If your goal is to win: Rely on advanced models like KenPom [1], Sagarin [2], and Nate Silver's rankings [3] to compare teams. Everyone has their pet heuristic (always pick experience, defense wins championships, mid-majors are underrated, Duke will always win/choke -- the list goes on), but the best data-driven methods consistently outperform everything else. Sometimes actual basketball knowledge like matchups or tendencies is useful, but I only use this information for games that the models tell me are toss ups. Understand the ramifications of your scoring system. If you use a standard ... Read More