Q: After running a simulation, the expected value of the output distribution is different than the expected value that is initially shown on the worksheet. Why?
A: Most often the simulated expected value will not equal the expected value calculated during a standard Excel recalc. If your model contains non-linearities (such as two distribution functions multiplied together) the expected value can be different (the expected value of a product is not necessarily the product of the expected values). In addition, you can put functionality into a model that might not be invoked when a calculation is made using only the mean values of the input distributions. For example, an IF statement might be imbedded in a worksheet which triggers additional costs if one distribution falls above a certain threshold. The impact of this additional cost might not occur when the expected value of the distribution is used, but does appear during a simulation. This, by the way, is one of the primary benefits of simulation--that additional costs or different events such as these are reflected in the simulated result but would not be shown in a result that is simply calculated using expected values.