The linked list would be weighted by length, essentially a typed array consisting of tuples of words. If looking for (X = Hell in a set with "Hello" and "Hop" in it, you could make the following deductions:
1. All the characters in X are in the character set. X might exist in the word set.
2. The largest, first word in the list, starting with "H" and skipped to through one or more pointers in the skip list, is 5 characters; too long to be X. It also matches by 4/5 characters, and It is "Hello", however we do not store this information or otherwise compute it due to the first condition.
3. The next word with the best match to the first word, or next shortest size is only 3 characters long. it is not long enough to be X, however there are no more words in the list which are larger or equal to the size of X, and thus X is definitely not in the list.