Do people exploit risk–reward structures to simplify information processing in risky choice?