A computational model makes predictions about a process from selected data and features, and its predictions are tested to check it. Abstraction keeps the details that matter for the question and hides the rest. A simulation is a simplified version of something more complex, made for a purpose; it uses changing values to show how a situation changes, and it can carry bias from what was left in or out. Simulations help people form and refine hypotheses, and random numbers can stand in for real-world variety.