When Donald Stephens first sat down to solve Wordle, he wasn't playing for fun — he was doing homework. Now, that homework has become a published research paper with a deceptively simple finding: the best guess in Wordle isn't always the one most likely to be correct.
Stephens, a doctoral student at Binghamton University, was part of a team led by Assistant Professor Congyu "Peter" Wu that applied Shannon entropy — a mathematical measure of uncertainty — to crack the popular New York Times puzzle. Their method solves Wordle 99% of the time, compared to just 90% for traditional approaches that focus on guessing common letters like "A," "E," and "R."
The breakthrough lies in a subtle but powerful reframe. Rather than starting with the most probable answer, the algorithm prioritizes guesses that eliminate the most possibilities, regardless of whether those guesses seem likely to be right. "A guess doesn't have to be the most likely answer; it simply has to be informative," Stephens explained. "By applying Shannon entropy, the objective shifts to maximizing the expected reduction in uncertainty rather than the probability of being right."
Wu first assigned the problem as a class project in the School of Systems Science and Industrial Engineering, challenging students to demonstrate how information theory could solve a real-world problem. Co-author Talal Aladaileh recalls the moment the assignment became something bigger. "The courses here don't just teach concepts; they push you to apply them in ways that have real, lasting impact," Aladaileh said. The project grew from a classroom exercise into a paper published in the Northeast Journal of Complex Systems — a testament, Wu noted, to the team's "deep understanding of class material and their talent as engineers."
The research arrives as Wordle celebrates its fifth anniversary, a milestone for a game that still draws millions of daily players. While casual players might not run entropy calculations alongside their guesses, the underlying principle applies: sometimes the path to victory isn't the obvious one. Wu described the team's innovation as transforming "a static measurement in a scientific domain into a dynamic solution that helps accomplish a popular task better."
For Stephens, Aladaileh, and their collaborators, the project illustrates something broader about how curiosity-driven learning can yield unexpected rewards. What began as an answer to a single assignment question has now been read, cited, and built upon by others in the academic community — all from a game played on five green squares.