Score is a decentralized computer vision subnet on Bittensor that transforms sports videos into structured, machine-readable data. It provides real-time analysis of player positions, ball trajectories, pitch geometry, and event context, enabling low-cost, high-accuracy visual insights for sports analytics, betting, scouting, and more. In use, miners submit video analysis results, while validators ensure quality through lightweight mechanisms such as frame filtering, keypoint detection, homography transforms, and CLIP-based semantic checks, with rewards distributed according to consensus.