I'm still trying to work out how to answer this question.

If Group $A$ has $\frac{1}{100}$ items defective, and Group $B$ has $\frac{1}{200}$ items defective. Overall probability of item being defective is therefore $\frac{3}{200} = 0.015$.

If the first item chosen works (is not defective) what is the probability it came from group $A$?

My thinking on this was just $1 - \frac{1}{100}$. Is that correct?

Secondly, what is the probability the second item works given the first one also works. Should I be applying conditional probability here?

  • $\begingroup$ How did you get $3/200$ as overall probability of being defective? If it is by adding the probabilities $1/100$ and $1/200$ then you are wrong. $\endgroup$ – drhab Mar 27 '18 at 11:39
  • $\begingroup$ @drhab I should have multiplied them instead? $\endgroup$ – kitkat1224 Mar 27 '18 at 12:05
  • $\begingroup$ No. If $p$ is the probability that the item comes from group $A$ and $1-p$ is the probability that the item comes from group $B$ then the probability of being defective is $p\cdot1/100+(1-p)\cdot1/200$ $\endgroup$ – drhab Mar 27 '18 at 12:16
  • $\begingroup$ so (p *1/100) + (1-p) * (1/200) ? $\endgroup$ – kitkat1224 Mar 27 '18 at 12:30
  • $\begingroup$ If "random" stands for fifty-fifty where it concerns the groups then the probability of being defective is $\frac12\frac1{100}+\frac12\frac1{200}$ $\endgroup$ – drhab Mar 27 '18 at 12:32

Assuming both groups are equally large, the a priori probabilities are $P(A) = P(B) = \frac{1}{2}$ .Given is (the event $d$ is for defective) $P(d|A) = \frac{1}{100}$ and so $P(\lnot d| A) = \frac{99}{100}$.

So standard Bayesian rules tell us:

$$P(A| \lnot d) = \frac{P(\lnot d|A)P(A)}{P(\lnot d|A)P(A) + P(\lnot d|B)P(B)}$$

which equals $\frac{2\cdot 99}{2\cdot 99+199}$ after simplification (a little less than a half).

  • $\begingroup$ How would I go about calculating the probability of the second item checked being good, given that the first one was? (I am assuming conditional probability :) $\endgroup$ – kitkat1224 Mar 27 '18 at 13:04
  • $\begingroup$ @kitkat1224 further conditioning. You know $P(A|\lnot d)$ and $P(B |\lnot d)$. Now condition P(\lnot d_2)$ in $A$ etc. $\endgroup$ – Henno Brandsma Mar 27 '18 at 13:11
  • $\begingroup$ @kitkat1224 so $P(\lnot d_2)= P(\lnot d_2|A)P(A|\lnot d_1) + P(\lnot d_2|B)P(B| \lnot d_1)$ $\endgroup$ – Henno Brandsma Mar 27 '18 at 13:14
  • $\begingroup$ so that's just giving me the probability that the second watch is not defective, and then I condition it? :) edit. i.e. which rules of probability as I using to work this out? $\endgroup$ – kitkat1224 Mar 27 '18 at 13:20
  • $\begingroup$ Can also approach this problem from the perspective of Probability(two good pieces) / P(first good part) = P(2nd good part | 1st part good) $\endgroup$ – kitkat1224 Mar 28 '18 at 3:36

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.