Do the 'unlucky' systematically underestimate high-variance strategies? — LessWrong