Measuring aversion and habit strength — LessWrong