0
$\begingroup$

Suppose that a bowl contains $100$ chips: $30$ are labelled $1$, $20$ are labelled $2$, and $50$ are labelled $3$. The chips are thoroughly mixed, a chip is drawn, and the number $X$ on the chip is noted. Also, assume that the first chip is replaced, a second chip is drawn, and the number $Y$ on the chip noted.

Compute $P(W=w)$ for every real number $w$ when $W=X+Y$.

I am really confused as to what $P(W=w)$ means. For example, if we considered $w = 1$, could we write $P(W=1)$ as $P(X = 1) + P(Y = 1)$?

At the back of the textbook where the solutions are written up for that question, it says that $P(W=2) = 0.09$, but doesn't mention anything about $P(W=1)$. Why is that, and how can I go about calculating $P(W=w)$? I know how to calculate $P(X=x)$ and $P(Y=y)$, as we're dealing with one random variable, but when it comes to two, I am just lost.

$\endgroup$
8
  • 1
    $\begingroup$ We have to calculate how many different ways $X$ and $Y$ can make the same $w$,... for example $w=1$ has probability zero as the smallest sum of $X+Y$ is $1+1=2$. For that matter, the overwhelming majority of real numbers have probability zero here: $w$ can only equal integers in the range $[1,6]$ $\endgroup$
    – FShrike
    Commented Jul 21, 2021 at 20:44
  • $\begingroup$ Can you please expand on "We have to calculate how many different ways $X$ and $Y$ can make the same $w$"? Does that imply that $P(W = w) = P(X = w) + P(Y = w)$? $\endgroup$
    – user812082
    Commented Jul 21, 2021 at 20:47
  • 2
    $\begingroup$ I don’t think it does so simply because take for example $w=4$: $X$ can be $1,2,3$ and $Y$ can be $3,2,1$ respectively, making for three pairs. I believe $P(W=4)=P(X=1)\times P(Y=3)+P(X=2)\times P(Y=2)+P(X=3)\times P(Y=1)$, but I am not 100% confident about that for the record $\endgroup$
    – FShrike
    Commented Jul 21, 2021 at 20:50
  • $\begingroup$ I see, thanks. So, as far as I understand, there is not formula for such expressions, and we simply have to carefully analyze the ways we can get $w$? $\endgroup$
    – user812082
    Commented Jul 21, 2021 at 20:59
  • 2
    $\begingroup$ I think so yes: the distribution of a combination of two variables is not necessarily the combination of the distributions. In other words, since the probability distribution of variables can be pretty much anything (with the constraints that it must sum to one and so on) $P$ really isn’t necessarily a linear function $\endgroup$
    – FShrike
    Commented Jul 21, 2021 at 21:05

1 Answer 1

1
$\begingroup$

We want to compute $P(W=w)$ where $W=X+Y$ and $P(W=w)$ is the probability of the event that the random variable $W $ is equal to the value $w$.

Let's first observe that the range of $W $ is the set $\{2,3,4,5,6\}$ since $X,Y$ take values in $\{1,2,3\} $. In particular, $P(W=1)=0 $ as this event cannot occur (this is true for any $w \notin \{2,3,4,5,6\} $).

A key observation to make is that the chips are replaced between drawings, so that each pull is independent of one another, i.e., the chance that I draw chip labeled with a 3 does not have any effect on which chip I draw next. This simplifies calculating the probabilities of each event since we can just multiply the probability of certain draws occurring.

I will calculate a few of these probabilities. For $P(W=2) $ we notice that the event that $W=2 $ can only occur in one way, that is, if $X=1 $ and $Y=1 $. We get

$$P(W=2) = P(X=1)P(Y=1) = \frac{30}{100}\cdot\frac{30}{100}=\frac{9}{100} $$

For $P(W=3) $, there are two ways of this occurring. We see that $ W=3$ if $X=1 $ and $Y=2 $ or if $X=2 $ and $Y=1$. Thus

$$P(W=3) = P(X=1)P(Y=2) + P(X=2)P(Y=1) = \frac{30}{100}\cdot\frac{20}{100} +\frac{20}{100}\cdot\frac{30}{100}=\frac{12}{100}. $$

Lastly, I will calculate $P(W=4) $ which can occur in three distinct ways,as Idiotic Shrike mentioned in the comments. That is, $X=1 $ and $Y=3 $, or $X=2 $ and $Y=2 $, or $X=3 $ and $Y=1 $. Therefore

$$ P(W=4) = P(X=1)P(Y=3) + P(X=2)P(Y=2)+P(X=3)P(Y=1)$$ $$ = \frac{30}{100}\cdot\frac{50}{100} +\frac{20}{100}\cdot\frac{20}{100}+\frac{50}{100}\cdot\frac{30}{100}=\frac{34}{100}$$

I will let you compute the rest.

$\endgroup$

You must log in to answer this question.