I was working on this problem and I can solve it by applying the same technique as suggested in this answer: Inclusion-Exclusion Problem?
"A widget inspector inspects 12 widgets and finds that exactly 3 are defective. The widgets then get all mixed up and the inspector has to find the 3 defective widgets again by testing one.Find the probability that the inspector will now have to test at least 9 widgets"
However, for my own sake I'm trying to also get the same answer by counting the ways the widget inspector doesn't find the defective widgets amongst the first 8 widgets using inclusion-exlusion, however, my answers don't match up so I was hoping someone could point out what I am doing wrong:
$$ \frac{\binom{12}{1}11!- \binom{12}{2}10!+\binom{12}{3}9!-\binom{12}{4}8!+\binom{12}{5}7!-\binom{12}{6}6!+\binom{12}{7}5!-\binom{12}{8}4!}{12!} $$
Edit (Reasoning for the numerator): I wanted to use inclusion-exclusion to count on the numerator the number of ways that the first 8 widgets were non-defective. I can see now that that's not what I have done at all. My reasoning for the first term in the numerator was that if the first widget was non-defective then the remaining widgets could be chosen in 11! ways and the first widgets could be chosen in 12C1 ways. I feel like this is were I have a gap in my understanding which I am trying to fill and Im struggling to make the proper connections.