I'm wondering how the different levels of "difficulty" are calculated and what the cut-off points are. As I understand it, they are not based on how difficult the characters are to write for skritterers, but rather some kind of frequency data. Is this correct?

I'm also wondering if characters and words are treated in the same way and how many that are included in each category, 'easy', 'medium', 'hard', 'rare' and 'super-rare'.

If the levels are based solely upon frequency data, could the rank of individual characters be made more explicit, e.g., by incorporating the rank from [1]? I notice that a lot of the characters in the top 500 of [1] are marked as hard, while some afterwards are marked as easy. So the levels have to be based on some other data – which one?

[1] http://lingua.mtsu.edu/chinese-computing/statistics/char/list.php?Which=MO

We combined standard frequency lists like the one you reference with an analysis of all textbooks in Skritter by level to come up with a combined, learner-weighted frequency metric that works for both characters and words (and now, sort of, sentences). So it's like the frequency list you'd actually want as a learner, where 你好 is more common than 经济. I don't have any readily available stats on how many characters are in each rank; really each has a toughness score ranging from ~4 to 16 and we've imposed labels on certain ranges of that.

The levels are a bit of a misnomer in that they refer to toughness/difficulty when they should probably refer to frequency/importance–we were originally planning to incorporate metrics on how often Skritterers missed each character and word, but it hasn't happened yet.

Hope this answers some of your questions, although I realize it doesn't give you the hard data you're looking for.

Thanks for a thorough answer!

Yea, as for the difficulty compare 孑 (jie2) => Level: hardest and 子 (zi3) => Level: easier. :D
The difficulty rating is a little bit misleading, yea, but I don't really see the point in them anyway. Would somebody (not) learn a character because it is marked easy/hard?

I get a kick out of having an increasing number of my kanii and words being "unknown".