Scientists Find Best Wordle Openers and a Winning Strategy That Doesn’t Involve Spamming Vowels

An entropy-based solver shows why the best guess is the one that narrows down your options the most.

Every morning, Wordle asks millions of players to make a small leap into the lexical dark: five blank squares, six chances, but no clues. Can you guess the five-letter word in just six tries?

Most players respond with instinct. The most common words contain some vowels, so trying words with A, E and O is a pretty sound strategy. Based on the game’s corrections, you can then try to feel your way toward the answer.

The Binghamton team tested a different way to play. Instead of choosing words that simply contain common letters, their computer program chose the word expected to eliminate the largest number of possible answers after Wordle returned its gray, yellow and green clues. The program used Shannon entropy, a mathematical measure of uncertainty, to build the winning Wordle strategy.

In simulations, that information-based strategy solved more than 99 percent of puzzles, compared with about 90 percent for a simpler approach based on letter frequency.

A Wrong Guess on Purpose Can Sometimes Be the Best Move

Wordle gives players six tries to find a hidden five-letter word. After each guess, gray tiles mark letters that are absent, yellow tiles mark letters that appear in the word but sit in the wrong place, and green tiles mark letters in the correct position.

“Let’s say you’re at a certain guess. The previous guesses will eliminate a whole bunch of options, and based on the remaining options, guessing some words will send you into a trajectory where information gain is speedier,” said Congyu “Peter” Wu, an assistant professor at Binghamton.

The researchers worked from Wordle’s list of 12,972 valid five-letter words, including 2,315 possible solution words. Because each of the five positions can return one of three outcomes — gray, yellow or green — each guess can produce 243 possible feedback patterns.

The team calculated how much each possible guess would reduce uncertainty across those patterns. The word “tares” came out on top, with an expected information score of about 6.23 bits. “Audio,” a word many players might like because it contains four vowels, scored lower, at about 4.87 bits.

×

Get smarter every day...

Stay ahead with ZME Science and subscribe.

Daily Newsletter
The science you need to know, every weekday.

Weekly Newsletter
A week in science, all in one place. Sends every Sunday.
No spam, ever. Unsubscribe anytime. Review our Privacy Policy.

Thank you! One more thing...

Please check your inbox and confirm your subscription.

“A subtle but important insight from the paper is that a guess doesn’t have to be the most likely answer; it simply has to be informative,” said Donald Stephens, a doctoral student at Binghamton.

“By applying Shannon entropy, the objective shifts to maximizing the expected reduction in uncertainty rather than the probability of being right. In practice, this approach can lead to solving the puzzle in fewer guesses.”

Why Math Beats Vibes

Shannon entropy comes from information theory, the field that the American computer scientist Claude Shannon helped launch in 1948. It now sits underneath much of our modern digital life, from data compression to communications systems. In simple terms, Shannon entropy measures how much uncertainty there is in a set of possible outcomes. In Wordle, high entropy means a guess is likely to split the remaining possible answers into smaller groups, making the next clue more useful.

Before the first guess, every possible answer is plausible. After each guess, the colored squares kill off some possibilities. A good guess produces feedback that splits the remaining answers into smaller, more useful groups.

RelatedPosts

What is entropy, and why is it always increasing?
City birds are smarter than country birds
You can arrange 128 tennis balls in 10^250 ways. A seemingly impossible problem solved
Physicists Say Time’s Arrow Could Move in Two Directions at Once

It follows that, in Wordle, the long-term winning strategy is to answer with five-letter words that reduce the search space.

The researchers compared their entropy method with a more intuitive letter-frequency strategy. A lot of players start with words that have common letters such as A, E and R. That’s not too bad. The study itself notes that most five-letter words in the list contain A, E or R, while letters such as J, Q, X and Z appear far less often. It also found that just under 70 percent of the valid English words have five unique letters, which helps explain why players often avoid repeated letters early.

But the letter frequency strategy has its own pitfalls. Words such as “least,” “stale,” “slate” and “steal” share the same letters in different orders. A strategy that chases letters without enough attention to positions can burn through guesses.

The entropy solver avoids some of those traps by updating after every round. It pays attention to the Wordle’s color feedback, then the program recommends the next word with the highest expected information among the remaining candidates.

A Better Solver, But Maybe Not a Better Game

To use the method in real time, a player would need to run software beside the puzzle and feed it each result. That may save a streak, but it’s also basically cheating.

So, using this solver defeats the purpose of the game, which is to have fun by beating it yourself, with all the frustrations involved. The project began as a class assignment, when Wu asked students to use information theory to solve a real problem. Wordle gave them an elegant test case: simple rules and a clean stream of feedback.

For most players, the lesson is to rethink the first guess. The best Wordle players aren’t necessarily those with the best vocabulary. They’re those who know how to narrow an intimidating pool of possibilities towards an easier guess.

The findings appeared in the journal Northeast Journal of Complex Systems.