In steady-state simulation the output data of the transient phase often cause a bias in the estimation of the steady-state results. A common method is to cut off this transient phase. This paper presents a new heuristic used to detect the warm-up period in steady-state simulation output. An evaluation procedure is used to compare the presented rule, called EVR, with the method MSER-m known as the most sensitive rule in detecting bias and most consistent rule in mitigating its effects. The rules are applied to the output generated by M/M/1 queuing process and the performance of the methods is tested at 4 different levels of utilizations. Various measures of goodness are used to assess the effectiveness, consistency and confidence of the methods.