Percentiles represent actual values, not ranges. Just because you receive a letter in the mail that says you scored in the 100th percentile doesn't make it correct.
It is also important to remember that percentiles are not meant to be used in data groups less than 100, they are a way of efficiently looking at LARGE groups of data.
Take an example of a test where 101 different students score 0, 1, 2, 3, 4, 5, 6, 7, 8, 9... all the way to 98, 99, and 100 points respectively.
The student who scored the 0 scored higher than 0% of the group, he is at the very bottom, the 0th percentile. Students below him = 0, total students = 101. 0/101 = 0%.
The student who scored the 1 scored higher than 0.99% of the group, he is in the 9.09th percentile. Students below him = 1, total students = 11. 1/101 = 0.99%
The student who scored the 2 scored higher than 1.98% of the group, he is in the 18.18th percentile. Students below him = 2, total students = 11. 2/101 = 1.98%
This continues on, let's skip to the student who scored the 99. Students below him = 99, total students = 101. 99/101 = 98.02% He is in the 98.02nd percentile.
Here comes the fun one, the student who scored the 100. Students below him = 100, total students = 101. 100/101 = 99.009%. He is in the 99.009th percentile.
By increasing the population of data, you can increase the denominator to very large numbers, but the highest the numerator can EVER reach is the denominator -1, because a value can never be higher than itself. You can have someone score higher than 9,999,999 other people, but because they are taking the exam as well, the total number of scores is out of 10,000,000. 9,999,999/10,000,000 = 99.99999% but it can NEVER reach 100% There is no such thing as the 100th percentile.