How Many Distinct Birthdays in a Room of k People?

  • Thread starter Thread starter mXSCNT
  • Start date Start date
AI Thread Summary
The discussion centers on calculating the expected number of distinct birthdays in a room of k people on a planet with n days per year. The formula presented for the expected number of distinct birthdays is j = n - n * (1-1/n)^k, which derives from the probability that a given day has no birthdays. The conversation also touches on combinatorial aspects, suggesting that the total number of distinct birthdays can be represented using combinations, specifically nCk. There is some confusion regarding the definition of distinct birthdays, with a clarification needed on whether it refers to unique birthdays or simply the presence of birthdays on specific days. The final conclusion emphasizes the need for precise definitions in calculating distinct birthdays accurately.
mXSCNT
Messages
310
Reaction score
1
This is a restatement of the vocabulary problem which I introduced in https://www.physicsforums.com/showthread.php?t=293553. Perhaps these terms will be more familiar/less ambiguous.

Suppose we are on a planet where each year has n days, and in a room with k people. If birthdays are uniformly distributed throughout the year, how many distinct birthdays, j, do we expect to find in the room? (Alternatively, how many birthdays n-j are NOT represented in the room, on average?)
 
Physics news on Phys.org
Here's the answer (someone else's idea): j = n - n * (1-1/n)^k. The probability that a given day is nobody's birthday is (1-1/n)^k, so the expected number of days that are nobody's birthday is n * (1-1/n)^k.
 
I think that, since its distinct birthdays (from what I understand, distinct bdays are only the ones that don't coincide on the same day), you will have nCk bdays total, thus the expected number on any given day would be (nCk)/n since they're uniformly distributed.

There will also be nC(n-k) bdays that don't happen, once again the expected number would be nC(n-k)/n since they're uniform. I'm not sure about this, but I think its intuitive.
 
Let X be the number of distinct birthdays, and for 1 \leq i \leq n define Xi to be 1 if there's at least one person whose birthday is on the ith day, and 0 otherwise. Then X = X1 + X2 + ... + Xn, so:

E(X) = \sum _{i=1} ^n E(X_i) = nE(X_1)

E(X1)
= Prob(at least one person has their birthday on day 1)
= 1 - Prob(no one has their birthday on day 1)
= 1 - (# of ways to arrange k birthdays amongst n-1 days, allowing repetition)/(# of ways to arrange k birthdays amongst n days, allowing repetition)
= 1 - \binom{n+k-2}{k} / \binom{n+k-1}{k}

So the final answer is:

\frac{nk}{n+k-1}
 
AKG said:
define Xi to be 1 if there's at least one person whose birthday is on the ith day, and 0 otherwise.
But doesn't the question as for distinct birthdays, ie Xi would be 1 if there is only one birthday on day i and 0 if there is not only one birthday on day i?
 
Back
Top