From Binomial to Poisson: A Beautiful Connection in Probability

Introduction

In probability theory, two of the most important distributions are the Binomial and Poisson distributions. At first glance, they seem unrelated:

• The Binomial distribution models the number of successes in a fixed number of trials (e.g., coin flips).
• The Poisson distribution models rare events occurring over time or space (e.g., meteor sightings per night).

But surprisingly, the Poisson distribution can be derived from the Binomial distribution under certain conditions! Let’s explore this fascinating connection.

---

1. The Binomial Distribution

The Binomial distribution describes the probability of getting exactly \( k \) successes in \( n \) independent trials, where each trial has success probability \( p \).

The probability mass function (PMF) is:

\(
P(X = k) = {n \choose k} p^k (1-p)^{n-k}
\)

Example:
If you flip a fair coin (\( p = 0.5 \)) 10 times (\( n = 10 \)), the probability of getting exactly 3 heads (\( k = 3 \)) is:

\(
P(X = 3) = {10 \choose 3} (0.5)^3 (0.5)^7 \approx 0.117
\)

This works well when \( n \) is small and \( p \) is not extremely tiny.

---

2. The Poisson Distribution

The Poisson distribution models the number of rare events occurring at a constant rate \( \lambda \) over a fixed interval.

Its PMF is:

\(
P(X = k) = \frac{\lambda^k e^{-\lambda}}{k!}
\)

Example:
If a call center receives an average of \( \lambda = 5 \) calls per hour, the probability of getting exactly 2 calls in the next hour is:

\(
P(X = 2) = \frac{5^2 e^{-5}}{2!} \approx 0.084
\)

This is useful when events are rare but can happen many times independently.

---

3. The Connection: Binomial → Poisson

Now, here’s the magical part: The Poisson distribution is a limiting case of the Binomial distribution when:

1. \( n \to \infty \) (number of trials becomes very large),
2. \( p \to 0 \) (success probability becomes very small),
3. But \( np = \lambda \) (a constant average rate).

Derivation (Simplified)

Start with the Binomial PMF:

\(
P(X = k) = {n \choose k} p^k (1-p)^{n-k}
\)

Since \( p = \frac{\lambda}{n} \) (because \( np = \lambda \)), substitute:

\(
P(X = k) = {n \choose k} \left( \frac{\lambda}{n} \right)^k \left( 1 – \frac{\lambda}{n} \right)^{n-k}
\)

Now, take the limit as \( n \to \infty \).

• Step 1: The binomial coefficient \( {n \choose k} \) behaves like \( \frac{n^k}{k!} \) for large \( n \).
• Step 2: \( \left( 1 – \frac{\lambda}{n} \right)^{n} \to e^{-\lambda} \) (a well-known exponential limit).
• Step 3: \( \left( 1 – \frac{\lambda}{n} \right)^{-k} \to 1 \) (since \( k \) is fixed).

Putting it all together:

\(
P(X = k) \approx \frac{n^k}{k!} \cdot \frac{\lambda^k}{n^k} \cdot e^{-\lambda} = \frac{\lambda^k e^{-\lambda}}{k!}
\)

Which is exactly the Poisson PMF!

---

4. When Does This Approximation Work?

The Poisson approximation is good when:

• \( n \geq 20 \) (large number of trials),
• \( p \leq 0.05 \) (small success probability),
• \( np \) is moderate (e.g., between 0.1 and 10).

Example:
A factory produces 10,000 light bulbs, each with a 0.1% defect rate (\( p = 0.001 \)).

• Exact Binomial Probability (5 defects):
  \(
  {10000 \choose 5} (0.001)^5 (0.999)^{9995} \quad (\text{hard to compute!})
  \)
• Poisson Approximation (\( \lambda = np = 10 )\):
  \(
  \frac{10^5 e^{-10}}{5!} \approx 0.0378
  \)

The Poisson version is much easier to calculate and is very close to the true Binomial probability.

---

5. Real-World Applications

This connection is not just theoretical—it’s widely used in:

• Telecommunications (modeling call arrivals),
• Biology (predicting mutations in DNA),
• Finance (estimating rare stock market crashes),
• Medicine (studying disease outbreaks).

Whenever you have many trials with tiny success probabilities, the Poisson distribution simplifies calculations!

---

Conclusion

What initially seems like two completely different distributions—Binomial (discrete trials) and Poisson (rare events)—are actually deeply connected. The Poisson distribution emerges naturally from the Binomial distribution when dealing with large numbers of rare events.

This elegant relationship shows how mathematics often unifies seemingly unrelated concepts into a beautiful framework. Next time you encounter a Poisson problem, remember: it’s just a Binomial distribution in disguise!