How can you gauge sampling error in simulation output?

• A. By running only a single trial and using that result.
• B. By pressing F9 to generate a different set of simulation trials and observing the variance in output statistics.
• C. By increasing the number of decimals reported.
• D. By comparing to a fixed deterministic forecast.

Answer: (B)

Sampling error is the variability in simulation results that comes from drawing only a finite set of random trials. To gauge it, you need to see how results change when you run the model again with a fresh set of random draws. Repeating the simulation and comparing the output statistics across runs lets you observe the spread caused by sampling. In many tools, pressing a key like F9 generates a new set of simulation trials, giving you a new sample to examine. By looking at how the mean, risk measures, or percentiles vary from run to run, you get a practical sense of the sampling error. If results are similar across replications, the sampling error is small; if they vary a lot, the error is larger. More replications yield a better estimate of the standard error of the statistic. Increasing decimals only changes display precision, and comparing to a fixed deterministic forecast doesn’t quantify the sampling error.