Birthday problem formula is wrong?

[CyberShot]

The birthday problem solution on Wikipedia seems to beP = Prob of having at least 2 people having the same birthday, given n

d = days in a year

n = # of people in room P = d! / (d^n * (d - n)!) To make calculations easy, let's assume there are 3 days in a year, thus 3 possible birthdays and 3 people in a room.

According to the formula, the prob of a match should be = 3! / (3^3 * 0!)

P = 6 / 27, or roughly 22.2% of the time there should be a match.

Does anyone else think this sounds way too low?

---

Now, I did it the long way, and I came up with what I think should be the answer

P = 21 / 27 or roughly 77.8 % of the time

I've also run a computer simulation to corroborate my analytical findings, and, in a room of 3 people, there was a match of at least 2 people between 77 - 82 % of the time out of a 100 runs.

 

[Filip Larsen]

In your expression for p, p means the probability that no one has the same birthday.

 

[uart]

The wikipedia article says :

\bar{p} = \frac{d!}{d^n \, (d-n)!}

The "bar" means it's the complementary probability. So p = 1 - \bar{p} and the correct probability for the question you posed is 1 - 6/27 = 21/27.