I was just wondering whether it is possible to somehow have skritter make me a list of words that consist of only charcters I know? Personally I would much rather focus on learning new words from the characters I know, rather than new characters every day. It would be awesome if skritter could recommend words to me that only have characters I already know in them. :-)

I'd like to see a feature like this as well. Right now I'm doing it on my own. I compiled a megalist of the Jouyou Kanji (Japanese). Whenever I learn a new kanji, I go to the details menu and search for compounds with kanji I already know. An auto-recommend function would be nice though...learning the compounds really helps reinforce what the main character means!

I agree, it would be nice if Skritter could automatically compile a list of "Words you might learn easily" based on single characters already learned well (high rates of retention). In my experience those sort of words are learned almost instantly and kept pretty consistently, but of course one has to be exposed to the concatenation at least once first.

For Chinese, there's this tool:

http://huygens.functor.nl/skritter/wordlist/

It hasn't been updated with a new Skritter word list in a while, so it doesn't have the extra words we've added since then. But most of that difference will be in rarer words.

That's a great tool - thanks for the link!

Now the issue of choosing which of the 12,000+ suggested words (from ~1000 characters I've learned) to add!

Just out of curiosity, what are the frequency measures? I guess a number like 32915 (中国) means extremely frequent, but do you know how that is derived? A count from common texts/literature?

It's a frequency count out of a standard Chinese corpus with 4,842,317 words in it (20,768 unique words).

Beware of adding too many words like that, though - I did the same thing for a while and now I've got hundreds of words in my studies that are basically useless to me as I don't ever hear them and I don't know how to use them.

I'd also advise against copying and pasting all these new words directly, especially if the results returned are of any large quantity. Perhaps carefully check them and feed them through a list?

Yea, I wouldn't just mass copy and paste, but rather quickly glance through the first 1000 or so and pick out <100 that seem particularly useful or that I've learned before but haven't added to Skritter. It's something I'll do occasionally (once month maybe) to try and add low-hanging fruit that is clearly intuitive or already studied but not in Skritter. Not interested in boosting my Skritter list by 一万 in one day, as I know the reviews that will bring tomorrow!