Meta
What are readability calculators basing their levels on? Is there any empirical data about what features of text make it more or less easy to read? Word choice is one thing, but structure and other less obvious things probably also play a part.
Most readability metrics use some combination of the average number of words per sentence and the average number of syllables per word (the latter being a crude metric of word rarity), though some compare against a whitelist of common words similar to (but usually larger than) the one XKCD's using. If people feel a need to use enough circumlocution to make up for removing rare words, the gains in the latter might not outweigh the losses in the former.
The details vary quite a bit by formula. I tried several readability analyses on portions of this post ...
If it's worth saying, but not worth its own post, even in Discussion, it goes here.