June 23, 2026

Guess Who’s Not Impressedle?

Researchers used math to crack Wordle

Wordle cracked by math — and the internet instantly yelled, “Didn’t we do this already?”

TLDR: Researchers say a math-based strategy can beat Wordle 99% of the time by choosing guesses that reveal the most clues. Commenters were split between “cool classroom project” and “come on, the internet solved this years ago,” turning a puzzle paper into a mini drama about originality.

A team at Binghamton University says it can solve Wordle — the daily five-letter guessing game owned by The New York Times — with a 99% success rate by picking guesses that reveal the most useful clues, not just the most obvious answer. In plain English: instead of trying to be right immediately, their method plays a little smarter and squeezes more information out of each try. The twist? To actually use it live, you’d need a side program feeding you the next move like a puzzle whisperer.

But the real action was in the comments, where the reaction was less “Nobel Prize!” and more “Wait… wasn’t this already a YouTube genre in 2022?” One commenter immediately linked the paper, while others pointed out that Wordle-solving-by-math has been floating around for ages, complete with viral videos and even wild experiments using people’s colored-square brag posts on social media to reverse-engineer the answer. Ouch.

Still, not everyone came to sneer. Some defended the project as exactly the kind of nerdy, relatable idea that gets people excited about math in the first place. One fan said Wordle is such a good teaching tool that they even use “solve 100 Wordles programmatically” in job interviews because candidates love it. And of course, the classic internet side quest appeared: a blogger popped in with a cheerful “hey, I did something similar years ago!” In other words, the comments had everything: déjà vu, humblebrags, educational hype, and just enough shade to keep it spicy.

Key Points

  • Researchers at Binghamton University developed a Wordle-solving method using Shannon entropy and reported a 99% success rate in simulations.
  • The approach chooses guesses based on expected information gain rather than simply selecting the most likely answer.
  • Wordle provides feedback through grey, yellow, and green tiles that indicate whether letters are absent, misplaced, or correctly placed.
  • The article says the entropy-based strategy outperformed a traditional common-letter strategy, which solved 90% of puzzles in simulations.
  • The project began as a class assignment on information theory and later became a published paper.

Hottest takes

"I thought this was old news?" — jezzamon
"isn’t that how everyone does it?" — kps
"it is a great way to get people interested" — gkoberger
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.