# Probability that $m$ balls will fall into first box

Suppose we have $n$ balls that are randomly distributed into $N$ boxes. Find the probability that $m$ balls will fall into the first box. Assume that all $N^m$ arrangements are equally likely.

### Attempt:

First, we notice that for the first box, we have $n$ choices, and for the second box we also have $n$ choices, and so on. Thus, we have $n^N$ ways to place the balls into the boxes. Pick $m$ balls out of the total $n$ balls, that gives ${n \choose m}$. In how many ways can these $m$ balls go into the first box? Well, in just ${1 \choose 1 }= 1$ ways. Thus,

$$P = \frac{ {n \choose m } }{n^N }$$

Is this correct?

• Yes!, in first like i assume you mean "we notice that for the first ball" – kayush Apr 16 '18 at 1:41

The number of ways to choose the $m$ items in the first box is $\binom{n}{m}$ and the number of ways to put the remaining $n-m$ items into the remaining $N-1$ boxes is $(N-1)^{n-m}$; therefore, the number of ways to put $m$ items into the first box is $$\binom{n}{m}(N-1)^{n-m}$$ So the probability is $$\frac{\binom{n}{m}(N-1)^{n-m}}{N^n}=\frac{\binom{n}{m}}{N^m}\left(1-\frac1N\right)^{n-m}$$ Note that the Binomial Theorem guarantees that \begin{align} \sum_{m=0}^n\binom{n}{m}(N-1)^{n-m} &=((N-1)+1)^n\\ &=N^n \end{align}

• why $N^n$ on the denominator. Isnt it the size of the sample size $n^N$? There are $1,2,3,...,N$ boxes. In the first box we can put $n$ balls, and in the second box we can put also the n balls and the same for the third, fourth, all the way to the $N$ box. So we have $n\times n \times .. \times n = n^N$ – James Apr 16 '18 at 2:41
• To count the number of ways to put $n$ items into $N$ boxes, the first item has the choice of $N$ boxes, the second item has the choice of $N$ boxes, ..., the $n^\text{th}$ item has the choice of $N$ boxes. That is $\underbrace{N\cdot N\cdots N}_{n\text{ items}}=N^n$. – robjohn Apr 16 '18 at 3:02
• I dont understand how this is possible? Why can we not just put all the balls in each box ? Im very confused – James Apr 16 '18 at 3:04
• For each possible arrangement of the balls in the boxes, each ball can only be in one box. That is, ball $1$ is in one of the $N$ boxes; ball $2$ is in one of the $N$ boxes; ... ball $n$ is in one of the $N$ boxes. A box can contain multiple balls; a ball cannot be in multiple boxes. – robjohn Apr 16 '18 at 3:53
• $n=1,N=3:\bbox[5px,border:2px solid black]{o|\phantom{o}|\phantom{o}}\ \bbox[5px,border:2px solid black]{\phantom{o}|o|\phantom{o}}\ \bbox[5px,border:2px solid black]{\phantom{o}|\phantom{o}|o}$. This is $N^n=3^1$ outcomes, not $n^N=1^3$. – robjohn Apr 16 '18 at 10:31

If $X$ is the number of balls that fall into the first box, then $P(X=m)$ follows the binomial distribution, where $p=1/N$.

If we're just concerned with the first box and all other boxes are equally likely, it's just a binomial with $n$ trials, $p=1/N.$Pr(m=k)={n \choose k}\,p^{k}(1-p)^{n-k}={n \choose k}\,N^{-k}(1-1/N)^{n-k}\$