There is a subject—naturally not math for computer science—in which 10% of the assigned problems contain errors. if you ask a teaching assistant (ta) whether a problem has an error, then they will answer correctly 80% of the time, regardless of whether or not a problem has an error. if you ask a lecturer, he will identify whether or not there is an error with only 75% accuracy.

we formulate this as an experiment of choosing one problem randomly and ask-ing a particular ta and lecturer about it. define the following events: