10 7 the bound is optimal when it is based on information from one previous execution