Mean and variance for the raincoat problem

3.3k Views Asked by At

Suppose $n$ people each have a hat which they check at a hat room. Suppose that hats are returned randomly. Let $Y$ be the number of people who get their own hats. Find $E(Y)$ and $Var(Y)$.

After few hours, I use some formula in combinatorics and generalize the event as

$f(y)$=(Number of ways of $y$ people get their hat)*(Number of ways of $n-y$ people derangement their hats)/(Total Possibilities)

$$\therefore f(y)=\frac{{n\choose y}*(n-y)!\bigg(1-1+\frac{1}{2!}+...+(-1)^y\frac{1}{(n-y)!}\bigg)}{n!}$$

Simplify $f(y)$, $$f(y)=\frac{1}{y!}\bigg(1-1+\frac{1}{2!}+...+(-1)^y\frac{1}{(n-y)!}\bigg)$$ where $0\leq y\leq n$

Then I applied this information which is $$\sum^n_{y=0}P(Y=y)=1$$

After 3 pages of working, then I get answer of page1 page2 page3 $$E(Y)=Var(Y)=1$$

Although I solved this problem, but new problems arise from my mind:

  1. Is there other ways to get the answers without have the knowledge of derangement formula?
  2. Any shorter and simpler ways to answer this question? (I feel my way of solving quite dumb)
  3. And how to show that $\sum^n_{y=0}P(Y=y)=1$ according to my $f(y)$?
1

There are 1 best solutions below

2
On BEST ANSWER

$Y = \sum_{i=1}^n Z_i$ where $Z_i = 1_{\{\text{i-th person went home with correct coat}\}}$

$E(Y) = \sum_{i=1}^n E(Z_i) = n\times 1/n=1$

For $Var(Y)$, note

$E (Y^2) = \sum_{i=1,j=1} E(Z_iZ_j) $.

Note that $E(Z_i^2)=\frac{1}{n}$, and $E(Z_iZ_j)= \frac{1}{n(n-1)}$ for $i,j$ distinct, then use

$Var(Y) = E(Y^2)-E(Y)^2$.

Note that the only thing I used here is linearity of expectation. I am sure this is the simplest way of doing this question (and probably the intended way for a probability course).