Financial Simulation Modeling in Excel

PUTTING TOGETHER A MORE DEVELOPED SIMULATION USING THE FUNDAMENTAL COMPONENTS

Thus far we have looked at individual components to a simulation, heavily focusing on critical basic elements such as pseudorandom number generation. While these concepts are extremely important as the basic building blocks for a simulation, we should review in more detail a few additional elementary concepts, namely iterations and sampling.

Iterations are the number of times the simulation runs. One complete cycle of the simulation is one iteration. When a low number of iterations is used there can be a large variance in the error from the simulation. In general a large number of iterations should be used to reduce such variance. What number is the right number of iterations? This is definitely project dependent and by no means definitive, but in finance you will often see Monte Carlo simulations running a minimum of 10,000 to 20,000 iterations. In some cases there have been models that we have worked on with up to 1,000,000 iterations assumed.

In certain situations we may only be able to sample from a population to infer results. We have already seen this with the coin toss example, where we know that the probability of getting a heads or tails is ½ or .5. We ran a simulation of tossing a coin and came up with probability results that were close to what is expected. The more samples that are included in a simulation, the closer we can estimate the actual probability. In summary, increasing the sampling ...