"Mensanator" <email@example.com> wrote > > I am not disputing that the probability changes at each i. What I am > disputing is whether that is the answer to the stated question. > If you performed the inspection on 10,000 buckets of parts (each with > exactly 1 defective part) you will find the defective part on the 13th > inspection approximately 100 times. You will find the defective part > on the 37th inspection approximately 100 times. You will find the > defective part on the 63rd inspection approximately 100 times. In fact, > for all values of i you will find the defective part approximately 100 times.
> How do you reconcile this with the fact that at inspection 13 the probability > of success is 1/88 whereas at inspection 37 the probability of success is > 1/64?
Nothing to reconcile. different questions have different answers.
> Assuming he meant "determine the probabilty before beginning the inspections" > then he is absolutely correct. The probability is 1/100 for all i.