Bayes Theorem where the number of tosses depends on tosses

105 Views Asked by At

There are two fair coins (i.e. Heads and Tails are equally likely for tosses of both). Coin $1$ is tossed $3$ times. Let $X$ be the number of Heads that occur. After this, Coin $2$ is tossed $X$ times. Let $Y$ be the number of Heads we get with Coin $2$. The probability $Pr(X ≥ 2|Y = 1)$ equals ?

I tried to solve it using Bayes' theorem such that

$Pr(X\geq 2|Y=1)={Pr(X\geq 2)\cdot[Pr(Y=1|X=2)+Pr(Y=1|X=3)]\over Pr(X>=2)\cdot[Pr(Y=1|X=2)+Pr(Y=1|X=3)]+Pr(X<2)\cdot[Pr(Y=1|X=0)+Pr(Y=1|X=1)]} $

and i got the answer $7/11$ which seems to be wrong. Can anyone help me understand why this method fails to work?

1

There are 1 best solutions below

0
On BEST ANSWER

$$P(X\geq 2 | Y=1) = \frac{P(X \geq 2 \text{ and } Y=1)}{P(Y=1)}$$

We can solve for the numerator and the denominator separately.

$X$ can be $0,1,2,$ or $3$. $Y$ can equal $1$ in $3$ different ways, because if $X=0$ then $Y$ cannot be $1$. Each of those is a product, so the denominator is a sum of three terms.

The numerator is easier. We can split that up into

$$P(X \geq 2 \text{ and } Y = 1) = P(Y = 1 | X \geq 2) \cdot P(X \geq 2)$$

The numerator is two cases, if $X=2$ and if $X = 3$, so the numerator will be a sum of two products:

$$P(Y=1|X=2)\cdot P(X=2) + P(Y=1|X=3)\cdot P(X=3)$$

From there you can put the answer together. It's hard to read your equation, but I think your problem is that you were trying to factor out a term where there is no common factor. You have to compute each specific product separately, then add.


I tried to solve it using Bayes' theorem such that

$Pr(X\geq 2|Y=1)={Pr(X\geq 2)\cdot[Pr(Y=1|X=2)+Pr(Y=1|X=3)]\over Pr(X>=2)\cdot[Pr(Y=1|X=2)+Pr(Y=1|X=3)]+Pr(X<2)\cdot[Pr(Y=1|X=0)+Pr(Y=1|X=1)]} $

Problem: $\Pr(Y=1\mid X\geq 2)\neq \Pr(Y=1\mid X=2)+\Pr(Y=1\mid X=3)$ as the Additive Law applies to unions of disjoint events, not conditions.

Solution: Deal with the union before the conditioning.

In summary:

$\begin{align}\Pr(X\geq 2\mid Y=1)&=\dfrac{\Pr(X=2,Y=1)+\Pr(X=3,Y=1)}{\Pr(X=1,Y=1)+\Pr(X=2,Y=1)+\Pr(X=3,Y=1)}\\[1ex]&=\tfrac{\Pr(X=2)\Pr(Y=1\mid X=2)+\Pr(X=3)\Pr(Y=1\mid X=3)}{\Pr(X=1)\Pr(Y=1\mid X=1)+\Pr(X=2)\Pr(Y=1\mid X=2)+\Pr(X=3)\Pr(Y=1\mid X=3)}\\[1ex]&=\dfrac{5}{9}\end{align}$