0
$\begingroup$

I have some trouble understanding the conditional probability of the following question.. Do pardon me even though it is a simple question..

Two fair dice are rolled and the sum on the faces is calculated. Find the probability : Pr(Sum of two dice is 8 | we know that the number on both dice is even)

This is my working


Let E be the event that the Sum of two dice is 8. $n(E) = {(2,6),(3,5),(4,4),(5,3),(6,2)}, P(E) = 5/36 $

Let F be the event that the number on both dice is even $n(F) = {(2,2),(2,4),(2,6),(4,2),(4,4),(4,6),(6,2),(6,4),(6,6)}, P(F) = 9/36 $

$P(E|F) = (5/36)/(9/36) = 5/9 $


However, in my textbook, while P(F) is the same answer as mine but P(E) is written as 3/36 and hence the overall answer is 1/3 I am unable to get my head around the part for P(E), other than 3/36 is obtained due to ${(2,6),(4,4),(6,2)}$ as these three are the even numbers..

Am I correct to assume as such?

$\endgroup$
  • 1
    $\begingroup$ You enumerated all possible outcomes given both die are even...only three of the $9$, $\{(2, 6), (4, 4), (6, 2)\}$, result in a sum of $8$, hence it's $\frac{3}{9} = \frac{1}{3}$. $\endgroup$ – Jared Apr 20 '16 at 4:11
  • $\begingroup$ @Jared okay.. I think I am having issues interpreting when given a conditional probability. This question is a good example of it though.. Many thanks $\endgroup$ – dissidia Apr 20 '16 at 4:34
2
$\begingroup$

The conditional probability $\mathbb{P}(E|F)$ is not equal to $\frac{\mathbb{P}(E)}{\mathbb{P}(F)}$, but rather to $\frac{\mathbb{P}(E\cap F)}{\mathbb{P}(F)}$.

There are three outcomes in $E\cap F$, namely $(2,6),(4,4)$, and $(6,2)$, hence the conditional probability is $$ \frac{\frac{3}{36}}{\frac{9}{36}}=\frac{1}{3}$$

$\endgroup$
  • $\begingroup$ My bad on the 'signs'. Will take note of that. I think I am still having issues understand when given a conditional prob. So suppose if now the question is stated something like Pr(Sum of two dice is 8 | we know that the number on both dice is <= 10), it will then be (5/36)/(33/36)? $\endgroup$ – dissidia Apr 20 '16 at 4:32
  • $\begingroup$ I'm not sure I understand, aren't the numbers on standard dice always at most 6? $\endgroup$ – carmichael561 Apr 20 '16 at 4:36
  • $\begingroup$ @carmichael561 Unless dissidia meant "... given that the sum is at most 10", which would indeed have a probability of $5/33$. $\endgroup$ – Graham Kemp Apr 20 '16 at 4:40
  • $\begingroup$ Perhaps that is it. $\endgroup$ – carmichael561 Apr 20 '16 at 4:41
  • 1
    $\begingroup$ @dissidia In this case you have $E\subset G$ so $E\cap G=E$ $$\Pr(E\mid G) = \dfrac{\Pr(E\cap G)}{\Pr(G)} = \dfrac{\Pr(E)}{\Pr(G)} = \dfrac{5}{33}$$ $\endgroup$ – Graham Kemp Apr 20 '16 at 4:45
1
$\begingroup$

You wish to know the probability that the sum is eight ($E$) given that the two dice are even ($F$). $\Pr(E\mid F)$

As you stated $E$ is the event set: $\{(2,6),(3,5),(4,4),(5,3),(6,2)\}$ and $F$ is the event set $\{(2,2),(2,4),(2,6),(4,2),(4,4),(4,6),(6,2),(6,4),(6,6)\}$

However, we want only want the conditional probability of $E$ when $F$.

Now the event of both dice being even and the sum equaling eight is: $F\cap E = \{(2,6) (4,4), (6,2)\}$

Since all outcomes of this experiment are equally likely $\Pr(E\mid F) = \dfrac {\lvert F\cap E\rvert}{\lvert F\rvert} = \dfrac{3}{9}$

$\endgroup$

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.