Halloween brings to mind the monsters that keep simulation writers up at night. One of the scariest - and we've all met it at least once - is the Creeping Scope. This boojum appears when projects seem to be going well; on time and under budget. "Since we're doing so well," the Creep insinuates, "why not add a little functionality. We're here anyway. It won't take that long." Then, quick as a flash of lightning over a bubbling cauldron, the goals, objectives and definition of success for the project transform themselves beyond recognition. The rest of the project can turn into a long, frustrating game of catch-up. In extreme cases, this can lead to abandoning a project altogether.

Here are three important steps to take in order to avoid starring in your own scary story:

Get the stakeholders to buy in early and often. Keeping people involved in the project and up to date on its progress will increase your chances for success. As a model becomes more and more complex, it will take more time for anyone other than the modeler to completely understand the workings of the model. This unavoidable fact reinforces the need to keep people involved in the assumptions, simplifications, scope, and limits of the model.

Build the model to solve a specific need. As basic as this sounds, many people new to simulation simply want to model the entire system. Instead, build the model around the questions you want answered. If the questions change, then your model may need to change as well. Nothing will kill your project faster than trying to take on too much.

Create a list of deliverables for the project and agree on them. It is very easy for your client (or your supervisors) to ask for features that sound simple enough but are extremely difficult to model. Make sure any change requests remain within the defined purpose and scope of the model, and then manage them carefully.

Happy Halloween.