Closed NickSto closed 2 years ago
My letter-freqs.tsv is based on all words in the list, but most of those are weird and would never be answers.
If I'm actually trying to guess words with letters in the right place, I should use common words.
Basically, use the Ghent data to weight words by their recognition.
Implemented in 6ae5cfd3. Did not make any difference at all!
My letter-freqs.tsv is based on all words in the list, but most of those are weird and would never be answers.
If I'm actually trying to guess words with letters in the right place, I should use common words.
Basically, use the Ghent data to weight words by their recognition.