1
$\begingroup$

1) Biased coin flipped twice. There is 6/9 probability that the coin will land heads for at least one of the two flips. What is the smallest number of times that I can flip the coin for at least 99% certainty that it will land heads at least once?

My answer: $n\left(\dfrac{6}{9}\right)^2 = .99 \Rightarrow n = 2.227$ so at least 3 flips

2) Biased coin lands head w/ probability p and start with empty urn. Flip coin n times. each time coin lands head, blue ball into urn. Each time it lands tail, green ball in urn. After finishing flipping the coin, draw k balls from urn one at a time. Replacing each ball you draw before drawing another. If all the k balls you drew are blue, what is the probability that all n balls in the urn are blue?

My answer: I am assuming it is in .99 to 1 range, because since all the balls drawn are blue, it is hard to know the probability for the green ball.

$\endgroup$
  • $\begingroup$ Do not attempt to deface your own question. As you can see, it's pointless anyway, because the original content can always be recovered. $\endgroup$ – heropup Jun 25 '14 at 18:03
0
$\begingroup$

Question 1. If $X$ is a random variable that counts the number of heads obtained in $n = 2$ coin flips, then we are given $\Pr[X \ge 1] = 2/3$, or equivalently, $\Pr[X = 0] = 1/3 = (1-p)^2$, where $p$ is the individual probability of observing heads for a single coin flip. Therefore, $p = 1 - 1/\sqrt{3}$.

Next, let $N$ be a random variable that represents the number of coin flips needed to observe the first head; thus $N \sim {\rm Geometric}(p)$, and we need to find the smallest positive integer $k$ such that $\Pr[N \le k] \ge 0.99$. Since $\Pr[N = k] = p(1-p)^{k-1}$, I leave the remainder of the solution to you as an exercise; suffice it to say, you will definitely need more than 3 coin flips.

Question 2. Your answer obviously must be a function of $p$, $n$, and $k$. It is not possible to give a numeric answer. Clearly, $X \sim {\rm Binomial}(n,p)$ represents the number of blue balls in the urn, and $n-X$ the number of green balls.

Next, let $Y$ be the number of blue balls drawn from the urn out of $k$ trials with replacement. Then $Y \mid X \sim {\rm Binomial}(k,X/n)$. You want to determine $$\Pr[X = n \mid Y = k] = \frac{\Pr[Y = k \mid X = n]\Pr[X = n]}{\Pr[Y = k]}.$$ It is trivial to see that $\Pr[Y = k \mid X = n] = 1$, and $\Pr[X = n] = p^n$. The denominator is the tricky part: You need to write $$\Pr[Y = k] = \sum_{x=0}^n \Pr[Y = k \mid X = x]\Pr[X = x],$$ by the law of total probability. Again, I leave the remainder of the solution for you to work out.

$\endgroup$
  • $\begingroup$ I am getting k to be a less than 3. 2/sqrt(3)*(1-1/sqrt(3))^(k-1)=.99 -> k = 1.178. What am I doing wrong? and thank you so much for your help $\endgroup$ – myth4ks Jun 20 '14 at 7:29
  • $\begingroup$ $$\Pr[N \le k] = \sum_{j=1}^k \Pr[N = j] = 1 - (1-p)^k.$$ Now substitute and solve. $\endgroup$ – heropup Jun 20 '14 at 7:36
  • $\begingroup$ I am sorry, but can you explain Pr[Y=k∣X=n]=1 and the denominator Pr[Y=k] on how you got those $\endgroup$ – myth4ks Jun 20 '14 at 8:01
  • $\begingroup$ $X$ counts the number of blue balls in the urn. $Y$ counts the number of blue balls drawn from the urn in $k$ trials. So, given that all the balls in the urn are blue, what is the probability that you will always draw a blue ball? $\endgroup$ – heropup Jun 20 '14 at 8:03
  • $\begingroup$ The probability is 1, since it only blue balls were only drew. $\endgroup$ – myth4ks Jun 20 '14 at 8:18

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.