The problem statement and solution is below. I understand that each die toss is independent but it doesn't make sense to me why the probability is multiplied by the number of sequences, as in the solution.
When you roll a die 18 times, it is possible to get a countable number of sequences such that a 1,2,3,4,5,6 each appear 3 times. Wouldn't the probability that each number appears exactly 3 times be equal to the ratio of the number of sequences in which 1,2,3,4,5,6 appear 3 times to the total number of sequences you can have (6^18)?

Yes, and the number of sequences is exactly what's calculated with the multinomial coefficient. You choose the positions in the sequence for the $1$'s, then for the $2$'s, and so forth. The collection of these sequences is counted by the multinomial coefficient.