thetoddluci0/anti-wordle-weaponry

Anti Wordle Weaponry

I'm the kind of person who sees a cool puzzle and thinks "I wonder if I can solve that". Of course, that's usually followed by "I wonder if I can get python to solve that faster".

Thus, Anti Wordle Weaponry was born. A great way to remove any enjoyment from a fun brain teaser. The original game can be found here

How the words are chosen

First, get a list of all words from NTLK.

Second, remove all words that aren't five letters in length.

Finally, calculate a weighted score. Currently, that's done by multiplying the frequency of each letter in each word by the 'uniqueness' of the word. For example, 'apple' has a raw score of 16267, and weighted score of 13013.6. The weighted score will only ever be lower, because we multiply by uniq/5, which in the case of apple is 4/5.

All of the above is done ahead of time using generate_data.py, because it's resource intensive, and dosen't change particulatly often.

Using this data, we can then make inteligent guesses by using the highest weighted word that fits the current game state.

anti-wordle-weaponry's Issues

'v' indicator does not move some letters to new location

Tested against the word moist, marking s valid in oreas moved the s to position 0. This would be marked as a valid letter, but still wrong location, but in subsequent guesses, the s remained at position 0.