Learn a five-step simulation workflow for decision-making: define random variables and probabilities, set up relationships between parameters, sample repeatedly to generate outcomes, analyze the results, and ask complex questions the model can answer. Use sensitivity analysis to test how changing one variable affects your outcome, or adjust a constant to hit a target number like a loan price.
Chapters
Now let's see how we can use simulation for decision making. Here is a more generalized version of the simple simulation workflow you used in the previous exercise. Steps one and two involve defining the outcomes for random variables and assigning probabilities. We could also have constraints entering the model in this step. Step three involves defining the relationship between model parameters. In step four, we repeatedly sample from the distributions of A and B to generate outcome. Finally, in step five, we analyze outcomes simulations. Let us ask nuanced questions of the model, questions which might not necessarily have easy analytical solutions. This is because we can easily modify model inputs and see how the final outcomes are affected. For example, we might want to ask how the outcome will change if the probability distribution of B changes. In this case, we would just simulate a new set of outcomes after changing B. Let's see what happens if we do that. We might find that as a result of the change in B, the distribution of outcomes has changed significantly. This might help us decide, for instance, that we should not have a particular stock in our portfolio. It helps us see how sensitive our model is to changes in B. Another case where we use simulations is when we might want to see what values of a particular input get us the desired output. For instance, we might want to find the lowest value of the constant C for which the mean of the outcome distribution is 5. In this case, all we need to do is just iteratively keep changing the value of the constant C and record the outcomes. We will do this many times and keep recording the outcomes. Finally, we can choose the value of C where the mean of the outcome distribution is 5. Using simulation for making decisions is as easy as that. Such a simulation could, for example, help us decide the right price at which it is still profitable to invest in a loan. Now, let's take all these ideas and work through some exercises where you will evaluate a lottery ticket.
