2
$\begingroup$

Two people arrive at a train station at a random time between 12pm and 1pm. They arrive independently of each other and their arrival times are uniformly distributed. What is the probability they arrive at exactly the same time? I am trying to use a graphical approach, and just have straight line (y=x). Wouldn't this yield a probability of zero, since the area of a line is 0?

$\endgroup$
  • 1
    $\begingroup$ yes, you're right. $\endgroup$ – Neil G Feb 11 '16 at 4:58
2
$\begingroup$

Let's start with simple case with discrete time. If you have two discrete random variables $X$ and $Y$ that that are iid and follow discrete uniform distributions with $k$ possible categories (arrival times from $0$ to $k-1$), then probability of $X=x$ is $1/k$ and the same for $Y$. Since $X$ and $Y$ are independent, then $\Pr(X=t \cap Y=t) = \Pr(X=t)\times\Pr(Y=t) = 1/k^2$.

With continuous variables things get more complicated (and I assume that if you are talking about time you think of continuous time). For continuous random variable $X$ probability of $X=x$ for any $x$ is equal to zero (that is why we use pdf's instead of pmf's). So for continuous line from $a$ to $b$ point $x$ occupies infinitesimal space. If you align two such lines $X$, $Y$ than the probability of $X=t$ and $Y=t$ occurring simultaneously goes to zero as probability of each of the events alone is zero. With a single variable we were talking about a line and with two variables we can think of it as of square, with $\Pr(X=x \cap Y=y)$ being a single point inside such square.

What you could do is to compute probability of two arrivals occurring within some time interval (say $\pm \varepsilon$), by using cumulative distribution functions

$$ \Pr(t-\varepsilon \le X \le t+\varepsilon \cap t-\varepsilon \le Y \le t+\varepsilon) \\ = \left( F_X(t+\varepsilon) - F_X(t-\varepsilon) \right) \times \left( F_Y(t+\varepsilon) - F_Y(t-\varepsilon) \right)$$

What translates to talking about area of smaller square within bigger square.

enter image description here

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.