# How to calculate the probability

I have $N>0$ coins that are thrown into $M>0$ boxes. Each box has the same chance to receive each coin. (i.e. the probability of a coin to land in one particular box is $1/M$)

How can I calculate the probability $p$, that each box contains at least one coin?

With $M > N$, $p = 0$
With $M = N$ $p$ would be $M!/M^N$
But with $N > M$ ?

Please give me some clues to solve this problem! Is it easier to just find $N = f(M)$ in order to have $p>90\%$?

-
Why $\displaystyle p=\frac{N!}{3^n}$ if M=N and what is $n$ in this? – Tengu Dec 4 '12 at 20:50
For $M = N$, I think it should be $M! / M^N$. – StuartHa Dec 4 '12 at 21:00
@Stuart I think you are right. At least if $\displaystyle \frac{N!}{3^N}$, the probability could be greater than 1. – Tengu Dec 4 '12 at 21:06
Oups. Of course it is M!/M^N for M=N. Thanks! – Jenny Dec 4 '12 at 21:21
See duplicate: math.stackexchange.com/questions/26528/… – StuartHa Dec 5 '12 at 7:12

First we fix $N=n>0$ coins, and let $Q(m)$ be the probability that at least one of the $m$ boxes does not have any coins. Then $p=1-q$. $Q(1)=0$

If $m=2$, $Q(2)=\displaystyle \frac 2 {2^n}$.

If $m=3$, $Q(3)\displaystyle =\frac{\begin{pmatrix}3\\2\end{pmatrix}(1-Q(1))1^n+ \begin{pmatrix}3\\1\end{pmatrix} (1-Q(2))2^n}{3^n}$.

If $m=4$, $Q(4)=\displaystyle\frac {\begin{pmatrix}4\\3\end{pmatrix}(1-Q(1))1^n+\begin{pmatrix}4\\2\end{pmatrix}(1-Q(2))2^n+\begin{pmatrix}4\\1\end{pmatrix}(1-Q(3))3^n}{4^n}$

If $m=m_0$,$Q(m_0)=\displaystyle \frac{\sum_{k=1}^{k=m_0-1}\begin{pmatrix}m_0\\m_0-k\end{pmatrix}(1-Q(k))k^n}{m_0^n}$ for $m_0\geq 3$.

I considered the number of boxes that could be vacant.

-
Thanks. That helped. – Jenny Dec 6 '12 at 11:55