NYT’s new WordleBot tries to gamify something that probably doesn’t need it

theverge.com:

We are several months into our new Wordle era (along with an expanded universe of clones like Worldle, Heardle, Numble, Poeltl, and many others) and a little over two months into The New York Times acting as stewards of the game after a “low seven figures” purchase. Besides altering the word list, the Times is introducing its biggest change yet by adding WordleBot, a tool that serves as a “daily companion” for your new favorite puzzle game.

Like Wordle itself, WordleBot operates in the browser and runs locally, evaluating the completed puzzles on whatever computer and stored session you’re using to play. As The New York Times describes it, the bot started as an attempt to mathematically prove the best starting word, developed with its The Upshot section that focuses on data visualizations for the news.

Of the 4,500 five-letter English words WordleBot chooses from, the NYT says “crane” in easy mode and “dealt” in hard mode are the ones that get the bot to the answer in the fewest number of guesses. However, you may play differently and know different words, which doesn’t make them the best starting guesses for you.

WordleBot tries to take into account both skill and luck. It has the list of all Wordle answers but doesn’t know which one is the answer on a particular day. As a result, what they’ve built is a kind of standardized comparison to measure how well you did at figuring out the day’s puzzle and suggest ways for you to improve or compare who did better even when you used the same number of guesses. Yes, analytic breakdowns are now a part of Wordle.

If you’re obsessed with efficiency and welcome the optimization advanced stats can provide, then pressing a button and seeing what the bot thinks of your process could