Birthday Problem: Probability that at least two also share the same weekday, if weekday uniformly distributed and independent of birth date?

144 Views Asked by At

Consider the following problem extension for the birthday problem: We now want know the probability that out of $n$ persons, at least two people were born on the same date and the same weekday (for example, both were born on the 02.01 and on a Tuesday). We also want to know for which $n$ that probability is great than $\frac{1}{2}$.

We can assume that birthdays and weekdays follow a uniform distribution (also, we can assume the absence of leap-years) and are independent events. That is, the probability to be born on a certain date is $\frac{1}{365}$ and to be born on a weekday is $\frac{1}{7}$.

1

There are 1 best solutions below

3
On BEST ANSWER

You are looking for $$1 - \dfrac{(2555)_n}{2555^n} \ge \dfrac{1}{2}$$

where $$(2555)_n = (2555)(2554)\cdots (2555-n+1)$$

Plugging this into Wolframalpha gives $n \ge 59.7852$.

Testing this out:

$$1 - \dfrac{(2555)_{60}}{2555^{60}} \approx 0.50252$$

$$1- \dfrac{(2555)_{59}}{2555^{59}} \approx 0.49076$$

If the notation is confusing, another notation would be, the probability that every person has a distinct birthday and birth weekday is:

$$\left(\dfrac{2555}{2555}\right)\left(\dfrac{2554}{2555}\right)\cdots \left(\dfrac{2555-n+1}{2555}\right)$$

So, the probability that at least two people share both a birthday and a birth weekday is the complement of that:

$$1-\left(\dfrac{2555}{2555}\right)\left(\dfrac{2554}{2555}\right)\cdots \left(\dfrac{2555-n+1}{2555}\right) = 1-\dfrac{(2555)_n}{2555^n}$$

In Excel, you can use the following formulas:

$$\begin{array}{rl}\text{Cell A1:} & \text{Number of people} \\ \text{Cell B1:} & \text{Prob All Distinct} \\ \text{Cell C1:} & \text{Prob at least two share}\end{array}$$

This gives you column headers. Next, we want calculations:

$$\begin{array}{rl}\text{Cell A2:} & \text{=ROW()} \\ \text{Cell B2:} & \text{=(2555-ROW()+1)/2555} \\ \text{Cell C2:} & \text{=1-B2} \\ \text{Cell A3:} & \text{=ROW()} \\ \text{Cell B3:} & \text{=B2*(2555-ROW()+1)/2555} \\ \text{Cell C3:} & \text{=1-B3}\end{array}$$

Copy the formulas in cells $\text{A3:C3}$ to $\text{A3:A100}$. You will see that at 60 people, the probability in column $C$ is greater than $0.5$ for the first time.