Human performance, psychometry, and baseball statistics — LessWrong