Load forecast (LF) is the prediction of the system load over a defined time interval that ranges from one hour to one week for power system operation purposes. Due to the random nature of the load and its dependency on numerous factors forecasting is never 100% accurate. Inaccuracies in the LF lead to inefficient daily system operation. On the one hand, if at a given period the load is under-forecasted the available capacity might not be enough to meet the demand and the spinning reserve (SR) requirements. This will result in a large expectation of energy not supplied in case of generating unit outages. On the other hand if at a given period the load is over-forecasted the schedule might consider unnecessary start-ups and the provision of excessive SR. This paper explores the economical impact of the LF errors on the daily power system operation. In this paper the probability of generating units outages is considered in order to estimate the energy not served due outages. The daily operating cost of the system is divided into three terms, the start-up cost, the dispatch cost and the cost of expected energy not served due to outages of generating units