Feature
New files raw/english-5letter-sortfreq-wordsonly.txt and raw/russian-5letter-sortfreq-wordsonly.txt 
coniain new dictionaries sorted by descending frequency appearance in common texts.
Replace dictionaries in word-bank*.json to content of the new files. Pick a set of distinct 100 words for each language for 
targets from new dictionaries using gaussian distribution with mean on the first word, and sigma on the 100th - so 
most frequent words appear more frequently than uncommon words.