Only check base strings whose length is evenly divisible by the target
string's length. Other lengths cannot be correct, so we don't need to
check them.

I used LPeg for both of these solutions, which is surely overkill. But I
wanted to experiment with it.

We just need a count of the words, which can be achieved by summing (via
x/reduce +) 1 for every instance we find.

Add comments to clarify what the algorithm is doing at each step.