The average of a set is its total divided by the number of items. Almost every exam question is really about the total: if the average of 5 numbers is 20, the total is 100, and that total is what you manipulate.
When one value is replaced, the change in average multiplied by the count gives the change in total. If the average of 10 students' marks rises by 2 after correcting one score, that score was understated by 20.
For consecutive numbers the average is simply the middle term, so the average of the first n natural numbers is (n + 1) / 2.