Wordle, the web-based word game that gives players six tries to guess one five-letter word every day, went viral recently. Players post the numbers of guesses they take on social media, with blocks of green, yellow, and back squares revealing the letters they got right or almost right along the way.
On Jan 31, 2022, The New York Times Company acquired Wordle from its creator, Josh Wardle, for an unspecified low seven-figure amount. Suffice to say that there’s something about the puzzle which struck a chord with many.
When it comes to puzzle games like this, it’s only a matter of time before someone optimises the solving process and figures out the best strategy to maximise one’s chances of guessing the right word as quickly as possible. Grant Sanderson, on his YouTube channel 3Blue1Brown, did just that. By using information theory, which involves heady statistics and mathematics, Sanderson found the best starting word to use in Wordle.
The word is Crane.
Of course, just knowing that won’t make everyone Wordle grandmasters because the process is more complicated than that. For the best results, one has to pick the ideal second word as well based on information gleaned from Crane. It’s about maximising the amount of information one can get from the least amount of guesses possible. For those interested in learning about information theory and entropy through the lens of trying to solve the game, the video makes those complex topics digestible.
In Sanderson’s video, he explains that his Wordle-solving program makes use of a list of almost 13,000 words, which is a larger list than what Wordle actually uses. It runs through simulations using various first words followed by the best second words before arriving at CRANE as the answer. As such, this is an approach that works much better for an AI solver than a human one.
Other people have tried to “solve” Wordle as well, with programmer Adam Kubaryk reaching SOARE and SAINE as his ideal opener, and strategist Rob Miller reaching TARES as his answer.
One thing is for sure though, that even if the game can be definitively solved, since people are not machines, we can still derive a lot of joy in guessing the word of the day at our own pace and with our own preferred method.
This article was first published in Geek Culture.