# A random variable that is a mix of discrete and continuous

Here's a question we got for homework:

It is given that at a certain bank there's 50-50 chance that when you
enter there's:

- no one waiting in line
- there's one man waiting, in which case the waiting time is
exponentially distributed.

What is the CDF of the total waiting tine?

Instruction:
Let X be the total waiting time, Y the number of people waiting.
For x>=0, use the total probability theorem for the CDF of X
Notice that X is not discrete nor continuous, but a mix of both.


Here's what I thought. As specified, Y can be either 0 or 1 people waiting. If there's no one waiting the waiting time is 0 which means P(X<=x) = 1 for all x>=0. If there's one man waiting then P(X<=x) = 1 − e^(−λx).

So, by the law of total probability,

P(X<=x) = P(X<=x|Y=1)P(Y=1) + P(X<=x|Y=0)P(Y=0)

Am I right so far? If I am right, then what is P(X<=x|Y=1)? Are the two variables independent?

Thanks!

-
You have said what it is! Recall that you wrote that if there's one man waiting, then $P(X \le x)=1-e^{-\lambda x}$. – André Nicolas Dec 17 '11 at 13:14

As you explained yourself, $P(X\leq x\mid Y=1)$ is the distribution function of the exponential distribution, so $$P(X\leq x\mid Y=1)=1 − e^{−\lambda x}.$$
$P(X\leq x\mid Y=0)$ looks differently: Since there is no line, the waiting time is zero and therefore $$P(X\leq x\mid Y=0)=1_{[0,\infty)}(x).$$
The two variables are certainly not independent. Otherwise, we would necessarily have $$P(X\leq x\mid Y=1)=P(X\leq x\mid Y=0),$$ which as we have seen is not the case.