Calculating the probability of one student scoring the highest in the class on an exam.

One basic problem is that although a student's test average is probably a good indicator of their expected exam score, we need some idea of the random variation that a student may have about that expected value.
Standard Deviation is the average distance away from their expected exam score.