A robotics team's wrong diagnosis turns debugging into a study of evidence, attachment, and test design.
The robotics room had one reliable object: a red kettle that clicked off after eighty-seven seconds. Our robot was less disciplined. Its left wheel drifted, its sensor saw ghosts, and three days before competition it began reversing whenever it found the boundary line. I spent kettle cycle after kettle cycle changing code. Each correction looked elegant on my laptop and failed on the floor. By midnight, I had written a recovery routine more complicated than the route itself.
Ana watched three runs before asking why I believed every failure was software. I wanted to defend my code, which made her question useful. We taped a grid to the floor, varied one condition at a time, and recorded the angle of each turn. The error appeared only under acceleration. Beneath the shell, a loose motor mount tilted under torque. I had been teaching software to compensate for a screw. We tightened the mount, deleted most of my program, and began a new experiment: how could a team detect which layer of a system was failing before choosing a fix?
That question now follows me beyond robots. In biology lab, I ask whether an odd result comes from the organism, the instrument, or the sampling method. While organizing meetings, I separate a communication failure from a disagreement about goals. I enjoy learning most when a confident explanation becomes testable and then becomes insufficient. The kettle's click still marks our pause: eighty-seven seconds to list competing causes and let the person least attached to our favorite theory design the next test. Curiosity, for me, is not collecting clever answers. It is building fairer ways for reality—and teammates—to prove my first answer wrong.
Original teaching sample written by the Writivia team. It is not a real student's submitted essay and does not imply admission to Stanford University.