There are two local factories that produce microwaves. Each microwave produced at a factory $A$ is defective with probability $0.05$ where as each microwave at factory $B$ is defective with $0.01$.Suppose you purchase two microwaves that were both produced at the same factory which is equally likely to have been either factory $A$ or $B$. If the first microwave you check is defective, what is the probability that second one will also be defective? I tried the following: Let $X$ be the random variable denoting the number of defective radios.

\begin{align} P(X=x)&=P(X=x|A)P(A)+P(X=x|B)P(B)\\ &=\frac{1}{2}\binom 2k 0.01^{k}(1-0.01)^{2-k}+\frac{1}{2}\binom2k 0.05^k(1-0.05)^{2-k}, \end{align}

and then I compute $\displaystyle \frac{P(X=2)}{1-P(X=0)}$. Is this reasonable?

  • $\begingroup$ Taking a random number of initial words from the problem is a HORRIBLY BAD AND MISLEADING way to construct question titles. Surely you can do better than that? $\endgroup$ – Henning Makholm Dec 8 '15 at 22:29
  • $\begingroup$ Did I answer your question? $\endgroup$ – Carlos Mendoza Dec 10 '15 at 2:28

You are being asked by a conditional probability. Let me first solve the problem using just notation of events. Let's define $D_i$ as the event that the $i$-th checked microwave is defective, and $A$ and $B$ the events associated with the type of factory the microwaves were bought. The desired probability is

\begin{align} P(D_2 \mid D_1) &\stackrel{a}{=} \frac{P(D_1 \cap D_2)}{P(D_1)}\\ &\stackrel{b}{=} \frac{P(A \cap D_1 \cap D_2)+P(B \cap D_1 \cap D_2)}{P(A \cap D_1)+P(B \cap D_1)}\\ &\stackrel{c}{=} \frac{P(A) P(D_1 \mid A) P(D_2 \mid A \cap D_1)+P(B) P(D_1 \mid B) P(D_2 \mid B \cap D_1)}{P(A \cap D_1)+P(B \cap D_1)}\\ &\stackrel{d}{=} \frac{P(A) P(D_1 \mid A) P(D_2 \mid A)+P(B) P(D_1 \mid B) P(D_2 \mid B)}{P(A) P(D_1 \mid A)+P(B) P(D_1 \mid B)}\\ &\stackrel{e}{=} \frac{P(D_1 \mid A) P(D_2 \mid A)+P(D_1 \mid B) P(D_2 \mid B)}{P(D_1 \mid A)+ P(D_1 \mid B)}\\ &\stackrel{}{=} \frac{0.05^2+0.01^2}{0.05+ 0.01}\\ &= \frac{13}{300} \end{align}

(a) Definition of conditional probability. (b)Total probability theorem. (c) Multiplication rule. (d) Given $A$ (or $B$), $D_1$ and $D_2$ are conditionally independent. (e) $P(A) = P(B)$.

I am not sure how the expression

$$\frac{P(X=2)}{1-P(X=0)} = \frac{13}{199}$$

could represent a conditional probability. The use of a random variable like $X$, instead of just events, could lead us to wrong conclusions. For example, $X=1$ when one of the following to possibilities occur: $D_1 \cap D_2^c$ or $D_1^c \cap D_2$. But $D_1^c \cap D_2$ is not possible because we already know that $D_1$ has occured!


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.