When reading a book, you detect each mistake with probability p, independent of other mistakes. Let M denote the amount of mistakes on a certain page and D be the number that you detect on that page. Write down P[D=k|M=m] and find for k>=0 P[D=k]. I've worked out for a textbook with n pages, number of mistakes on each page is poisson RV with parameter a, independent of mistakes on all other pages, that the expected number of pages with no mistakes is 1 - e^-a. I tried using P[D=k|M=m] = P(D=k union M=m)/P(M=m), any ideas how else to go about this?