather than model each source separately forever, analysts often want the distribution of the total. The Poisson sum formula gives the answer immediately.

Core rule: For independent Poisson random variables, the rate parameters add. That means the combined mean is the sum of the individual means, and the combined variance is also the sum of the individual variances.

Why the sum stays Poisson

The Poisson distribution is special because it is stable under addition when independence holds. There are several ways to prove this, but the most common are through the probability generating function, the moment generating function, or direct convolution. The generating function approach is especially clean. The probability generating function of a Poisson(λ) random variable is G_X(t) = exp(λ(t – 1)). For independent variables, the generating function of the sum is the product of the individual generating functions. Therefore:

G_S(t) = Π exp(λ_i(t – 1)) = exp((Σ λ_i)(t – 1)).

That is exactly the generating function of a Poisson random variable with parameter Σ λ_i. So the sum must be Poisson with the combined rate.

The exact formula

• If X ~ Poisson(λ), then P(X = k) = e^-λ λ^k / k!, for k = 0, 1, 2, …
• If S = ΣX_i and the X_i are independent Poisson(λ_i), then S ~ Poisson(Σλ_i).
• Therefore P(S = k) = e^-Σλi (Σλ_i)^k / k!.
• The expected value is E[S] = Σλ_i.
• The variance is Var(S) = Σλ_i.

Special case: sum of identical Poisson variables

A very common question is what happens if all the Poisson random variables have the same rate λ. In that case, if X₁, X₂, …, X_n are independent and each one is Poisson(λ), then their sum is Poisson(nλ). This shows up in operations research, actuarial models, queueing, epidemiology, reliability analysis, and telecommunications. For example, if four independent counters each record failures at an average rate of 1.5 per hour, then the total failures across the four counters during one hour follow Poisson(6).

Step by step method for calculating the sum

1. Identify each independent Poisson random variable and its rate λ.
2. Verify independence. Without independence, the simple addition rule may fail.
3. Add the rates together to get the total parameter: λ_S = Σλ_i.
4. State the total distribution: S ~ Poisson(λ_S).
5. If needed, compute exact probabilities with P(S = k) = e^-λS λ_S^k / k!.
6. If needed, compute cumulative probabilities by summing point probabilities from 0 up to the desired threshold.

Worked example with numerical interpretation

Suppose a support center receives chatbot escalations from two independent systems. System A has an average of 2.4 escalations per hour, and System B has an average of 3.1 per hour. Let X ~ Poisson(2.4) and Y ~ Poisson(3.1). Since X and Y are independent, the total S = X + Y follows Poisson(5.5). That means the expected total number of escalations in one hour is 5.5, and the variance is also 5.5.

If you want the probability of exactly 5 total escalations, you use the Poisson formula with λ = 5.5 and k = 5. The result is P(S = 5) = e^-5.5 5.5⁵ / 5! ≈ 0.1714. This is the practical power of the sum rule: instead of convolving two separate distributions manually, you jump straight to one Poisson calculation.

Comparison table: adding Poisson rates

Scenario	Individual Rates	Combined Rate λS	Mean of Sum	Variance of Sum	Example Point Probability
Two moderate streams	2.4 and 3.1	5.5	5.5	5.5	P(S = 5) ≈ 0.1714
Three small streams	0.8, 1.1, 1.6	3.5	3.5	3.5	P(S = 3) ≈ 0.2158
Four identical stations	1.5, 1.5, 1.5, 1.5	6.0	6.0	6.0	P(S = 6) ≈ 0.1606
High volume aggregate	4.2, 5.0, 3.8	13.0	13.0	13.0	P(S = 13) ≈ 0.1102

When the formula applies and when it does not

The formula for the sum of Poisson random variables depends critically on independence. This is the condition users most often overlook. If two count processes influence one another, share a hidden common driver, or are measured in overlapping ways, then adding their λ values may not produce a correct Poisson model for the total. In real projects, dependence can arise from seasonality, shifts in demand, common weather conditions, synchronized system behavior, or duplicate event logging.

Use the sum formula when:

• Each count process is modeled as Poisson.
• The processes are independent over the interval of interest.
• The rates are defined over the same time window, area, or exposure unit.
• You want the distribution of the total count from all sources combined.

Be careful when:

• The variables are correlated or share common shocks.
• The rates change during the interval and are not homogeneous.
• The data are overdispersed relative to Poisson assumptions.
• Counts include structural zeros, truncation, or reporting artifacts.

Poisson sum versus normal approximation

As the combined rate gets larger, the Poisson distribution becomes more symmetric and can often be approximated by a normal distribution with mean λ_S and variance λ_S. However, the exact Poisson formula is still preferred when precision matters, especially for small or moderate rates and tail probabilities. The table below compares exact Poisson probabilities with normal approximations using continuity correction for selected values.

Combined λS	Target k	Exact Poisson P(S = k)	Approx. Normal Probability	Absolute Difference
5.5	5	0.1714	0.1669	0.0045
10.0	10	0.1251	0.1234	0.0017
15.0	15	0.1024	0.1018	0.0006
20.0	18	0.0848	0.0841	0.0007

How this formula is used in real analysis

In practice, the sum of Poisson random variables is used whenever independent count streams are aggregated. In a hospital setting, different departments may contribute independent arrival counts to a shared triage unit. In a cloud system, separate microservices may generate independent alert counts that need to be modeled as a total incident load. In manufacturing, defects from multiple production lines may be combined to estimate the distribution of total defects per shift. In insurance, claims from different independent segments may be summed to build a portfolio-level count model. The mathematical convenience of adding λ values allows fast planning, capacity estimation, and risk assessment.

Common mistakes students and analysts make

• Adding probabilities instead of rates. You add λ parameters, not separate point probabilities.
• Ignoring independence. Dependence can break the Poisson closure property.
• Mixing time units. If one λ is per hour and another is per day, convert them to the same unit first.
• Confusing mean and standard deviation. For Poisson, the mean and variance are equal, but the standard deviation is the square root of λ.
• Forgetting that k must be a non-negative integer. Poisson probabilities are defined only for count outcomes.

Deriving the probability of the sum directly

For two variables, you can also derive the formula by convolution. Let X ~ Poisson(λ₁) and Y ~ Poisson(λ₂) be independent. Then for a non-negative integer k:

P(X + Y = k) = Σ P(X = i)P(Y = k – i), where the sum runs from i = 0 to k.

Substitute the Poisson probabilities and simplify:

P(X + Y = k) = Σ [e^-λ1 λ₁ⁱ / i!] [e^-λ2 λ₂^k-i / (k-i)!]

= e^-(λ1+λ2) / k! Σ [k! / (i!(k-i)!)] λ₁ⁱ λ₂^k-i.

The summation is a binomial expansion of (λ₁ + λ₂)^k. That gives:

P(X + Y = k) = e^-(λ1+λ2) (λ₁ + λ₂)^k / k!,

which is exactly the pmf of Poisson(λ₁ + λ₂).

Authoritative references

For formal definitions, derivations, and applied context, review these high-quality academic and government resources:

• NIST Engineering Statistics Handbook: Poisson Distribution
• Penn State University STAT 414: The Poisson Distribution
• University of California, Berkeley: Poisson Processes and Distribution Notes

Bottom line

The formula used to calculate the sum of Poisson random variables is straightforward but powerful: add the λ values, keep the Poisson family, and then compute probabilities from the new total rate. In symbols, if the variables are independent and Poisson distributed, then S = ΣX_i is Poisson with parameter Σλ_i. That one line lets you move from multiple count processes to a single exact model for the total. If you are building dashboards, operational forecasts, reliability studies, or classroom solutions, this rule is one of the fastest ways to simplify a count problem without losing mathematical correctness.

Educational note: this calculator assumes independence and a standard Poisson model. For dependent counts or overdispersed data, consider more advanced models such as mixed Poisson, negative binomial, or hierarchical count models.