I created a Wordle helper tool to help me beat my brother at Wordle. Firstly, I performed a letter frequency analysis of all the words in the Wordle dictionary to give each word a score based on the frequency and position of the letters in it. For instance, “yeast” and “tasty” contain similar letters but since y is much more common at the end of the word than the start, “tasty” scores slightly higher with 459 compared to “yeast” at 405.
The score is calculated as the sum of the number of letters in each position, for example yeast would be y1 = 6 (as there are 6 words in the Wordle dictionary that start with a “y”) + e2 = 164 + a3 = 283 + s4 = 128 + t5= 229 = 810 and divided by 2.The score was divided by two so that the maximum score < 999 which made displaying the ordered score in the command window easier. It then returns an ordered list of all possible words.
The algorithm does not necessarily solve the word in less steps than a person as luck plays a large role, but it is more robust. For instance, starting with one of the best starting words “slate” and starting with one of the worst words “index” has a small effect.