A study involving 30,954 people mowing the same virtual lawn revealed insights into human problem-solving strategies. While most participants came close to the optimal path, with 52% within five moves and 16% achieving perfection, the variety in approaches was astounding. Participants found 14,589 different ways to cover the 49 squares, yet the median efficiency remained high at 91%.

This experiment mirrored classic computer science problems like Coverage Path Planning and the Traveling Salesman Problem, where finding the shortest route is key. Humans demonstrated an impressive ability to approximate optimal solutions, even with increasing complexity. The study highlighted that as problems grow, both humans and algorithms rely on heuristics or "good enough" strategies when perfect solutions become impractical.

Analysis of individual play, like that of participant Bones, showed that initial pauses for planning were common. Critical decisions at forks in the path significantly impacted efficiency. The best performers, like Sarah, intuitively planned to tackle dead-end sections first to avoid backtracking, a strategy derived from anticipating the lawn's structure rather than just size. This decomposition and compression of learned strategies allowed for consistent performance across increasingly complex lawns.