The best way to understand simulation is to run through an example. Let’s do a simulation that runs 5 years into the future. We will do the simulation 1000 times and we will use the last 5 years of historical data as a basis for return data when performing the simulation.
When we run this simulation, Stock Rover will perform 1000 individual independent simulations, where each simulation projects portfolio return over a future 5 year period. The simulation will actually operate on a monthly basis, generating random returns for each of the 60 months in the future period. The returns are based on monthly data from the historical period selected. There are also additional parameters that shape the results of the sampling run, such as the bullishness setting.
Note that 1000 simulations is a reasonable number of simulations to use. Increasing the number of simulations will result in more simulated portfolios from which Stock Rover can extrapolate, which is a good thing. But with that said, increasing the number significantly may not yield meaningfully different results, as 1000 samples will provide a very high statistical confidence level.
For each simulation, each of these steps are run for each of the 60 months:
In our example, Stock Rover is returning 1000 simulated portfolios that span the 60 months, where each of those months is populated based on the 8 steps outlined above. The monthly returns for each simulation are based on the weighted size of each ticker relative to the size of the portfolio.
Below we see 2 consecutive months from a simulation. These values were pulled from the Balance of Key Portfolios chart in the Future Simulations tool. Using the tooltip, we selected a specific portfolio and the months February 2027 and March 2027. February 2027 is showing a negative return based on the sample month of January 2015 and March 2027 is showing a positive return based on the sample month of December 2021.