I use Live Writer for several blogs I write, some in English, some in Hebrew. While the Word Count feature works fine for English text, it's completely broken for non-Latin characters in Hebrew text, where it counts every character
as a separate word - two 4-letter words count as 8 words. This, naturally, makes the word-count somewhat useless.
I don't know how the word-count is performed internally, but it should be relatively to use the Unicode information to distinguish between Letters, regardless of script, and non-letter whitespace which signifies a new word starting.
Thanks,
Avner Kashtan