A chessboard holds still; the floor does not
Chess gives a computer a clean snapshot: every piece is in a known square, and each legal move leads to another well-defined position. A program can compare possible moves, estimate which positions are promising, and repeat that process before choosing. The number of possibilities grows enormously, so chess engines rely on search shortcuts and evaluation rules—but the board itself does not shift while they think.
Walking presents a different kind of problem. A robot must estimate its body’s position and motion from sensors, then use that estimate to decide how to move. The information is imperfect: a camera may misjudge distance, and a foot sensor may register contact only after the foot has landed. Meanwhile, the robot’s body keeps moving.
Every step changes the next decision
To stay upright, a walking machine continually adjusts its legs and torso as its balance changes. A small slip or unexpected bump can alter where its weight is, which changes what movements will work next. The controller must respond quickly, often before it can calculate a perfect plan.
This is a feedback loop: sense the body and surroundings, make a correction, then sense again. Unlike choosing a chess move, the correction has physical consequences. A mistaken move can send the robot off balance, and the surface or contact with the ground may not behave exactly as its model predicts.
That does not make walking beyond computers. It means success depends on coordinating perception, prediction, and rapid control in the real world—not just searching a defined set of choices. A chess engine can take time to weigh a move; a walking robot has to keep its footing while it thinks.
