I’ve been on try 6 in wordle many a time with no end in sight. Although I sure love looking through a dictionary, I decided to take a data science approach by synthesizing a wordle solving tool using a letter frequency analysis. You can try it by downloading the wordle solver from GitHub and running it yourself.

Image credit. . .nobody? (AI generated image)Image credit. . .nobody? (AI generated image)

It takes a three step approach:

  1. Use the game rules to generate a set of possible solutions
  2. Rank remaining words based on their apperance in the Oxford dictionary’s “3000 words every English speaker should know”
  3. Rank remaining words based on their likelyhood to eliminate other remaining words (letter frequency analysis)
  4. Assign a compound score to each word
  5. Sort by score (reccomended word) and allow user to play next word

Enjoy!